irc.oftc.net #zumastor log beginning Sat Mar 1 00:00:01 PST 2008
2008-03-01 02:36 -!- pgquiles(~pgquiles@90.Red-83-34-134.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-01 09:18 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-01 11:49 <flipz> http://www.linuxtoday.com/news_story.php3?ltsn=2008-02-29-008-26-RV-HW-0003 <- lol
2008-03-01 16:40 -!- Iggy_Koopa(~iggy_koop@pool-71-126-228-65.bstnma.east.verizon.net) has joined #zumastor
2008-03-01 16:40 <Iggy_Koopa> hey
2008-03-01 16:40 <Iggy_Koopa> Is anyone active?
2008-03-01 17:47 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-01 21:30 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-01 21:41 <flipz> hi iggy_koopa
2008-03-01 23:27 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Sun Mar 2 00:00:01 PST 2008
2008-03-02 12:34 -!- charlesn1(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-02 13:41 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-02 14:27 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Mon Mar 3 00:00:01 PST 2008
2008-03-03 00:50 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-03 05:50 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has left #zumastor
2008-03-03 05:50 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-03 07:50 -!- willn(~wan@pinball.ccs.neu.edu) has joined #zumastor
2008-03-03 07:50 <willn> 
ACTION blinks

2008-03-03 10:44 <flipz> good morning
2008-03-03 10:49 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-03 12:41 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-03 14:17 <willn> Someone needs to update or remove our Todo file
2008-03-03 14:23 <shapor> hm, it could just point to the issue tracker now
2008-03-03 14:23 <shapor> it predates our use of the code.google.com issue tracker
2008-03-03 14:41 <flipz> Todo could just say: See our issue tracker (url)
2008-03-03 14:42 <shapor> is there an echo in here? :P
2008-03-03 14:43 <flipz> heh
2008-03-03 14:43 <flipz> hollow head -> great echo chamber
2008-03-03 14:47 <flipz> charlesnw of ubuntu?
2008-03-03 14:47 <flipz> somebody left me a nice silver copy of a fake steve jobs book on my desk
2008-03-03 14:47 <flipz> 
ACTION suspects tim_vimm with a co-conspirator

2008-03-03 14:48 <tim_vimm> wasn't me  
2008-03-03 14:48 <flipz> ooh
2008-03-03 14:48 <tim_vimm> who could it have been?
2008-03-03 14:48 <tim_vimm> is it a signed copy?
2008-03-03 14:48 <flipz> fake steve jobs? shapor?
2008-03-03 14:49 <flipz> let me check
2008-03-03 14:49 <flipz> signed
2008-03-03 14:49 <flipz> how bout that
2008-03-03 14:50 <flipz> probably just somebody found the one I left lying around
2008-03-03 14:50 <tim_vimm> how was it signed?
2008-03-03 14:50 <flipz> I got it figured out
2008-03-03 14:50 <tim_vimm> lol
2008-03-03 14:50 <flipz> now need to read it
2008-03-03 14:50 <tim_vimm> should put that in the todo file
2008-03-03 14:50 <flipz> and get back to him even
2008-03-03 14:51 <tim_vimm> he's ventured into weird waters lately
2008-03-03 14:51 <flipz> wierd =  me
2008-03-03 15:05 <flipz> going to post the alternative interface to device mapper to lkml tomorrow morning I think
2008-03-03 15:05 <flipz> something like "[ANNOUNCE] ddsetup alternative interface to device mapper"
2008-03-03 15:05 <flipz> or maybe "[RFC] ddsetup alternative interface to device mapper"
2008-03-03 15:07 <flipz> "Q: if you have an axe with a rusty head and a rotten handle, how do you repair it? A: it is a two step process.  First you replace the handle, then you replace the head"
2008-03-03 15:19 <flipz> static void exit_errno(char *action)
2008-03-03 15:19 <flipz> {
2008-03-03 15:19 <flipz> 	error("ddsnap: could not %s: %s (%i)", strerror(errno), errno);
2008-03-03 15:19 <flipz> }
2008-03-03 15:19 <flipz> static void exit_error(char *action, char *reason)
2008-03-03 15:19 <flipz> {
2008-03-03 15:19 <flipz> 	error("ddsnap: could not %s because %s", reason);
2008-03-03 15:19 <flipz> }
2008-03-03 15:29 <flipz> next major cleanup from me really really has to be getting the user space message codes out of the kernel header
2008-03-03 15:38 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-03 16:01 <flipz> unknown_message_handler() is called 12 times in ddsnap.c
2008-03-03 16:01 <flipz> :p
2008-03-03 16:46 <flipz> ok, there goes my plea for Andrew Hutton to ignore the fact that we are two weeks late proposing a paper, and take a submission on LVM3
2008-03-03 17:19 <charlesnw> flipz: charlesnw the one and only :)
2008-03-03 17:25 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has left #zumastor
2008-03-03 17:31 -!- charlesn1(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-03 17:39 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-03 18:16 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-03 19:13 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-03 19:49 <flipz> the 9" eee will have 1 GB memory, up to 12 GB flash, and 1/3 higher price
2008-03-03 19:50 <flipz> I'm in :-)
2008-03-03 20:03 <shapor> sounds cool
2008-03-03 20:03 <shapor> i'm not a huge fan of the form factor
2008-03-03 20:03 <flipz> I can type fast on it, and plug it into a vga projector
2008-03-03 20:04 <flipz> 90% of what I need on a road trip
2008-03-03 20:04 <flipz> got to see how well uml kernel development works
2008-03-03 20:04 <flipz> I suspect, ok, recent bloatware compiler is going to be a little slow
2008-03-03 20:05 <flipz> bloatware compiler compiling bloatware kernel
2008-03-03 22:28 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Tue Mar 4 00:00:01 PST 2008
2008-03-04 00:06 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-04 02:01 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-04 04:30 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-04 04:30 <dank> pgquiles, you there?
2008-03-04 04:33 <pgquiles> dank: hi
2008-03-04 04:35 <dank> So, are you really going to use a 5 second snapshot interval in production?
2008-03-04 04:35 <pgquiles> dank: no
2008-03-04 04:36 <pgquiles> dank: but the math is 5 seconds snapshot interval for a 5GB volume is roughly equivalent to 30 min or 1 hour snapshot period in a 2TB volume
2008-03-04 04:37 <pgquiles> the problem is unmounting, applying the delta and mounting takes some time
2008-03-04 04:38 <pgquiles> if there are quite some changes in the origin volume, the delta will take some time to be applied
2008-03-04 04:38 <pgquiles> therefore unavailability of the volume suddenly is important
2008-03-04 04:38 <pgquiles> one interesting thing to note is this problem was not present a couple of weeks ago
2008-03-04 04:39 <pgquiles> dank: I'm testing 5GB/5seconds replication as a worst-case scenario, if zumastor is able to deal with that, I can quite safely assume I won't have problems with 1TB or 2TB volumes and 1 hour replication periods (probably even longer)
2008-03-04 04:42 <dank> OK.  That wasn't clear in the bug report.  I'll add that info.
2008-03-04 04:42 <pgquiles> dank: oh, sorry
2008-03-04 04:43 <pgquiles> anyways, some sort of lsof-checking should be done before unmounting/applying delta/mounting
2008-03-04 04:44 <pgquiles> consider the following scenario: a zumastor volume is used as the origin for an FTP server which supports resuming
2008-03-04 04:44 <pgquiles> you start downloading the file
2008-03-04 04:45 <pgquiles> a new version of the file replaces the old file, filesystem is unmounted, delta applied, filesystem is remounted and the FTP client resumes the download
2008-03-04 04:45 <pgquiles> you successfully download the file but the file is actually corrupt
2008-03-04 04:46 <dank> lsof wouldn't help in that case
2008-03-04 04:46 <pgquiles> why?
2008-03-04 04:46 <dank> Nothing will.  You have to adopt a policy that files, once created, never change.
2008-03-04 04:47 <dank> Only immutable files can be resumed.
2008-03-04 04:47 <pgquiles> immutable files? should I replace my hard disks with stone tables? :-D
2008-03-04 04:47 <dank> HTTP allows resuming after the user has dropped the connection, so no open file is available to pin the snapshot.
2008-03-04 04:48 <dank> Where I work, we have exactly such a policy; our big software distribution volume is write-once.
2008-03-04 04:49 <dank> It's enforced like this: a daemon scans the volume, looking for files older than 60 minutes, and when it finds them, it chowns them and makes them read-only.
2008-03-04 04:49 <dank> That way users can never be surprised by files changing in the middle of a resumed download.
2008-03-04 04:50 <dank> (This is an internal corporate volume; I have no idea how external download volumes are managed.)
2008-03-04 04:50 <pgquiles> I was thinking of something simpler: if lsof | grep /dev/mapper/zumatest != null, don't unmount and apply the delta
2008-03-04 04:51 <pgquiles> but that may lead to never applying deltas
2008-03-04 04:51 <pgquiles> :-/
2008-03-04 04:51 <dank> Back to the 5 second replication issue.  If a 1TB volume was offline for 5 seconds during switch to a new snapshot, would that be ok?
2008-03-04 04:51 <pgquiles> mmm it's not that easy
2008-03-04 04:52 <pgquiles> the offline period should be proportional to the replication period
2008-03-04 04:52 <dank> I'm not asking that.
2008-03-04 04:52 <dank> I
2008-03-04 04:52 <pgquiles> I know :-)
2008-03-04 04:52 <pgquiles> yes, 5 seconds if perfectly acceptable
2008-03-04 04:52 <dank> I'm asking what the absolute upper bound is on a volume being offline during the switch to a new snapshot.
2008-03-04 04:52 <pgquiles> s/if/it's
2008-03-04 04:53 <dank> Some users might not like 5 seconds.  Good to know it's ok with you for now.
2008-03-04 04:53 <pgquiles> for me 5 seconds is fine
2008-03-04 04:54 <pgquiles> I will have a 1hour replication period in-LAN and 1 day replication period over WAN, so 5 seconds is very good
2008-03-04 04:54 <dank> ok
2008-03-04 04:54 <pgquiles> it's 5am there!?
2008-03-04 04:54 <pgquiles> :-D
2008-03-04 04:55 <dank> Now, about the "lsof" idea.   Perhaps we can provide a hook to let you implement that policy yourself.  i.e. when we notify apps of the incoming snapshot, maybe they should be able to say "not yet"
2008-03-04 04:55 <pgquiles> yes, that would be nice
2008-03-04 04:55 <dank> Yeah, my 4 year old son couldn't sleep in his own bed, so he's sleeping in ours, and we're up.
2008-03-04 04:55 <pgquiles> :-)
2008-03-04 04:56 <pgquiles> hooks would be nice, although that's low priority IMHO
2008-03-04 04:56 <dank> right
2008-03-04 04:57 <dank> !$&@?!! issue tracker is not taking my update...
2008-03-04 04:57 <pgquiles> 502 for me :-(
2008-03-04 05:01 <pgquiles> lunch time here, good luck with the kid :-)
2008-03-04 05:01 <dank> whew, it's in.  Bon appetit.
2008-03-04 06:31 <willn> 
ACTION blinks

2008-03-04 06:38 <dank> Say, Will, I heard a rumor you were in Boston.  Is it true they eat ice cream there in winter?
2008-03-04 07:41 <willn> dank: It is.
2008-03-04 07:59 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-04 08:21 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-04 08:52 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-04 10:23 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-04 10:23 -!- pgquiles_(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-04 11:01 <flipz> hmm, could it be I have never posted the device mapper interface rewrite to our list?
2008-03-04 11:27 <pgquiles_> flipz: I recall something about the rewriting but I cannot remember if it was on IRC or mailing list
2008-03-04 11:27 <pgquiles_> flipz: what the status of libddsnap/libzumastor? (ddlink?)
2008-03-04 11:34 <flipz> pgquiles, libddsnap etc will get a little closer when the new device mapper interface becomes available
2008-03-04 11:44 <pgquiles_> flipz: do you have an ETA for that new devmapper interface? (I'm anxious to use lib* :-) )
2008-03-04 11:46 <flipz> pgquiles_, eta is this afternoon
2008-03-04 11:46 <flipz> I'm writing the lkml post right now
2008-03-04 11:48 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-04 12:00 <pgquiles_> flipz: cool
2008-03-04 12:38 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-04 12:52 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-04 16:22 -!- cbsmith(~xman@207.47.98.132.static.nextweb.net) has joined #zumastor
2008-03-04 17:43 -!- cbsmith(~xman@207.47.98.132.static.nextweb.net) has joined #zumastor
2008-03-04 19:13 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-04 19:32 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-04 20:56 -!- cbsmith(~xman@adsl-71-133-80-65.dsl.irvnca.pacbell.net) has joined #zumastor
2008-03-04 22:32 <flipz> ddsetup post is finally finished... almost
2008-03-04 22:33 <flipz> need to upload code examples to zumastor.org
2008-03-04 22:55 <tim_vimm> a peek at the future of ddsetup post?
2008-03-04 22:55 <flipz> something like that
2008-03-04 22:55 <flipz> shapor, ping?
2008-03-04 23:14 <flipz> [RFC] An alternative interface to device mapper
2008-03-04 23:48 <flipz> why do I not see my files checked into svn showing up on zumastor.org?
2008-03-04 23:50 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-04 23:55 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Wed Mar 5 06:05:34 PST 2008
2008-03-05 00:42 <flipz> now to do the third proofread... add addresses... hit send
2008-03-05 00:42 <flipz> I hope I can stay awake
2008-03-05 01:11 <flipz> gah, I can't read through it again
2008-03-05 01:11 <flipz> there is always one more typo
2008-03-05 01:11 <flipz> well... maybe grab a cup of tea and read through it again
2008-03-05 01:13 <cbsmith> anyone still up?
2008-03-05 01:15 <flipz> not me
2008-03-05 01:15 <cbsmith> lol
2008-03-05 01:15 <cbsmith> I was reading over Jiaying's discovery with the x4500 bug.
2008-03-05 01:16 <cbsmith> It occurs to me that doing O_DIRECT with sector size writes is going to be pretty slow with any RAID-5 or RAID-6 config, no?
2008-03-05 01:16 <flipz> it should not be happening
2008-03-05 01:16 <flipz> bug somewhere
2008-03-05 01:17 <cbsmith> Ah, so there is an intent to write stride sized bits to the dasd?
2008-03-05 01:18 <cbsmith> s/dasd/disk/
2008-03-05 01:18 <cbsmith> Damn, my old days with mainframes are showing.
2008-03-05 01:18 <flipz> no, it is just that ddsnap does not use sector size transfers anywhere, neither does ext3
2008-03-05 01:20 <cbsmith> Well, with ext3 you can configure the stride size, so i guess if it was misconfigured (although that would require some fairly profound PEBKAC-ery) it could happen.
2008-03-05 01:26 <cbsmith> flipz: So, the zumastor journal writes couldn't end up being really small?
2008-03-05 01:26 <flipz> no
2008-03-05 01:26 <cbsmith> I'll double check the machine then,  but I'm all but certain it was set up with a stride larger than 512 byes, and probably the correct stride.
2008-03-05 01:28 <flipz> shoud I say: [RFC] An alternative interface to device mapper
2008-03-05 01:28 <flipz> or: [RFC] An alternative interface for device mapper
2008-03-05 01:29 <cbsmith> flipz: Hmm.... "for" implies more that it is device mapper that is interfacing with it. "to" implies more that it is other code that is interfacing with it.
2008-03-05 01:29 <flipz> done
2008-03-05 01:29 <flipz> pulled the trigger
2008-03-05 02:10 <flipz> http://lkml.org/lkml/2008/3/5/82
2008-03-05 04:21 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-05 05:47 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-05 05:48 <dank> shapor, the irc log didn't roll over?
2008-03-05 05:58 <shapor> dank, hrm
2008-03-05 06:01 <shapor> probably related to:
2008-03-05 06:01 <shapor> Password expires                                        : Mar 04, 2008
2008-03-05 06:02 <shapor> that could also explain: 23:48 < flipz> why do I not see my files checked into svn showing up on zumastor.org?
2008-03-05 06:08 <shapor> fixed
2008-03-05 07:39 <willn> 
ACTION pokes packaging

2008-03-05 08:45 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-05 08:47 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-05 09:01 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-05 10:33 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-05 11:32 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-05 12:14 -!- charlesn1(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-05 12:35 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-05 13:52 <flipz> shapor, ping?
2008-03-05 14:00 <shapor> flipz pong
2008-03-05 14:00 <flipz> so...
2008-03-05 14:00 <flipz> how about we have a static reason[200] buffer
2008-03-05 14:00 <flipz> and you just strcpy your error message there?
2008-03-05 14:07 <pgquiles> flipz: after reading your e-mail about ddlink, it seems to me it's too low level for what I wanted
2008-03-05 14:08 <flipz> what did you want?
2008-03-05 14:08 <pgquiles> something higher level, at the ddsnap level at least, zumastor level would have been even better
2008-03-05 14:09 <flipz> to do what?
2008-03-05 14:09 <pgquiles> a GUI
2008-03-05 14:09 <pgquiles> without scraping zumastor output
2008-03-05 14:09 <pgquiles> ddsnap would be built on top of ddlink, isn't it?
2008-03-05 14:10 <flipz> entirely correct
2008-03-05 14:10 <pgquiles> if I were to develop a GUI using ddlink now, I'd be essentially reimplementing ddsnap
2008-03-05 14:10 <flipz> ddlink is not what you want
2008-03-05 14:10 <flipz> ddlink is strictly for userspace/kernel communication
2008-03-05 14:11 <pgquiles> but I do not mean ddlink looks wrong to me, it looks very good ('sane' is the word I'm actually thinking :-) )
2008-03-05 14:11 <flipz> we use unix domain sockets for communicationing with a ddsnap server, that is a pretty nice interface to build on
2008-03-05 14:12 <pgquiles> 'we' meaning zumastor communicates with ddsnaps through unix sockets?
2008-03-05 14:12 <flipz> zumastor calls ddsnap, which communicates over the unix sockets
2008-03-05 14:12 <flipz> a C program could leave out the "calling ddsnap" step
2008-03-05 14:13 <pgquiles> mmm I've not read the source but, wouldn't that mean I'm reimplementing ddsnap and zumastor?
2008-03-05 14:14 <flipz> depends what you want to do
2008-03-05 14:14 <flipz> for a program that just wants to control snapshotting, it might be ideal
2008-03-05 14:14 <flipz> if you want to do replication... well, ddsnap does that and perhaps that is the bit that should become a library
2008-03-05 14:14 <flipz> a replication library
2008-03-05 14:15 <flipz> would you like to propose a library api on the mailing list?
2008-03-05 14:15 <pgquiles> I don't know ddlink, devmapper, ddsnap or zumastor enough to propose something sensible, I believe
2008-03-05 14:16 <flipz> what I would suggest is waiting a little while for the device mapper interfaces to be sanitized, and ddsnap to be simplified as a result
2008-03-05 14:17 <pgquiles> ok
2008-03-05 14:17 <flipz> then we will be in position to define a sane library for replicating, snapshot create, revert, status, etc
2008-03-05 14:17 <pgquiles> more or less the same commands 'zumastor' has
2008-03-05 14:17 <flipz> yes
2008-03-05 14:18 <pgquiles> IMHO ddsnap should be a dummy application with one-liner functions which call libddsnap functions, little more than an option parser to call libddsnap functions
2008-03-05 14:18 <pgquiles> same thing for zumastor and a hypotetical libzumastor
2008-03-05 14:19 <flipz> you need one more ingredient to have a sane library: other users
2008-03-05 14:19 <pgquiles> of course
2008-03-05 14:19 <flipz> a specific proposal would help there
2008-03-05 14:20 <flipz> I propose this test to determine fitness of a ddsnap library: would you want to link your storage application to it?
2008-03-05 14:21 <flipz> I am thinking, suppose I was Sun, would I want to link ZFS to libdevmapper?
2008-03-05 14:21 <flipz> ...resounding no
2008-03-05 14:21 <flipz> and why not?
2008-03-05 14:21 <flipz> hmm, because it only addresses one class of block devices, and not very well?
2008-03-05 14:22 <flipz> because it presents an absurd interface?
2008-03-05 14:22 <pgquiles> so what's your proposal? linking to a lower level (ddlink) or linking to a higher level (libzumastor)?
2008-03-05 14:25 <pgquiles> the problem with linking to ddlink is, as every GUI would be reimplementing ddsnap, the slightest bug in that GUI means your NAS is not a zumastor NAS but a zumastorblah NAS
2008-03-05 14:26 <pgquiles> and that's assuming the bug in 'zumastorblah' does not lead to data loss or corruption
2008-03-05 14:31 <flipz> I don't have a proposal yet, that is why I am trying to convince somebody else to make one
2008-03-05 14:32 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-05 14:32 <flipz> we should rewrite zumastor in c++ and modularize while doing that, not necessarily create a library but at least separate interface from backend
2008-03-05 14:33 <flipz> so we can have multiple interfaces
2008-03-05 14:33 <flipz> then if we did that, we might have a user for libddsnap
2008-03-05 14:33 <flipz> does that make sense?
2008-03-05 14:33 <pgquiles> it certainly does
2008-03-05 14:33 <cbsmith> 
ACTION likes the idea of getting to flex some C++ muscles

2008-03-05 14:34 <pgquiles> this weekend I was reading about ZFS and the zfs command and I was thinking all the time 'zfs' seems roughly equivalent to 'ddsnap'
2008-03-05 14:34 <flipz> interesting
2008-03-05 14:34 <cbsmith> Yeah, and flipz hates the zfs command ;-)
2008-03-05 14:34 <pgquiles> i. e. you could have 'zumastor' with a ddsnap backend, 'zumastor' with a ZFS backend, etc
2008-03-05 14:35 <flipz> I did not really look at the zfs command very closely
2008-03-05 14:35 <pgquiles> probably true for Hammer or even git 
2008-03-05 14:35 <flipz> I critizied the use of environment variables I think
2008-03-05 14:35 <pgquiles> http://eigenclass.org/hiki/gibak-backup-system-introduction
2008-03-05 14:35 <cbsmith> As I catch up on the irc logs, I'm finding pgquiles proposal very interesting.
2008-03-05 14:36 <pgquiles> cbsmith: I didn't propose anything, flipz is desperately trying to get someone to make a proposal and I'm trying to avoid it by all means :-P
2008-03-05 14:37 <cbsmith> pgquiles: You're thinking about it, and flipz is luring me in to working on a proposal. ;-)
2008-03-05 14:38 <pgquiles> :-D
2008-03-05 14:39 <pgquiles> I started to work on a GUI (it was just a toy, with no error handling and very tied to my specific configuration)
2008-03-05 14:40 <pgquiles> then, as zumastor needs to run as root, I diverted to a libsudo
2008-03-05 14:41 <cbsmith> pgquiles: You don't need a libsudo, just a client-server kind of split, probably connected by pipes, which fits very nicely with your modular backend notion.
2008-03-05 14:42 <pgquiles> cbsmith: how do you manage zumastor running as root and the GUI as non-root, then?
2008-03-05 14:42 <pgquiles> suid for a webapp looks awful to me
2008-03-05 14:43 <flipz> pgquiles, ok I remember now, you had a permsissions issue that I thought we could address
2008-03-05 14:43 <flipz> not by you using ddlink, but by us using it to let ddsnap run as a normal user
2008-03-05 14:43 <shapor> flipz: hm static
2008-03-05 14:43 <pgquiles> yes, but that's a different issue
2008-03-05 14:43 <cbsmith> pgquiles: You have the zumastor bits running as a daemon on one end of the pipes, and the GUI running as a member as some "snappshot admin" group on the other end.
2008-03-05 14:43 <flipz> I still think we need to do this
2008-03-05 14:44 <flipz> shapor, errno is static too
2008-03-05 14:44 <shapor> yeah
2008-03-05 14:44 <pgquiles> ddsnap running as non-root still implies scraping ddsnap output
2008-03-05 14:44 <shapor> errstr :)
2008-03-05 14:44 <flipz> pgquiles, and we better do something about that scraping
2008-03-05 14:44 <flipz> two issues so far, any more?
2008-03-05 14:45 <pgquiles> not that I can come with now but give me a minute an infinite rows in the issue tracker database :-P
2008-03-05 14:45 <flipz> 1) ddsnap runs as nonroot 2) should not have to parse human oriented output from ddsnap
2008-03-05 14:50 <shapor> what he wants is probably not just ddsnap
2008-03-05 14:50 <shapor> but zumastor
2008-03-05 14:50 <shapor> zumastor needs an api
2008-03-05 14:51 <cbsmith> which would work better if zumastor wasn't a bash script ;-)
2008-03-05 14:51 <flipz> I think so
2008-03-05 14:51 <flipz> quick summary is: wants to write an api for zumastor
2008-03-05 14:51 <cbsmith> But in general, I think zumastor should be refactored as a very thin wrapper around an API, as is traditional for Unix.
2008-03-05 14:52 <flipz> that translates into: if you want to do it now, you will be hacking some bash.  Not the worst thing in the world (second worst maybe)
2008-03-05 14:52 <flipz> if you want to do it in the fullness of time, then we need to convert to a real language
2008-03-05 14:53 <cbsmith> flipz: Yes, exactly.
2008-03-05 14:54 <pgquiles> do you mean API as in #include <zumastor.h> or as in "scrape error output of system('zumastor command options'), here is the error list"?
2008-03-05 14:55 <cbsmith> pgquiles: I mean it in the sense of #include <zumastor.hpp>
2008-03-05 14:55 <flipz> pgquiles, I misspoke
2008-03-05 14:55 <pgquiles> cbsmith: I love you
2008-03-05 14:55 <cbsmith> I think everyone else understands that as the meaning.
2008-03-05 14:56 <cbsmith> 
ACTION blushes

2008-03-05 14:56 <flipz> I meant to say pgquiles wants to write a gui, not an api
2008-03-05 14:56 <pgquiles> me and every NAS vendor out there thinking in using zumastor :-)
2008-03-05 14:57 <flipz> how nice it would be
2008-03-05 14:58 <cbsmith> We just need pgquiles to write an API spec. ;-)
2008-03-05 14:58 <pgquiles> :-D
2008-03-05 14:58 <flipz> picked up a major gem of knowledge last week at fast: heard that ext3 is twice as fast as zfs on the same hardware
2008-03-05 15:00 <pgquiles> good to know
2008-03-05 15:01 <pgquiles> what about reiser4? sadly since the trial started reiser4 fans seem to have vanished
2008-03-05 15:03 <cbsmith> more like gone to ground
2008-03-05 16:32 <pgquiles> time to sleep
2008-03-05 16:32 <cbsmith> pgquiles: I bet
2008-03-05 22:42 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Thu Mar 6 00:00:01 PST 2008
2008-03-06 00:49 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-06 01:17 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 08:18 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 09:36 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-06 10:05 <willn> meeting?
2008-03-06 10:42 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 10:52 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 10:52 <dkegel> Met.
2008-03-06 10:56 <dkegel> Deferring bugs clarifies the mind.
2008-03-06 11:10 <flipz> meet or not?
2008-03-06 11:15 <dkegel> We done met already.
2008-03-06 11:16 <dkegel> Blame Jiaying for the early hour.  Amazingly, we had pretty good attendance, considering.
2008-03-06 11:18 <flipz> see you in a bit
2008-03-06 11:18 <flipz> I could have made it, really
2008-03-06 11:18 <flipz> with my new scramjet shoes
2008-03-06 11:43 <dkegel> I thought you bought a pair of those last year.
2008-03-06 11:43 <flipz> these are my inline scramjets
2008-03-06 11:43 <flipz> got them week before last
2008-03-06 11:44 <flipz> now if I can figure out how to keep the sun off my head while travelling at speed I will use them now
2008-03-06 11:44 <flipz> but I can't so I won't
2008-03-06 11:44 <flipz> back to the old scramjets
2008-03-06 11:45 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 11:50 <flipz> my lkml post is number 8 google hit for "alternative interface"
2008-03-06 11:51 <flipz> the jet turbine racecar quote made it to this week's lwn kernel quotes section
2008-03-06 12:10 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 12:18 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 13:13 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 13:26 -!- natalie(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 13:28 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 13:42 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 13:54 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 14:28 <dkegel> "zumastor replicate volname" seems to always be ignored for me on the first try
2008-03-06 14:30 <dkegel> Doing a sleep 10 first also fixes it, so it seems there's a race condition?
2008-03-06 14:30 <cbsmith> creepy
2008-03-06 14:30 <cbsmith> Wait, doing a "sleep 10" and then invoking the first time works?
2008-03-06 14:33 <cbsmith> FYI, I just confirmed in with Dan (in the real world) that things work if he sleep 10's after defining the volumes, but before his first replication invocation.
2008-03-06 14:33 <cbsmith> Sounds like that asynchronous command execution is hurting us again.
2008-03-06 15:33 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 15:55 <flipz> cbsmith, be better have concrete plan what to do with that aio deathray
2008-03-06 15:55 <flipz> jiayingz has some specific ideas
2008-03-06 15:56 <flipz> me too
2008-03-06 15:56 <flipz> s/be/we/
2008-03-06 15:57 <flipz> syncifying zumastor commands better be number one bug for 0.8
2008-03-06 16:00 <cbsmith> flipz: why don't you guys share those plans with me next week and we can see where to go from there.
2008-03-06 16:01 <flipz> ready to share any time
2008-03-06 16:01 <flipz> isn't jiaying 6 feet away at the moment?
2008-03-06 16:01 <cbsmith> yeah, I'm just trying to focus on getting the last bug cleared out for 0.7
2008-03-06 16:02 <flipz> she talked about adding a parallel thread for writeout, but aio is better, and just poll... I hope...
2008-03-06 16:02 <cbsmith> it should be
2008-03-06 16:02 <flipz> 
ACTION gets the sinking feeling you can't just poll

2008-03-06 16:02 <flipz> is libaio pollable?
2008-03-06 16:02 <cbsmith> it is
2008-03-06 16:03 <flipz> what is the call?
2008-03-06 16:03 <cbsmith> they have an API that returns back all the completion events, and another one that waits for completion events for X seconds.
2008-03-06 16:03 <flipz> what is the name of the man page for that poll api?
2008-03-06 16:03 <flipz> 
ACTION looks at aio_read

2008-03-06 16:04 -!- natalie(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 16:04 <flipz> aio_read looks like it
2008-03-06 16:04 <flipz> well
2008-03-06 16:04 <flipz> ahem
2008-03-06 16:04 <cbsmith> io_getevents
2008-03-06 16:04 <cbsmith> aio_read is to submit a read... IIRC
2008-03-06 16:04 <cbsmith> oh, there is io_poll too :-)
2008-03-06 16:04 <flipz> oh great, not referenced from the aio man page
2008-03-06 16:04 <flipz> that is a man page bug
2008-03-06 16:05 <cbsmith> but io_getevents is probably closer to what you want
2008-03-06 16:05 <cbsmith> There is some unfortunate overlap between libaio and POSIX aio man pages, IIRC
2008-03-06 16:05 <flipz> no man page for io_poll
2008-03-06 16:05 <flipz> all of aio seems to be one big unfortunate
2008-03-06 16:06 <cbsmith> lol
2008-03-06 16:06 <cbsmith> Except when it comes to performance/efficiency. :-)
2008-03-06 16:06 <flipz> remains to be seen
2008-03-06 16:06 <cbsmith> io_getevents is the actual syscall, but it's the one I've used.
2008-03-06 16:06 <flipz> with all that gack, how can it compete with in-kernel endio?
2008-03-06 16:07 <cbsmith> flipz: well, for starters... it isn't in the kernel... :-)
2008-03-06 16:07 <flipz> how do you poll io_getevents
2008-03-06 16:07 <flipz> this is not obvious to me
2008-03-06 16:07 <flipz> I mean, with sys_poll
2008-03-06 16:08 <flipz> cbsmith, and so this whole braindamaged interface was created just because endio could not be exported?
2008-03-06 16:08 <cbsmith> I've never tried that. AFAIK you don't, because the whole deal with AIO is that it isn't tied to fd's and per-fd offsets.
2008-03-06 16:08 <flipz> that's my point
2008-03-06 16:08 <flipz> pure braindamage
2008-03-06 16:08 <cbsmith> flipz: It has other advantages.
2008-03-06 16:08 <flipz> it's fucked up
2008-03-06 16:09 <flipz> can't poll == can'
2008-03-06 16:09 <cbsmith> Anyway, you can poll by submitting a poll command, but it is way better to just ask the kernel what events have completed.
2008-03-06 16:09 <flipz> can't poll == can't integrate with our polling loop
2008-03-06 16:09 <flipz> how do you poll this by submitting a poll command?
2008-03-06 16:09 <flipz> can we get an rt signal?
2008-03-06 16:10 <cbsmith> we can, but you'd not like how it'd be done. It'd be an ugly hack.
2008-03-06 16:10 <cbsmith> POSIX AIO supports that, and 50% of the reason why I'd not want to use it would be to avoid the hack.
2008-03-06 16:11 <flipz> so how do we integrate with our polling loop? separate thread for aio?
2008-03-06 16:11 <flipz> I can live with that
2008-03-06 16:11 <flipz> it sucks, but then I can fix it too
2008-03-06 16:11 <flipz> this is a job for ddlink
2008-03-06 16:11 <flipz> which is pollable
2008-03-06 16:11 <cbsmith> The whole point of doing AIO would be to avoid the polling paradigm. If we're really married to our polling loop, then we should stick with it.
2008-03-06 16:11 <cbsmith> ..and not do aio.
2008-03-06 16:11 <flipz> in fact I can just toss this whole piece of junk
2008-03-06 16:12 <flipz> cbsmith, it is not just us that is wedded to the polling paradigm
2008-03-06 16:12 <cbsmith> Polling is a comparatively naive interface. It tells you which fd's have had things completed on them, but what you really want (IMHO) is which IO's have completed.
2008-03-06 16:12 <dkegel> And that's a very efficient kind of polling.
2008-03-06 16:12 <flipz> it is considered better practice than the ten thousand thread sickness
2008-03-06 16:12 <flipz> it is
2008-03-06 16:12 <flipz> polling was made in heaven
2008-03-06 16:13 <cbsmith> dkegel: Sure, but it doesn't meld so well with our existing polling loop.
2008-03-06 16:13 <flipz> for lazy people who do not like to think but want fast programs anyway
2008-03-06 16:13 <flipz> in other words, for me
2008-03-06 16:13 <dkegel> flipz, "polling" is usually understood to include methods where you just ask the kernel what events are available.
2008-03-06 16:13 <flipz> ok, we have a plan I think, we will have a separate thread, we will pipe the aio events to our main loop, then when the next round of work on ddlink happens we will do it nicely
2008-03-06 16:13 <flipz> dkegel, makes sense?
2008-03-06 16:14 <dkegel> I don't know yet.  I'd better read our loop first.
2008-03-06 16:14 <flipz> it's a classic poll oriented sever design
2008-03-06 16:14 <flipz> server even
2008-03-06 16:15 <dkegel> Why do you need another thread?  
2008-03-06 16:15 <cbsmith> flipz: I think that kills most of the benefits of using libaio (I guess it would still allow us to have multiple IO's in flight to the same fd), but we'd be making two trips to the kernel to find out that an IO completed.
2008-03-06 16:15 <cbsmith> dkegel: To funnel things in to a pipe so that he can poll. :-)
2008-03-06 16:15 <flipz> [RFC]: char *reason[200]; /* detailed error message, nonnull means errono is nonzero */
2008-03-06 16:16 <flipz> errno even
2008-03-06 16:16 <dkegel> I'm still not clear why we need another thread.
2008-03-06 16:16 <flipz> cbsmith, really it does not
2008-03-06 16:16 <flipz> dkegel, because aio is not pollable
2008-03-06 16:16 <flipz> stupid
2008-03-06 16:16 <flipz> correct action is to fix that
2008-03-06 16:17 <flipz> but not this week
2008-03-06 16:17 <cbsmith> flipz: It is pollable, just not with sys_poll. :-)
2008-03-06 16:17 <flipz> so for now we simulate a fixed version
2008-03-06 16:17 <dkegel> flipz, aio delivers buckets of events.  That's called polling.
2008-03-06 16:17 <flipz> cbsmith, that is waht "not pollable" means
2008-03-06 16:17 <flipz> ok, let's be clear, when I say "poll" I mean poll()
2008-03-06 16:17 <flipz> or select()
2008-03-06 16:17 <flipz> equivalently
2008-03-06 16:18 <flipz> so the ddlink way of doing this will be exactly as efficient, but pollable, and with way less syscalls
2008-03-06 16:18 <dkegel> I see the problem clearly now.  Lessee, somebody had a fix for that proposed recently...
2008-03-06 16:19 <dkegel> What's the ddlink way of doing this?
2008-03-06 16:19 <flipz> I seem to recall
2008-03-06 16:19 <flipz> ring buffer in kernel, ddlink call to return the entire contents of the ring buffer
2008-03-06 16:19 <flipz> ddlink itself is pollable
2008-03-06 16:19 <dkegel> http://lwn.net/Articles/216155/
2008-03-06 16:19 <flipz> basically just export kernel endio
2008-03-06 16:20 <flipz> suparna is one lone voice of reason in the wilderness
2008-03-06 16:20 <flipz> mostly her accurate analysis is ignored
2008-03-06 16:21 <cbsmith> dkegel: yeah, we could use IO_CMD_EPOLL_WAIT
2008-03-06 16:21 <flipz> epoll...
2008-03-06 16:21 <flipz> whole new set of problems
2008-03-06 16:22 <flipz> anyway, if nobody flames my [RFC] above, I'm just going with it ;)
2008-03-06 16:22 <dkegel> http://marc.info/?l=linux-kernel&m=99689413517263&w=2
2008-03-06 16:22 <cbsmith> flipz: I am not sure I agreed that it'd be the same number of trips to the kernel. It should be one less.
2008-03-06 16:23 <flipz> ?
2008-03-06 16:23 <cbsmith> With polling, you normally have one trip for the poll, and another to either read()/write().
2008-03-06 16:24 <cbsmith> With aio you do one trip to get the data.
2008-03-06 16:24 <flipz> dkegel, so I will not preach to you as you are the converted
2008-03-06 16:24 <dkegel> I went seropositive around the turn of the millenium
2008-03-06 16:24 <flipz> cbsmith, see "return the entire contents of the ring buffer" above
2008-03-06 16:25 <dkegel> but see http://marc.info/?l=linux-aio&m=117116242415980&w=2
2008-03-06 16:25 <flipz> anyway, compared to waiting synchronously, even 5 syscalls per transfer would be a relief
2008-03-06 16:25 <cbsmith> flipz: Sorry, I thought you plan was to poll the ddlink fd and then read the contents. If you mean to just do blocking reads then we are in violent agreement.
2008-03-06 16:26 <flipz> in fact we have all this wanking about saving nanoseconds, and no usable aio
2008-03-06 16:26 <flipz> you poll the ddlink, and read a whole batch of completions at once
2008-03-06 16:26 <dkegel> I loves that kind of interface
2008-03-06 16:27 <flipz> so then if there is only one completion it is two calls but you don't care because there is no heavy traffic
2008-03-06 16:27 <dkegel> But it's not the "ddlink" way of doing things.  It's also the epoll way of doing things, and probably several other ways.
2008-03-06 16:27 <cbsmith> flipz: Yeah, the poll() is then an extra syscall compared to straight aio. :-)
2008-03-06 16:27 <flipz> if there are lots of completions, the interarrival time will be measured in nanoseconds
2008-03-06 16:27 <flipz> cbsmith, correct but not relevant to efficiency
2008-03-06 16:27 <cbsmith> But anyway, I am fine with however you want to get the completion events.
2008-03-06 16:27 <flipz> it is not a reason to adopt a horrible api
2008-03-06 16:28 <flipz> so the proposal is, to simulate the right way of doing things for the immediate future
2008-03-06 16:28 <flipz> then to implement the right thing in due course
2008-03-06 16:28 <flipz> which will have minimal effect on ddsnapd in the changeover
2008-03-06 16:29 <cbsmith> flipz: I agree with the critique of the AIO effort in general. IMHO they should have gone with the SGI KAIO codebase in the short term and come up with a generalized completion event mechanism, and then migrated POSIX AIO to run on top of that.
2008-03-06 16:29 <flipz> in either of the changeovers actually
2008-03-06 16:29 <flipz> we will do it right with the help of ddlink
2008-03-06 16:29 <dkegel> http://people.redhat.com/mingo/syslet-patches/ looks dormant
2008-03-06 16:29 <dkegel> is ddlink the 2nd coming?
2008-03-06 16:29 <flipz> I hope
2008-03-06 16:29 <flipz> will it is not syslets
2008-03-06 16:29 <flipz> but it is actually open ended
2008-03-06 16:30 <flipz> you could do all kinds of weird things with it
2008-03-06 16:30 <flipz> including something like syslets
2008-03-06 16:30 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 16:30 <flipz> you could implement a little forth interpreter in kernel and send code to it with ddlink, that would be weird :)
2008-03-06 16:31 <flipz> good for at least 3 new root holes
2008-03-06 16:31 <cbsmith> The main reason I thought AIO is a) multiple in flight writes to the same fd and b) completion notification of the actual events, rather than a "we did something with this fd" notification. So long as we preserve those traits, most of the other ugliness isn't likely to hurt too badly.
2008-03-06 16:31 <cbsmith> flipz: Yeah, don't talk to the ACPI guys about ddlink too much... They just might use it. ;-)
2008-03-06 16:32 <flipz> cbsmith, we propose to preserve those traits
2008-03-06 16:32 <flipz> omg
2008-03-06 16:32 <cbsmith> flipz: yes, I agree.
2008-03-06 16:33 <cbsmith> flipz: What I am saying is the other aspects of it don't bother me much one way or the other. Routing over the pipe is fine so long as we've got a reason for doing it.
2008-03-06 16:33 <flipz> ok, I am going to do the ddlink forth interpreter for next april fool's
2008-03-06 16:33 <flipz> in the mean time, got to make char *reason a reality
2008-03-06 16:33 <cbsmith> flipz: Port the ACPI interpreter to use ddlink instead. I swear people will scream in the dark. :-)
2008-03-06 16:33 <cbsmith> literally :-)
2008-03-06 16:34 <flipz> forth gets the same reaction
2008-03-06 16:34 <flipz> but is actually useful
2008-03-06 16:34 <flipz> you can implement acpi in forth, that should do it
2008-03-06 16:34 <dkegel> syslets seem live... http://marc.info/?l=linux-kernel&w=2&r=1&s=syslets&q=b
2008-03-06 16:34 <cbsmith> flipz: That's the problem with forth: it is useful. The other thing is most forth code isn't broken. These are reasons why the ACPI one would be far more terrifying. ;-)
2008-03-06 16:35 <flipz> acpi implemented in forth
2008-03-06 16:35 <dkegel> http://marc.info/?l=linux-kernel&m=119698330331612&w=2
2008-03-06 16:35 <cbsmith> flipz: Yeah, that one wins. :-0
2008-03-06 16:35 <flipz> could only be beaten by perl in kernel
2008-03-06 16:36 <cbsmith> flipz: s/perl/php/
2008-03-06 16:36 <flipz> zab is pretty sharp
2008-03-06 16:36 <cbsmith> flipz: ...and bring back khttpd ;-)
2008-03-06 16:36 <flipz> but syslets will not goin into mainline this year <- anybody take a bet on that?
2008-03-06 16:37 <dkegel> so, the ol' one thread for disk I/O, and one thread for network I/O for us?
2008-03-06 16:37 <flipz> dkegel, sort of
2008-03-06 16:37 <flipz> it's... and one thread to massage aio badness into pollable goodness
2008-03-06 16:38 <flipz> then we do the actual logic in the main polling loop as god intended
2008-03-06 16:38 <flipz> maybe
2008-03-06 16:38 <dkegel> Don't go hating on everybody.  It's not becoming.
2008-03-06 16:38 <dkegel> Zen...
2008-03-06 16:39 <flipz> I don't hate aio at all, just the aio interface
2008-03-06 16:39 <flipz> and ben is my bud
2008-03-06 16:39 <flipz> never mind he made this mess ;)
2008-03-06 16:39 <willn> 
ACTION blinks

2008-03-06 16:39 <flipz> logically justified every step of the way
2008-03-06 16:40 <cbsmith> we could just use aio for everything.... ;-)
2008-03-06 16:40 <dkegel> ... if only it supported networking.
2008-03-06 16:41 <cbsmith> dkegel: You mean like the ancient SGI KAIO patch? ;-)
2008-03-06 16:42 <dkegel> Ah, the lost years...
2008-03-06 16:42 <flipz> cbsmith, that is exactly what we should do
2008-03-06 16:42 <flipz> however we need to add the bits in one at a time
2008-03-06 16:42 <flipz> we need to be able to #ifdef our optimizations on/off at every step
2008-03-06 16:43 <flipz> because losing stability is just not what we need right now
2008-03-06 16:43 <dkegel> BTW aio for disk I/o means having to throttle ourselves...
2008-03-06 16:43 <cbsmith> stability is for the weak.... and those who need their data. :-)
2008-03-06 16:43 <flipz> dkegel, aio does not mean that the kernel will not throttle
2008-03-06 16:44 <flipz> the contrary
2008-03-06 16:44 <flipz> fits the bio.throttle model perfectly
2008-03-06 16:45 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-06 16:48 <flipz> [RFC] error("could not %s %s%s (%i)", action, strlen(reason) ? "because " : "", reason, strerror(errno));
2008-03-06 16:49 <dkegel> It would be messy to issue a bazillion aio operations and have them sitting around taking up memory.  The higher level you do the throttling, the better...
2008-03-06 16:50 <dkegel> In general.  Not talking about anything in particular, just pontificating.
2008-03-06 16:53 <flipz> snprintf(reason, sizeof(reason), "%s length mismatch: expected >= %zu, actual %u", request, reply_len, head.length);
2008-03-06 16:53 <flipz> real live use of above mechanism
2008-03-06 16:53 <flipz> this is falling into place now
2008-03-06 16:54 <flipz> like this:
2008-03-06 16:54 <flipz> 		snprintf(reason, sizeof(reason), "%s length mismatch: expected >= %zu, actual %u", request, reply_len, head.length);
2008-03-06 16:54 <flipz> 		errno = EINVAL
2008-03-06 16:54 <flipz> 		return -1;
2008-03-06 16:54 <flipz> semicolon damage notwithstanding
2008-03-06 16:54 <flipz> can obviously make that one nice function
2008-03-06 16:55 <flipz> but 3 lines is lightyears better than incumbent, so leave that for next time
2008-03-06 16:56 <flipz> we also need an ignore_message()
2008-03-06 16:56 <flipz> that just reads it
2008-03-06 16:56 <flipz> we have something like that, but it's not right
2008-03-06 16:56 <flipz> should just read it and throw away, not malloc and return the result
2008-03-06 17:00 <flipz> static int ignore(fd, len)
2008-03-06 17:00 <flipz> {
2008-03-06 17:00 <flipz> 	char junk[1];
2008-03-06 17:00 <flipz> 	int err, i;
2008-03-06 17:00 <flipz> 	for (i = 0; i < len; i++)
2008-03-06 17:00 <flipz> 		if ((err = read(fd, junk, 1)) == -1)
2008-03-06 17:00 <flipz> 			return err;
2008-03-06 17:00 <flipz> 	return 0;
2008-03-06 17:00 <flipz> }
2008-03-06 17:00 <dkegel> Hey, progress.  "zumastor replicate zumatest vm2 --wait" hangs forever.
2008-03-06 17:00 <flipz> forever is a little too long
2008-03-06 17:01 <dkegel> And no progress is ever made, I think.  This is all in the "didn't wait long enough after volume creation" case.
2008-03-06 17:04 <flipz> big fat bug
2008-03-06 17:15 <shapor> so apparently most people still dont know you can do kerberized nfs v3 in linux
2008-03-06 17:16 <shapor> someone was just asking me about it today, and what needs to be patched for it to work
2008-03-06 17:16 <shapor> there should really be a howto on setting it up
2008-03-06 17:16 <dkegel> Didn't jiaying write one?
2008-03-06 17:17 <shapor> i thought so but i couldnt find it
2008-03-06 17:17 <shapor> also google doesnt find it
2008-03-06 17:22 <dkegel> zumastor/doc/kerberos-howto.html ?
2008-03-06 18:53 <flipz> does google need to index us better?
2008-03-06 19:12 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-06 19:26 <shapor> no, theres just no doc
2008-03-06 21:53 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Fri Mar 7 00:00:02 PST 2008
2008-03-07 00:29 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-07 00:53 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-07 00:58 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-07 00:59 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-07 01:55 <flipz> 
ACTION planning to sleep as long as possible starting now

2008-03-07 01:55 <flipz> got to shake this cold sometime
2008-03-07 01:59 <flipz> google now find 265 links pointing at zumastor, or 885 with similar links not omitted
2008-03-07 01:59 <flipz> big increase from a short time ago
2008-03-07 05:55 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-07 07:38 <willn>  So
2008-03-07 07:38 <willn> Ubuntu does not do a 2.6.22-18 kernel
2008-03-07 07:40 <willn> We've got a choice of 2.6.22.14 or 2.6.24.11
2008-03-07 07:40 <willn> I don't fully remember, but I think drake said that we dont apply to 2.6.24 proper, so are we ok with 2.6.22-14 ?
2008-03-07 08:05 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-07 10:25 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-07 10:52 <flipz> 2.6.24.11 is a future kernel
2008-03-07 10:53 <flipz> willn, we were running on 2.6.24 not too long ago, just need to get the patches into tree
2008-03-07 11:04 <flipz> jiayingz, there?
2008-03-07 11:05 <jiayingz> hi flipz
2008-03-07 11:05 <flipz> started working with the error handler patch yet?
2008-03-07 11:05 <flipz> otherwise I will make some minor changes and post to the list
2008-03-07 11:05 <jiayingz> not yet
2008-03-07 11:06 <flipz> ok, will post in just a few minutes then
2008-03-07 11:06 <jiayingz> i think you took my old error handling patch
2008-03-07 11:06 <flipz> yes
2008-03-07 11:06 <flipz> it's in there
2008-03-07 11:06 <jiayingz> but I posted a new one
2008-03-07 11:06 <flipz> significant changes?
2008-03-07 11:07 <jiayingz> not significant, but it includes more changes
2008-03-07 11:07 <flipz> that's ok I think
2008-03-07 11:08 <flipz> your get_reply still works the same way
2008-03-07 11:08 <flipz> except no code_error
2008-03-07 11:12 <jiayingz> flipz, I just emailed the right error handling patch
2008-03-07 11:13 <jiayingz> maybe we can merge our two patches
2008-03-07 11:13 <flipz> will do
2008-03-07 11:13 <flipz> if I can do it in 5 minutes
2008-03-07 11:15 <jiayingz> may not be that easy
2008-03-07 11:15 <flipz> jiayingz, we made very similar changes, my expect is slightly better factored than your get_reply_header though
2008-03-07 11:15 <flipz> so there is probably little to merge
2008-03-07 11:16 <flipz> more uses of get_reply
2008-03-07 11:16 <flipz> I will merge those
2008-03-07 11:17 <jiayingz> how about i commit my patch first, then you can work on it directly?
2008-03-07 11:17 <jiayingz> the patch has been tested
2008-03-07 11:17 <jiayingz> it should not break any tests
2008-03-07 11:18 <jiayingz> then i can work on snapshot revert separately
2008-03-07 12:04 <flipz> jiayingz, finished the merge
2008-03-07 12:04 <flipz> shapor, I lost my recipe for making svn do function context in diffs
2008-03-07 12:05 <flipz> again
2008-03-07 12:05 <flipz> 
ACTION makes the sign of the beast at svn

2008-03-07 12:06 <flipz> how can I set the "don't be stupid" mode?
2008-03-07 12:06 <shapor> svn diff --diff-cmd diff -x -rup
2008-03-07 12:06 <flipz> thankyou
2008-03-07 12:07 <flipz> was just reading the help output...
2008-03-07 12:07 <flipz> slowly
2008-03-07 12:07 <shapor> echo 'alias svndiff="svn diff --diff-cmd diff -x -rup"' > ~/.bashrc
2008-03-07 12:07 <flipz> capital idea
2008-03-07 12:10 <shapor> er >> that is ;)
2008-03-07 12:16 -!- zumbi(~zumbi@32.132.219.87.dynamic.jazztel.es) has joined #zumastor
2008-03-07 12:19 <flipz> tried to haxor my bashrc
2008-03-07 12:26 <flipz> http://slashdot.org/article.pl?sid=08/03/07/1854247 <- lol
2008-03-07 12:26 <flipz> quisling de icaza gets his comeupance
2008-03-07 13:01 <cbsmith> I'm not sure if that really counts as a comeupance
2008-03-07 13:26 -!- pgquiles(~pgquiles@144.Red-83-38-36.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-07 13:47 <cbsmith> interesting reading: http://lwn.net/Articles/272048/
2008-03-07 15:07 <dkegel> I'm about to do some performance benchmarking on zuma5 and zuma6...
2008-03-07 15:08 <dkegel> gettin' tired of benchmarking vmware machines on a laptop :-)
2008-03-07 15:43 <willn> 5 and 6 are where?
2008-03-07 15:43 <willn> nyc?
2008-03-07 16:02 <dkegel> yep.  I installed the latest zumastor on both, no problem.
2008-03-07 16:02 <dkegel> But changing the partition table was a little rockier... zuma6 is now down :-(
2008-03-07 16:13 <dkegel> Back up, Shapor power cycled it, that's all it needed...
2008-03-07 17:14 <dkegel> So, zumastor replication is faster than rsync... at least if you're replication >=100k small files that haven't changed since last snapshot.
2008-03-07 17:15 <dkegel> replicating
2008-03-07 21:51 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has left #zumastor
2008-03-07 22:18 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Sat Mar 8 00:00:01 PST 2008
2008-03-08 01:10 -!- MaZ1(~MaZe@216-239-45-4.google.com) has joined #zumastor
2008-03-08 01:14 -!- MaZ1(~MaZe@216-239-45-4.google.com) has left #zumastor
2008-03-08 03:26 -!- pgquiles(~pgquiles@144.Red-83-38-36.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-08 08:06 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-08 08:22 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-08 09:22 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-08 12:19 -!- cbsmith(~xman@h-74-1-65-148.lsanca54.covad.net) has joined #zumastor
2008-03-08 17:10 -!- flips(~phlipz@phunq.net) has joined #zumastor
2008-03-08 22:11 -!- flips(~phlipz@phunq.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Sun Mar 9 00:00:01 PST 2008
2008-03-09 00:31 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-09 09:26 -!- charlesn1(~charles@75.84.92.80) has joined #zumastor
2008-03-09 11:23 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-10 01:27 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-10 02:07 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-10 02:09 <erwan_taf> http://xkcd.com/394/ ;o)
2008-03-10 09:38 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-10 11:10 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-10 11:10 <dkegel> I'm adding lvm snapshots to my benchmark script, running on zuma5 and zuma6.:
2008-03-10 11:11 <cbsmith> how is lvm doing?
2008-03-10 11:11 <cbsmith> never mind. I reread that and got it right. ;-)
2008-03-10 11:13 <dkegel> I was considering renaming it LOLBENCH, but that seems to be taken.
2008-03-10 11:13 <cbsmith> lol
2008-03-10 11:14 <cbsmith> LOLZBENCH? ;-)
2008-03-10 11:14 <cbsmith> YUR FILEZ R REPLIC8NG
2008-03-10 11:15 <dkegel> WERE IN YOUR SNAPS TAKING YOUR SHOTS
2008-03-10 11:16 <cbsmith> lol
2008-03-10 11:16 <cbsmith> I could see this going on for some time.
2008-03-10 11:26 <flipz> 
ACTION is confused about what time it is

2008-03-10 11:29 <erwan_taf> 19h25 here ;)
2008-03-10 11:34 <dkegel> Time to get more hot water for my cocoa here
2008-03-10 11:41 <willn> 
ACTION blinks

2008-03-10 12:37 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-10 13:11 <willn> If anyone in SMO wants a Canoli, and didn't get one, come poke me.
2008-03-10 13:12 <dkegel> yum.  My arteries, who are about to die, salute you.
2008-03-10 14:00 <flipz> canoli?
2008-03-10 14:00 <flipz> dkegel, how are we sitting for 0.7?
2008-03-10 14:00 <dkegel> cannoli, I guess.  
2008-03-10 14:01 <flipz> still don't know what it is
2008-03-10 14:01 <dkegel> http://en.wikipedia.org/wiki/Cannoli
2008-03-10 14:02 <flipz> re 0.7: I need to get the follow up post to last week's device mapper stuff written for tuesday this time, then my plate is clear I think
2008-03-10 14:02 <flipz> can reasonably do something about that untested thing ;)
2008-03-10 14:02 <flipz> I might take a run a aio if nobody beats me to it
2008-03-10 14:03 <flipz> there is also the outstanding protocol cleanup... the journal cleanup, still not checked int.... the rest of ben's patches...
2008-03-10 14:04 <dkegel> Issue 26 is the only blocker for 0.7.  While we wait for that, Jiaying is writing howto doc for mulilayer replication.  We might want to test and maybe fix volume resize.
2008-03-10 14:04 <dkegel> flipz, are you going to do revert?
2008-03-10 14:04 <flipz> snapshot retention I guess
2008-03-10 14:04 <flipz> I expect jiaying will beat me to it, but if not then yes
2008-03-10 14:04 <dkegel> Right.
2008-03-10 14:05 <flipz> http://arstechnica.com/news.ars/post/20080101-solid-state-cloud-storage-on-tap-for-a-power-hungry-2008.html <- very worthwhile post on storage future
2008-03-10 14:06 <flipz> we sure do fit into that
2008-03-10 14:06 <flipz> see "commoditization of network storage"
2008-03-10 14:06 <dkegel> amen
2008-03-10 14:18 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-10 14:19 <dkegel> Any new users here?
2008-03-10 14:19 <flipz> like charlesnw?
2008-03-10 14:19 <dkegel> say
2008-03-10 14:19 <flipz> anyway, I do need a little more testing on that bio stacking code posted earlier
2008-03-10 14:20 <dkegel> be still my heart
2008-03-10 14:20 <flipz> I think I ran it a total of three times, on a device mapper linear target
2008-03-10 14:20 <flipz> didn't oops
2008-03-10 14:20 <flipz> good enough for the likes of me ;)
2008-03-10 14:28 <charlesnw> dunno if i am a new user :)
2008-03-10 14:29 <dkegel> Are you new at heart?  :-)
2008-03-10 14:31 <charlesnw> perhaps 
2008-03-10 14:31 <charlesnw> :)
2008-03-10 14:37 <dkegel> Who was that masked man?
2008-03-10 15:06 <pgquiles> dkegel: no idea but he's been hanging here for at least a week
2008-03-10 15:07 <pgquiles> he might be this charles: http://www.thewybles.com/~charles/
2008-03-10 15:08 <dkegel> Ah.  Yes, that fits.
2008-03-10 15:09 <dkegel> I'm impressed that you pegged him so quick...?
2008-03-10 15:11 <pgquiles> dkegel: I just googled "charlesnw" :-D
2008-03-10 15:11 <pgquiles> by the way, from his blog:
2008-03-10 15:11 <pgquiles> Please join me and other Los Angeles A list bloggers in spreading caution about companies such as G00G that grow at such a rapid rate without the necessary controls to sustain that level of growth. With the economy slowing down like it is, G00G is hurting everyone.
2008-03-10 15:11 <pgquiles> making friends :-D
2008-03-10 15:11 <dkegel> http://twitter.com/charlesnw has some good quotes, too.
2008-03-10 15:12 <dkegel> http://upcoming.yahoo.com/user/113874/ too
2008-03-10 15:14 <pgquiles> yikes, he's 24 and married
2008-03-10 15:14 <pgquiles> that's quite unusual, at least here in Spain
2008-03-10 15:16 <pgquiles> btw, I'm talking about zumastor this Thursday in a local LUG meeting
2008-03-10 15:16 <dkegel> cool!  Got a link?
2008-03-10 15:16 <pgquiles> http://www.polinux.upv.es/drupal/node/960
2008-03-10 15:16 <dkegel> Also, are you waiting for the "revert to snapshot" feature?
2008-03-10 15:16 <pgquiles> sadly, very few info and only in Spanish
2008-03-10 15:16 <pgquiles> dkegel: no
2008-03-10 15:17 <dkegel> Send that link to the mailing list!
2008-03-10 15:17 <pgquiles> ok
2008-03-10 15:18 <dkegel> Charles is an interesting character.  (Charles, you have a typo in your "More Google Holes" post...)
2008-03-10 15:19 <dkegel> In particular, he calls himself a "top blogger".  
2008-03-10 15:21 <pgquiles> mmm he has a "quote" about an idea on monetizing the iphone-wifi duo... he might find my wificast idea interesting ( http://www.elpauer.org/index.php?p=216 )
2008-03-10 15:32 <pgquiles> dkegel: what version of rsync are you using for the benchmarks?
2008-03-10 15:32 <dkegel> 2.6.6
2008-03-10 15:32 <pgquiles> dkegel: quite old
2008-03-10 15:32 <dkegel> The boxes are running dapper more or less
2008-03-10 15:33 <dkegel> Good point
2008-03-10 15:33 <pgquiles> 3.0 improves directory recursion
2008-03-10 15:36 <pgquiles> I'm uploading sources of rsync 3.0.0 for Dapper to my PPA
2008-03-10 15:47 <dkegel> I'm building it now
2008-03-10 15:50 <pgquiles> dkegel: i386 package is still queued, amd64 is already available
2008-03-10 15:51 <dkegel> Running my benchmark now
2008-03-10 16:07 <pgquiles> I'll see the results tomorrow, time to sleep here
2008-03-10 16:07 <dkegel> Don't think it helped
2008-03-10 16:07 <dkegel> #### volsize 1000, 100 dirs, 1 kB/file, first replication
2008-03-10 16:07 <dkegel> real    0m40.158s
2008-03-10 16:07 <dkegel> Second replication
2008-03-10 16:08 <dkegel> real    0m10.045s
2008-03-10 16:08 <dkegel> First replication with rsync
2008-03-10 16:08 <dkegel> real    0m15.054s
2008-03-10 16:08 <dkegel> Second replication with rsync
2008-03-10 16:08 <dkegel> real    0m13.702s
2008-03-10 16:08 <dkegel> First replication with lvm snapshot and rsync
2008-03-10 16:08 <dkegel> real    0m20.156s
2008-03-10 16:08 <dkegel> Second replication with lvm snapshot and rsync
2008-03-10 16:08 <dkegel> real    0m22.762s
2008-03-10 16:08 <dkegel> That's not much different 
2008-03-10 16:10 <pgquiles> first replication with rsync is quite faster now
2008-03-10 16:10 <pgquiles> from 40 min to 15 min
2008-03-10 16:11 <dkegel> no... first, that's seconds, and second, you're comparing it to zumastor
2008-03-10 16:11 <dkegel> Bother rsync 2.6 and rsync 3.0 take about 15 seconds to replicate this 100 directory/1000 file per directory test.
2008-03-10 16:11 <pgquiles> ouch, I can't read :-D
2008-03-10 16:13 -!- cbsmith(~xman@adsl-71-133-80-65.dsl.irvnca.pacbell.net) has joined #zumastor
2008-03-10 16:15 <dkegel> We need an ideas page so we can apply for Summer of Code!
2008-03-10 16:15 <dkegel> Anybody have good ideas for student projects?
2008-03-10 16:16 <pgquiles> mmm a Google project applying for Google's SoC? :-D
2008-03-10 16:16 <dkegel> Sure, why not?  :-)
2008-03-10 16:16 <dkegel> How about online resize?
2008-03-10 16:16 <pgquiles> dkegel: nic
2008-03-10 16:16 <pgquiles> e
2008-03-10 16:17 <pgquiles> zumastor with a zfs backend would be an interesting experiment, too :-)
2008-03-10 16:18 <dkegel> go to bed :-)
2008-03-10 16:18 <pgquiles> :-D
2008-03-10 16:18 <pgquiles> supporting windows 'previous versions' in konqueror and nautilus would be interesting, too, but those are more gnome+samba and kde+samba
2008-03-10 16:19 <dkegel> Or +nfs.
2008-03-10 16:19 <dkegel> Or local.
2008-03-10 16:20 <pgquiles> right
2008-03-10 16:27 <pgquiles> I've added the project proposal to KDE's wiki
2008-03-10 16:28 <pgquiles> read you tomorrow
2008-03-10 16:31 <dkegel> Can you send me a link?
2008-03-10 17:46 -!- cbsmith(~xman@64.148.65.14) has joined #zumastor
2008-03-10 18:36 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-10 22:35 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-10 23:05 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-10 23:43 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Tue Mar 11 00:00:01 PDT 2008
2008-03-11 00:19 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-11 00:42 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-11 03:53 <flipz> well that took a lot more work than I had hoped
2008-03-11 03:54 <flipz> http://lkml.org/lkml/2008/3/11/148 [RFC] Stacking bio support
2008-03-11 04:12 <shapor> you and your long posts
2008-03-11 05:07 <flipz> what are you doing up at this hour?
2008-03-11 05:08 <flipz> first comment re the bio stacking post is a friendly one
2008-03-11 06:13 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-11 06:13 <dank> And the second comment asks if you have a public git tree...
2008-03-11 09:57 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-11 11:15 <willn> flipz: Can you see about making our patches apply (and boot) with a 2.6.24 kernel (trying for 2.6.24-11 from Ubuntu)
2008-03-11 11:24 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-11 11:41 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-11 14:18 <willn> flipz: ping
2008-03-11 15:13 <flipz> willn, pong
2008-03-11 15:13 <flipz> willn, will do
2008-03-11 15:16 <willn> flipz: Thanks
2008-03-11 15:16 <willn> Feel free to send trial patches along the way, i'm warming up the proper builder
2008-03-11 15:19 <flipz> kay
2008-03-11 15:44 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-11 16:04 <flipz> http://lkml.org/lkml/2008/3/11/399 <- how to calculate nines of relability
2008-03-11 16:23 <pgquiles> flipz: I proposed something like RamBack about 2 years ago, only I called it "Snapshot System State Management", AKA "Poor man's UPS"
2008-03-11 16:23 <pgquiles> flipz: http://www.elpauer.org/index.php?p=216 (the last one)
2008-03-11 16:23 <pgquiles> http://lkml.org/lkml/2006/5/21/111
2008-03-11 16:23 <flipz> parallel thinking indeed
2008-03-11 16:24 <flipz> I wrote one then proposed it ;)
2008-03-11 16:25 <flipz> well, these are not actually snapshots to backing store, and that is what the current wanking on lkml is about
2008-03-11 16:26 <cbsmith> Yeah, I wanted the ability to treat bits of RAM like BB RAID controller RAM. Similar idea.
2008-03-11 16:26 <flipz> it only becomes snapshots when the line power goes off
2008-03-11 16:28 <pgquiles> the implementation I had thought of for Poor Man's UPS was quite simple: just reuse the current hibernation code, with some optimization to avoid writing unchanged RAM pages. Re-opening sockets, etc was supposed to be the complex part.
2008-03-11 16:30 <cbsmith> I just wanted to reboot without messing with whatever segment of RAM had been set aside for write caching. Seems relatively trivial really.
2008-03-11 16:31 <flipz> re-establishing hardware state is Hard
2008-03-11 16:31 <flipz> anyway, all these hacks are possible
2008-03-11 16:31 <flipz> just a small matter of coding
2008-03-11 16:33 <flipz> willn, did you mean 2.6.24.2-11?
2008-03-11 16:33 <flipz> I might as well go for 2.6.24.3
2008-03-11 16:34 <cbsmith> flipz: If RAM never loses power though it should stay intact right? The only problem you might have is the memory being corrupted by whatever bug crashed the system in the first place.
2008-03-11 16:34 <flipz> don't have bugs in the kernel
2008-03-11 16:34 <flipz> PC's will reboot without hurting ram
2008-03-11 16:35 <flipz> microsort used that back in OS/2 days to switch back from protected mode to real mode
2008-03-11 16:35 <cbsmith> flipz: Well, the "bug" could be in the kernel, the hardware, or in a driver. I might believe no bugs in the first one, but the other two are a source of much consternation in my life. ;-)
2008-03-11 16:35 <flipz> don't have bugs in the kernel, hardware or driver
2008-03-11 16:36 <flipz> translation: get decent hardware and no binary only drivers
2008-03-11 16:36 <flipz> and mature open source drivers
2008-03-11 16:37 <cbsmith> Lars does have one point though: a lot of databases go to a great deal of trouble to ensure transactional integrity. Some database customers even want it. That might be a problem if we break that feature.
2008-03-11 16:37 <flipz> a great point for some other argument
2008-03-11 16:37 <cbsmith> 
ACTION makes a note not to use pchdtv on a 64-bit kernel ;-)

2008-03-11 16:37 <flipz> completely misses the point that some systems depend on a battery for their integrity
2008-03-11 16:37 <cbsmith> flipz: Isn't it the same argument, or am I lacking context?
2008-03-11 16:37 <flipz> such as many raid controllers
2008-03-11 16:37 <flipz> not the same argument at all
2008-03-11 16:38 <flipz> just sounds similar enough to confuse people not paying attention
2008-03-11 16:38 <cbsmith> Well, I wasn't intending to post to the thread.... I just thought we ought to have a way to answer that concern. Certainly some of the people interested in Zumastor have wanted to use it for databases.
2008-03-11 16:39 <flipz> everybody needs to read the rocket post
2008-03-11 16:39 <cbsmith> flipz: Yeah, I read that. :-)
2008-03-11 16:39 <flipz> so if you want to ride a bicycle, so so, but do not expect it to be allowed into the grand prix
2008-03-11 16:40 <cbsmith> Yeah, I think it'd be good if for 2.0 we gave people the option of riding the bicycle.
2008-03-11 16:40 <flipz> they already have it
2008-03-11 16:40 <flipz> see echo 1 >/proc/driver/ramback/name
2008-03-11 16:41 <flipz> gives ramdisk performance plus fully consistent, sucky write performance
2008-03-11 16:41 <flipz> gives ramdisk read performance plus fully consistent, sucky write performance
2008-03-11 16:41 <flipz> just as cox and lmb want
2008-03-11 16:42 <flipz> and stated clearly in the original post, which shows how attentively they read it
2008-03-11 16:42 <flipz> let alone reading the code
2008-03-11 16:42 <flipz> feh
2008-03-11 16:43 <cbsmith> 
ACTION is still digging up the tree to get full context

2008-03-11 16:51 <cbsmith> 
ACTION now has full context

2008-03-11 16:52 <cbsmith> flipz: Okay, I think your approach makes total and obvious sense.
2008-03-11 16:52 <flipz> :)
2008-03-11 16:53 <flipz> cbsmith, don't tell anybody I bribed you to say that
2008-03-11 16:53 <cbsmith> lol
2008-03-11 16:53 <cbsmith> The only thing I want is a version 2.0 that does what I was talking about. ;-)
2008-03-11 16:53 <flipz> post same on lkml and I will deliver the bribe
2008-03-11 16:53 <cbsmith> hehe
2008-03-11 16:54 <flipz> everybody has the right to create version 2.0, magic of open source
2008-03-11 16:54 <flipz> immediate concern is that 1.0 be solid as the proverbial canadian shield
2008-03-11 16:55 <cbsmith> flipz: Yup, completely agree. 1.0 should be solid and *simple* (which is directly related to solid, but worth emphasizing ;-).
2008-03-11 16:58 <cbsmith> flipz: If you deliver a Violin system to my house, I'll post to the LKML. ;-)
2008-03-11 16:59 <flipz> yes
2008-03-11 16:59 <flipz> if anything, removing code + excessive algorithmic decoration is the right direction to go at this point
2008-03-11 17:00 <flipz> cbsmith, your price per post is impressive
2008-03-11 17:00 <flipz> google up, must have been my post
2008-03-11 17:01 <flipz> "Or to quote Captain Hunt from Andromeda:
2008-03-11 17:01 <flipz> Slipstream: it's not the best way to travel faster than light, it's just 
2008-03-11 17:01 <flipz> the only way.[1]
2008-03-11 17:01 <flipz> At least that's what first crossed my mind after reading the above. ;-)" -- private mail
2008-03-11 17:02 <tim_vimm> cbsmith: if I could deliver a violin system to your house, I'd still be a violinist
2008-03-11 17:02 <flipz> not the only positive private mail I got on this, It seems the people who like this feel intimidated by the usual flamers
2008-03-11 17:02 <cbsmith> tim_vimm: lol
2008-03-11 17:03 <cbsmith> Well, maybe I'll post, but I doubt folks will listen to me much anyway. I haven't posted to the LKML in like 2 years.
2008-03-11 17:04 <flipz> cbsmith, posts from strangers traditionally carry more weight than posts from the usual crickets
2008-03-11 17:04 <flipz> lol, tim_
2008-03-11 17:04 <tim_> guess that makes it official, huh?
2008-03-11 17:04 <flipz> only that
2008-03-11 17:04 <tim_> hehe
2008-03-11 17:04 <flipz> scramjet time, want to meet up tim_?
2008-03-11 17:05 <tim_> did you get into the EA event?
2008-03-11 17:05 <cbsmith> flipz: alright. I'll fire something off tonight then. One thing I'm really good about: I have little fear about flamage beyond losing the time it takes to read and respond to it. ;-)
2008-03-11 17:05 <flipz> no
2008-03-11 17:05 <flipz> s'ok
2008-03-11 17:05 <flipz> got work to do
2008-03-11 17:05 <flipz> my buds will tell me about it
2008-03-11 17:05 <tim_> think it starts at 6:30
2008-03-11 17:05 <flipz> cbsmith, would be much appreciated
2008-03-11 17:05 <flipz> meet at goog asap then?
2008-03-11 17:06 <flipz> I presume shapor got his rsvp in
2008-03-11 17:06 <tim_> i'm involved in a long winded phone call at the moment unfortunately
2008-03-11 17:06 <tim_> shapor's status is the same as yours
2008-03-11 17:07 <tim_> must_read_the_postcard nextime  ;-)
2008-03-11 17:07 <tim_> siggraph always has a rsvp for the high profile events
2008-03-11 17:08 <cbsmith> I gotta head out now (this conversation 'cause me to lose track of time), but I'll catch ya tomorrow.
2008-03-11 20:59 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-11 21:58 <flipz> ziffdavis, the old wintel rag has died
2008-03-11 21:58 <flipz> writing on the wall for wintel I hope
2008-03-11 21:59 <shapor> pc magazine, right?
2008-03-11 22:01 <flipz> right
2008-03-11 22:01 <flipz> died just like comdex, totally a microsoft mouthpiece
2008-03-11 22:01 <flipz> who is left?
2008-03-11 22:01 <flipz> best buy...
2008-03-11 22:01 <flipz> dell is quitting the microsoft thug club
2008-03-11 22:01 <flipz> hopefully before they die too
2008-03-11 22:01 <flipz> 
ACTION owns dell stock

2008-03-11 22:03 <flipz> 2.6.24.2 patches apply to 2.6.24.11, no big surprise
2008-03-11 22:03 <flipz> did anybody ever boot with the 2.6.24.2 patches?
2008-03-11 22:03 <shapor> i think drake did
2008-03-11 22:04 <shapor> and it was broken
2008-03-11 22:04 <flipz> what broke?
2008-03-11 22:04 <shapor> dmsetup hanging or crashing maybe, i dont remember exactly
2008-03-11 22:04 <flipz> hmm
2008-03-11 22:05 <flipz> I guess I will eyeball the bio.throttle patch
2008-03-11 22:05 <flipz> that one was merged by hand
2008-03-11 22:05 <flipz> correctly I thought
2008-03-11 22:09 <flipz> yes, looks like drake did the bio.throttle merge correctly
2008-03-11 22:09 <flipz> not bad
2008-03-11 22:53 <flipz> 2.6.23.3 compiles and boots with ddsnap patches
2008-03-11 22:53 <flipz> now lets's see if ddsnap does its thing
2008-03-11 22:53 <flipz> err
2008-03-11 22:53 <flipz> 2.6.24.3
2008-03-11 22:59 <willn> flipz: Reportedly bio throttle causes the kernel to never boot
2008-03-11 22:59 <flipz> booted for me
2008-03-11 23:00 <flipz> I need a problem report if otherwise
2008-03-11 23:00 <willn> I'll double check with drake tommorow. I'm building an "official" ubuntu zumastor kernel 
2008-03-11 23:01 <flipz> ok
2008-03-11 23:02 <flipz> just verified that I did boot a patched kernel ;)
2008-03-11 23:02 <flipz> always worth double checking that
2008-03-11 23:02 <willn> I'm also working on the package builder in openvz, but my head hurts too much to finish it tonight.
2008-03-11 23:02 <flipz> ok, well I am on call for you
2008-03-11 23:02 <flipz> and problem you have is my first priority
2008-03-11 23:02 <willn> Was it stock 2.6.24 with zumastor patches? Or did you include ubuntu patches as well?
2008-03-11 23:02 <flipz> stock 2.6.24.3
2008-03-11 23:03 <flipz> ubuntu patches are unnecessary as far as I know
2008-03-11 23:03 <willn> http://archive.ubuntu.com/ubuntu/pool/main/l/linux/linux_2.6.24-12.20.diff.gz is what i'm applying before zumastor
2008-03-11 23:03 <flipz> try it without of course
2008-03-11 23:04 <willn> I can't. (Official kernel packages)
2008-03-11 23:04 <flipz> I'll take a boo, see what's in there
2008-03-11 23:04 <flipz> if there is a problem, see "problem report" above.  I will check for any oddities in the unbuntu patches that touch fs/bio.c or block/ll_rw_block.c
2008-03-11 23:05 <willn> If I recall, there is some other bio throttling rolled in at the pre-zumastor level. Drake thought perhaps there would be a conflict
2008-03-11 23:05 <flipz> 3 meg of diff against mainline, getting out of control
2008-03-11 23:05 <willn> at any rate, i'll follow up with him in the morning
2008-03-11 23:05 <flipz> not quite out of control, but getting there
2008-03-11 23:05 <flipz> other bio throttling?
2008-03-11 23:06 <flipz> not from us
2008-03-11 23:06 <flipz> well
2008-03-11 23:06 <willn> Yeah, i'm not sure of the specifics. 
2008-03-11 23:06 <flipz> let me check our dm-ddsnap patch
2008-03-11 23:06 <willn> I'll plead the 5th
2008-03-11 23:07 <flipz> I see somebody changed zumastor/scripts/install completely, I hope it is an improvement
2008-03-11 23:07 <flipz> because it used to work perfectly ;)
2008-03-11 23:07 <flipz> well it ran
2008-03-11 23:07 <flipz> seemed to work
2008-03-11 23:07 <flipz> I won't ask questions :)
2008-03-11 23:08 <flipz> whoops, no
2008-03-11 23:08 <flipz> there were several errors
2008-03-11 23:08 <flipz> tar: usr: Cannot stat: No such file or directory
2008-03-11 23:08 <flipz> tar: Error exit delayed from previous errors
2008-03-11 23:08 <flipz> possible regression
2008-03-11 23:08 <willn> what's the purpose of that little thing
2008-03-11 23:09 <flipz> which little thing?
2008-03-11 23:09 <willn> install
2008-03-11 23:09 <flipz> it's actually the base linux way of doing things
2008-03-11 23:09 <flipz> distros load on piles of gack that are foreign to traditional linux people
2008-03-11 23:10 <willn> The lack of any header leads me to ask these questions
2008-03-11 23:10 <flipz> scripts/install or ./install or make install are the usual ones
2008-03-11 23:10 <flipz> the first two typical for script installs
2008-03-11 23:10 <flipz> header?
2008-03-11 23:10 <flipz> you mean explanation?
2008-03-11 23:10 <flipz> traditionally found in the README or INSTALL
2008-03-11 23:10 <willn> Header in the script, a quick line to say what its for, who wrote it, what the license is, anything really
2008-03-11 23:11 <flipz> seems to have done the right thing, just with some bogus error
2008-03-11 23:11 <flipz> ah, that is normally in the README
2008-03-11 23:11 <flipz> the install script might say "read the README"
2008-03-11 23:11 <willn> I would usually expect a line or two in any script
2008-03-11 23:11 <flipz> true
2008-03-11 23:12 <flipz> file a bug or just check it in?
2008-03-11 23:12 <flipz> I much prefer the latter for trivial things
2008-03-11 23:14 <flipz>  zumastor define volume foo
2008-03-11 23:14 <flipz> /bin/zumastor: /lib/zumastor/common: line 116: conditional binary operator expected
2008-03-11 23:14 <flipz> /bin/zumastor: /lib/zumastor/common: line 116: syntax error near unexpected token `=~'
2008-03-11 23:14 <flipz> /bin/zumastor: /lib/zumastor/common: line 116: `        [[ $# = 1 && $1 && ! $1 =~ [^0-9] ]]'
2008-03-11 23:14 <flipz> /bin/zumastor: unable to load common zumastor function
2008-03-11 23:14 <flipz> the install failed
2008-03-11 23:14 <flipz> somebody broke it
2008-03-11 23:21 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-11 23:40 <flipz> woohoo, nano works in my 100 MB uml root :)
2008-03-11 23:40 <flipz> big step up from joe
2008-03-11 23:40 <flipz> 
ACTION deletes joe

2008-03-11 23:40 <flipz> less, however does not work
2008-03-11 23:40 <flipz> how lame is that
2008-03-11 23:52 <flipz> bash --version
2008-03-11 23:52 <flipz> GNU bash, version 2.03.0(1)-release (i386-pc-linux-gnu)
2008-03-11 23:52 <flipz> current is 3.1.something
2008-03-11 23:52 <flipz> let's see if it will run
2008-03-11 23:53 <flipz> 2/3rds of a MB
2008-03-11 23:53 <flipz> my how everybody has forgotten the meaning of tight code
2008-03-11 23:55 <shapor> well its because they've added support for regex's ;)
2008-03-11 23:55 <shapor> GNU bash, version 2.05b.0(1)-release (i386-pc-linux-gnu)
2008-03-11 23:55 <shapor> also no =~
2008-03-11 23:59 <flipz> root@usermode:~# bash --version
2008-03-11 23:59 <flipz> GNU bash, version 3.1.17(1)-release (i486-pc-linux-gnu)
2008-03-11 23:59 <flipz> Copyright (C) 2005 Free Software Foundation, Inc.
2008-03-11 23:59 <flipz> woohoo
2008-03-11 23:59 <flipz> works now
2008-03-11 23:59 <flipz> nobody broke the install
2008-03-11 23:59 <flipz> I love everybody
irc.oftc.net #zumastor log beginning Wed Mar 12 00:00:01 PDT 2008
2008-03-12 00:00 <flipz> ok, enough for tonight
2008-03-12 00:00 <flipz> by tomorrow I will have a zumastor test environment to a hacker's taste
2008-03-12 00:00 <flipz> less networking, which I still do not undertstand at all when uml gets involvedf
2008-03-12 00:01 <flipz> only less proved hard to just copy from my debian sid /bin
2008-03-12 00:01 <flipz> that is pathetic, that nano should work and less not work
2008-03-12 00:01 <flipz> some symbol  problem
2008-03-12 00:02 <flipz> I probably have an old machine around I get grab a less from ;)
2008-03-12 00:03 <flipz> root@usermode:~# zumastor define volume foo
2008-03-12 00:03 <flipz> /bin/zumastor: zumastor not running, start with '/etc/init.d/zumastor start' first
2008-03-12 00:03 <flipz> 
ACTION is pleased

2008-03-12 00:04 <flipz>  /etc/init.d/zumastor start
2008-03-12 00:04 <flipz> /etc/init.d/zumastor: line 24: /lib/lsb/init-functions: No such file or directory
2008-03-12 00:04 <flipz> /etc/init.d/zumastor: unable to load LSB init function
2008-03-12 00:04 <flipz> 
ACTION is less pleased

2008-03-12 00:04 <flipz> so where did init.d/zumastor go?
2008-03-12 00:08 <flipz> actually, less works
2008-03-12 00:08 <flipz> just has a scary link error:
2008-03-12 00:08 <flipz> root@usermode:~# less test4
2008-03-12 00:08 <flipz> less: Symbol `ospeed' has different size in shared object, consider re-linking
2008-03-12 00:08 <flipz> I wonder if I can relink the elf
2008-03-12 00:10 <flipz> oh wow, tons of people hitting this one out of the web
2008-03-12 00:10 <flipz> smells like a fsckup
2008-03-12 08:31 <willn> 
ACTION blinks

2008-03-12 08:53 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-12 09:04 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-12 10:03 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-12 10:20 -!- erwan_taf(~erwan@81.80.43.67) has left #zumastor
2008-03-12 11:45 -!- MaZ1(~MaZe@216-239-45-4.google.com) has joined #zumastor
2008-03-12 11:46 -!- MaZ1(~MaZe@216-239-45-4.google.com) has left #zumastor
2008-03-12 11:48 -!- MaZ1(~MaZe@216-239-45-4.google.com) has joined #zumastor
2008-03-12 13:04 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-12 13:32 <jiayingz> hi cbsmith
2008-03-12 13:38 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor
2008-03-12 14:25 <cbsmith> jiayingz: belated "hi" back. I forgot to look at my IRC client ;-)
2008-03-12 14:27 <jiayingz> cbsmith, np. i already asked you the question in person :)
2008-03-12 15:43 <flipz> hola
2008-03-12 15:43 <cbsmith> flipz: hula
2008-03-12 15:48 <willn> flipz: Can you send out the emails from the meeting?
2008-03-12 16:02 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-12 16:03 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-12 16:38 <flipz> "LinuxWorld 2008 is looking for BOF topics and .org Exhibitors"
2008-03-12 16:38 <flipz> http://lwn.net/Articles/272691/
2008-03-12 16:38 <flipz> what say we show up in force?
2008-03-12 16:51 <cbsmith> might be fun
2008-03-12 17:10 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor
2008-03-12 17:34 <willn> I'll be back east by then :/
2008-03-12 18:06 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor
2008-03-12 21:46 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-12 21:46 <cbsmith> Why is my inbox stuffed with install failures?
2008-03-12 22:09 <willn> The build/test box had a lockfile mistake
2008-03-12 22:16 <cbsmith> nice
2008-03-12 22:32 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-12 22:33 <dank> http://lwn.net/SubscriberLink/273030/b5544a86ac5e6678/
2008-03-12 22:33 <dank> "Perhaps it's time for some wild ideas. And there is no better source for such ideas than Daniel Philips,"
2008-03-12 22:33 <dank> :-)
2008-03-12 22:33 <flipz> wow
2008-03-12 22:34 <flipz> 
ACTION blushes

2008-03-12 22:34 <dank> flipz, you have a 2nd career as a pr man
2008-03-12 22:34 <flipz> picked up on that rocket ship thing I guess
2008-03-12 22:34 <dank> you know how to get into lwn
2008-03-12 22:36 <dank> last two comments seem a bit off - they say this device is not relevant to the violin
2008-03-12 22:38 <cbsmith> Someone owes me 5 bucks ;-)
2008-03-12 22:38 <flipz> already responded to both
2008-03-12 22:38 <flipz> who does?
2008-03-12 22:38 <cbsmith> I forget. ;-)
2008-03-12 22:38 <cbsmith> I bet someone that the ramback thread would make LWN. ;-)
2008-03-12 22:38 <flipz> heh, and they foolishly took you up
2008-03-12 22:38 <flipz> every patch makes lwn
2008-03-12 22:38 <flipz> unless it is trivial
2008-03-12 22:39 <cbsmith> only ~2 a week
2008-03-12 22:40 <dank> "Your editor fears that this work could suffer the same fate as many of Daniel's other patches: they can contain brilliant ideas and great coding but just don't quite survive the encounter with the real, messy world."
2008-03-12 22:40 <dank> Ouch
2008-03-12 22:40 <flipz> no problem
2008-03-12 22:40 <flipz> somebody asked for a dd tree
2008-03-12 22:41 <flipz> they get it
2008-03-12 22:41 <flipz> great writeup
2008-03-12 22:43 <dank> When are you going to publish your git tree?
2008-03-12 22:44 <flipz> I'm playing with git right now
2008-03-12 22:44 <flipz> hopefully before the end of the week
2008-03-12 22:44 <flipz> well
2008-03-12 22:44 <flipz> coming soon
2008-03-12 22:44 <flipz> say, before next week's LWN ;)
2008-03-12 22:47 <cbsmith> flipz: lol
2008-03-12 22:55 <flipz> I think I better host the bit tree on my own server for a few days, and everybody can try pulling from it etc
2008-03-12 22:56 <flipz> get it worked out, then move it onto kernel.org
2008-03-12 22:57 <cbsmith> 
ACTION falls out of chair reading RMS post about bzr vs. git. ;-)

2008-03-12 23:49 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Thu Mar 13 00:00:01 PDT 2008
2008-03-13 00:03 <dank> Hi Maze
2008-03-13 00:14 <flipz> hey, maze back
2008-03-13 00:15 <flipz> ok, I have a base git tree, 479 MB
2008-03-13 00:16 <flipz> now to add some code to it
2008-03-13 00:18 <flipz> http://www.acis.ufl.edu/~ming/dmcache/index.html
2008-03-13 00:19 <dank> nifty
2008-03-13 00:19 <dank> 'nite
2008-03-13 00:33 <MaZe> hey ;-
2008-03-13 00:37 <flipz> git-diff
2008-03-13 00:37 <flipz> diff --git a/CREDITS b/CREDITS
2008-03-13 00:37 <flipz> index da0a56e..60c0edc 100644
2008-03-13 00:37 <flipz> --- a/CREDITS
2008-03-13 00:37 <flipz> +++ b/CREDITS
2008-03-13 00:37 <flipz> @@ -2752,9 +2752,12 @@ D: Xscale PXA architecture
2008-03-13 00:37 <flipz>  D: unified SMC 91C9x/91C11x ethernet driver (smc91x)
2008-03-13 00:37 <flipz>  S: Montreal, Quebec, Canada
2008-03-13 00:37 <flipz> -N: Ken Pizzini
2008-03-13 00:37 <flipz> -E: ken@halcyon.com
2008-03-13 00:37 <flipz> -D: CDROM driver "sonycd535" (Sony CDU-535/531)
2008-03-13 00:37 <flipz> +N: Daniel Phillips
2008-03-13 00:37 <flipz> +E: phillips@phunq.net
2008-03-13 00:37 <flipz> +E: phillips@google.com
2008-03-13 00:38 <flipz> +D: ddlink device driver control interface
2008-03-13 00:38 <flipz> +D: ddsetup device mapper setup interface
2008-03-13 00:38 <flipz> +D: ddsnap cluster block device snapshot
2008-03-13 00:38 <flipz>  N: Stelian Pop
2008-03-13 00:38 <flipz>  E: stelian@popies.net
2008-03-13 00:38 <flipz> ok, first change was to add myself to the CREDITS file ;)
2008-03-13 00:38 <flipz> funny I'm not in it already
2008-03-13 00:38 <flipz> better send in a patch
2008-03-13 00:38 <cbsmith> You're not?
2008-03-13 00:38 <flipz> oddly enough
2008-03-13 00:39 <flipz> do appear in the source code
2008-03-13 00:39 <shapor> why'd you remove Ken Pizzini
2008-03-13 00:39 <flipz> that is, my old email appears
2008-03-13 00:39 <flipz> um
2008-03-13 00:39 <flipz> accident
2008-03-13 00:39 <flipz> let's put him back
2008-03-13 00:39 <cbsmith> shapor: settling an old grudge. ;-)
2008-03-13 00:39 <shapor> thats what i assumed :)
2008-03-13 00:40 <flipz> $ git-diff
2008-03-13 00:40 <flipz> diff --git a/CREDITS b/CREDITS
2008-03-13 00:40 <flipz> index da0a56e..fe249f9 100644
2008-03-13 00:40 <flipz> --- a/CREDITS
2008-03-13 00:40 <flipz> +++ b/CREDITS
2008-03-13 00:40 <flipz> @@ -2756,6 +2756,13 @@ N: Ken Pizzini
2008-03-13 00:40 <flipz>  E: ken@halcyon.com
2008-03-13 00:40 <flipz>  D: CDROM driver "sonycd535" (Sony CDU-535/531)
2008-03-13 00:40 <flipz> +N: Daniel Phillips
2008-03-13 00:40 <flipz> +E: phillips@phunq.net
2008-03-13 00:40 <flipz> +E: phillips@google.com
2008-03-13 00:40 <flipz> +D: ddlink device driver control interface
2008-03-13 00:40 <flipz> +D: ddsetup device mapper setup interface
2008-03-13 00:40 <flipz> +D: ddsnap cluster block device snapshot
2008-03-13 00:40 <flipz> +
2008-03-13 00:40 <flipz>  N: Stelian Pop
2008-03-13 00:40 <flipz>  E: stelian@popies.net
2008-03-13 00:40 <flipz>  P: 1024D/EDBB6147 7B36 0E07 04BC 11DC A7A0  D3F7 7185 9E7A EDBB 6147
2008-03-13 00:40 <flipz> there we go
2008-03-13 00:40 <flipz> git is fast
2008-03-13 00:41 <flipz> whoops, got to add a whitespace line
2008-03-13 00:41 <flipz> hmm, no
2008-03-13 00:41 <flipz> it was irc that removed it
2008-03-13 00:42 <cbsmith> flipz: yes, it's the C++ of dcvs systems ;-)
2008-03-13 00:42 <flipz> so far it did exactly what I expected
2008-03-13 00:42 <flipz> and did it blindingly fast
2008-03-13 00:42 <flipz> course I haven't done much
2008-03-13 00:42 <flipz> but still, it's whacking the tail of svn
2008-03-13 00:48 <cbsmith> flipz: Yeah, that's kind of like saying "C++ is whacking the tail of Cobol!". ;-)
2008-03-13 00:48 <flipz> much like that
2008-03-13 00:49 <cbsmith> flipz: The good thing is that you are making assertions that few are likely to argue with, which is impressive for dcvs (or language) debates. ;-)
2008-03-13 00:49 <flipz> git tutorial doesn't work
2008-03-13 00:49 <flipz> commands have changed
2008-03-13 00:50 <cbsmith> flipz: "C++ is not easy to use, and the compiler I have doesn't match the standard." ;-)
2008-03-13 00:50 <flipz> 
ACTION can't believe that somebody challenged his assertion that a benchmark ran 25 times faster on a ramdisk than read disk

2008-03-13 00:50 <flipz> that's what email filters are for I suppose
2008-03-13 00:51 <cbsmith> flipz: I'm sure it is possible to come up with some kind of contrived benchmark where it is only 20x faster. ;-)
2008-03-13 00:51 <flipz> s/read/real/
2008-03-13 00:57 <flipz> each git command is a sepate binary -> bad taste
2008-03-13 00:57 <flipz> should have learned from busybox
2008-03-13 00:58 <flipz> also leads needing to type a lot of extra hyphens
2008-03-13 02:12 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor
2008-03-13 03:17 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor
2008-03-13 07:21 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Fri Mar 14 00:00:01 PDT 2008
2008-03-14 20:53 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-14 20:53 <shapor> testfire
2008-03-14 20:53 <shapor> ok, irc logs fixed
2008-03-14 20:56 <shapor> hrm, my isp had some network/latency issues yesterday
2008-03-14 20:56 <shapor> i had to kill the irc bot and let it restart
2008-03-14 22:59 <flipz> the aio example from the man page runs
2008-03-14 22:59 <flipz> after some trivial fixes
2008-03-14 22:59 <flipz> uses a deprecated error output style that ld complains about
2008-03-14 23:03 <flipz> the linux aio interface uses callbacks instead of signals for completion
2008-03-14 23:04 <flipz> still not as good as a pollable fd, but not as bad as signals
2008-03-14 23:06 <flipz> the simple example in the man page is 181 lines long
2008-03-14 23:06 <flipz> to cp one file to another
2008-03-14 23:13 <shapor> this would be much easier if ddsnapd was in the kernel
2008-03-14 23:14 <shapor> simplier at least
2008-03-14 23:36 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Sat Mar 15 00:00:01 PDT 2008
2008-03-15 03:26 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-15 10:36 -!- zumbi_(~zumbi@155.132.219.87.dynamic.jazztel.es) has joined #zumastor
2008-03-15 10:45 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-15 11:23 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-15 11:39 <willn> Blarg. Forgot CONFIG_VERSION_SIGNATURE in the zumastor builds, so the whole thing failed overnight
2008-03-15 11:39 <willn> The good news is, that with 400k inodes and 30G of disk, resources are no longer a problem
2008-03-15 11:42 <willn> Uses ~2G and 100k inodes to do the source package re-build
2008-03-15 13:28 -!- cbsmith(~user@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 13:38 -!- cbsmith(~user@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 13:43 -!- cbsmith(~user@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 14:00 -!- cbsmith(~user@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 14:12 -!- cbsmith(~user@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 14:18 <flipz> so it was  inodes... wow, need 400k to build zumastor, who is eating inodes?
2008-03-15 14:19 <flipz> shapor, agreed about ddsnapd being in the kernel
2008-03-15 14:20 <flipz> it is really tempting to just start working on that
2008-03-15 14:20 <flipz> well
2008-03-15 14:20 <flipz> let's try aio
2008-03-15 14:20 <flipz> somebody needs to
2008-03-15 14:21 <flipz> the resulting code should be fairly similar to how it would be in kernel anyway
2008-03-15 14:33 <willn> flipz: not to build zumastor, to build the entire ubuntu 2.6.24 kernel with a zumastor flavor
2008-03-15 14:33 <flipz> right, still, just to build a kernel...
2008-03-15 14:34 <willn> to build the kernel ~6 or 7 times, plus docs, headers, source packages, etc
2008-03-15 14:34 <flipz> I guess the kernel build must be going tempfile crazy
2008-03-15 14:34 <flipz> or .file crazy, more like
2008-03-15 14:34 <willn> only ever 3000 simultaneous files in the instance
2008-03-15 14:44 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-15 15:01 <willn> When are we slating to release 0.7?
2008-03-15 15:40 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-15 16:04 <flipz> willn, I guess Jiaying put in the last patch on friday...
2008-03-15 16:04 <flipz> I need to ack, shapor already did
2008-03-15 16:04 <flipz> wasn't there a bug that needed to be reopened on initial volume zeroing?
2008-03-15 16:05 <flipz> otherwise, we are into release testing monday I think
2008-03-15 17:33 <willn> That bug was fixed ~15 minutes after I brought it up
2008-03-15 17:33 <willn> Drake put in a unittest to confirm
2008-03-15 17:34 <willn> /s/unit/functional /
2008-03-15 18:32 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-15 21:28 <willn> Blah. Failed again, have to check the office tommorow and see why
2008-03-15 22:37 -!- charlesnw(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-15 22:49 -!- charlesn1(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Sun Mar 16 00:00:01 PDT 2008
2008-03-16 03:19 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-16 04:42 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-16 10:52 -!- charlesn1(~charles@cpe-75-84-92-80.socal.res.rr.com) has joined #zumastor
2008-03-16 11:17 <willn> Missing config options... Joy.
2008-03-16 11:30 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-16 12:13 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-16 13:57 <flipz> shapor, around?
2008-03-16 13:57 <flipz> willn, cool re 15 minute bug
2008-03-16 15:46 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-16 15:47 <pgquiles_> willn: ping
2008-03-16 15:54 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-16 15:57 <flipz> ddsetup is now ported to 2.6.24.3
2008-03-16 15:57 <flipz> I guess this will be the target kernel version for the next little while
2008-03-16 15:57 <flipz> lvm3 patches also ported, and posted
2008-03-16 15:58 <flipz> since we do not have a current user of 2.6.24.2, maybe we should forward port those too
2008-03-16 15:59 <flipz> or I could backport everything from 2.6.24.3
2008-03-16 15:59 <flipz> votes...
2008-03-16 16:00 <pgquiles__> flipz: next Ubuntu (which is a Long Term Support release) will carry 2.6.24.12.13
2008-03-16 16:01 <flipz> 2.6.24.12 isn't out yet
2008-03-16 16:01 <flipz> typo?
2008-03-16 16:01 <pgquiles__> mmm something's wrong
2008-03-16 16:01 <pgquiles__> yeah, just noticed
2008-03-16 16:01 <pgquiles__> no, I guess the trailing 12.13 is ubuntu's stuff
2008-03-16 16:01 <pgquiles__> I'm checking now
2008-03-16 16:01 <flipz> should be dash there, not dot
2008-03-16 16:02 <pgquiles__> indeed
2008-03-16 16:02 <pgquiles__> the metapackage shows 2.6.24.12.13 but the actual package says 2.6.24-12
2008-03-16 16:02 <pgquiles__> http://packages.ubuntu.com/hardy/linux-image-2.6.24-12-386
2008-03-16 16:03 <flipz> wow, it would be weird to base off of 2.6.24.nothing
2008-03-16 16:03 <flipz> I wonder what they really mean
2008-03-16 16:03 <pgquiles__> according to the changelog, they merged 2.6.24.3
2008-03-16 16:03 <flipz> lends weight to the idea we should too
2008-03-16 16:03 <pgquiles__> ubuntu's numbering is not obvious sometimes
2008-03-16 16:04 <flipz> willn, if we aren't using 2.6.24.2 yet, I would prefer to delete and replace with 2.6.24.3
2008-03-16 16:04 <flipz> perhaps the 4 section linux kernel numbering breaks their scripts
2008-03-16 16:04 <flipz> :p
2008-03-16 16:05 <pgquiles__> what's debian using? actually, what's the roadmap for next debian stable?
2008-03-16 16:07 <flipz> checking
2008-03-16 16:08 <pgquiles__> 2.6.24.something, too
2008-03-16 16:09 <pgquiles__> 2.6.24.1 according to the changelog
2008-03-16 16:09 <pgquiles__> mmm 2.6.24-1, actually
2008-03-16 16:09 <pgquiles__> I guess if that's 2.6.24.1 :-?
2008-03-16 16:09 <flipz> right, they don't say either
2008-03-16 16:10 <flipz> probably 2.6.24.anything is ok
2008-03-16 16:10 <flipz> api changes within a stable point release tend to get rejected
2008-03-16 16:12 <pgquiles__> I'm uploading a new version of the kernel for gutsy to my Launchpad PPA, btw
2008-03-16 16:12 <pgquiles__> it'd be nice if that were available from zumastor-team's PPA instead of mine
2008-03-16 16:13 <flipz> ppa?
2008-03-16 16:14 <flipz> ...personal package archive
2008-03-16 16:14 <pgquiles__> :-)
2008-03-16 16:15 <flipz> I don't have an opinion on that, willn or drake maybe
2008-03-16 16:36 <pgquiles__> wtf!!! what's debian using? actually, what's the roadmap for next debian stable?
2008-03-16 16:36 <pgquiles__> ooops
2008-03-16 16:36 <pgquiles__> http://video.google.com/videoplay?docid=323909610753051544
2008-03-16 16:36 <pgquiles__> star wars holiday special
2008-03-16 16:36 <pgquiles__> :-D
2008-03-16 16:36 <pgquiles__> chewbacca goes home for christmas to visit his family :-D
2008-03-16 16:36 <pgquiles__> omg
2008-03-16 18:07 <willn> pgquiles__: which kernel version are you building?
2008-03-16 18:15 <pgquiles__> willn: 2.6.22-14.52
2008-03-16 18:16 <willn> ah, i'm working on the 2.6.24 packages from hardy
2008-03-16 18:16 <pgquiles__> willn: ok
2008-03-16 18:16 <pgquiles__> are you making them the ubuntu way?
2008-03-16 18:16 <willn> flipz: Did you say we have patches for 2.6.24.3/.
2008-03-16 18:17 <willn> pgquiles__: Yea. I spent a couple days putting together openvz scripts for clean building
2008-03-16 18:17 <pgquiles__> willn: great
2008-03-16 18:17 <pgquiles__> willn: I use pbuilder (which uses UML, IIRC)
2008-03-16 18:19 <willn> pgquiles__: We've got cowbuilder inside openvz
2008-03-16 18:19 <willn> A little paranoid, but I err on the safe side
2008-03-16 18:22 <pgquiles__> I've been having some trouble lately with pbuilder and the Launchpad PPA
2008-03-16 18:22 <willn> oh?
2008-03-16 18:22 <pgquiles__> for some reason, open-iscsi installs fine in a pbuilder in my machine but it does not in the PPA
2008-03-16 18:23 <willn> Nothing helpful in their build logs?
2008-03-16 18:23 <pgquiles__> that completely fucks my qemu-cvs packages
2008-03-16 18:23 <pgquiles__> not really, I sent an e-mail to the launchpad-users mailing list but nobody answered :-/
2008-03-16 18:23 <pgquiles__> http://thread.gmane.org/gmane.comp.cms.launchpad.user/3314
2008-03-16 18:24 <pgquiles__> oh, libvirt-cvs was it, not qemu-cvs
2008-03-16 18:28 <willn> pgquiles__: which kernel config did you fork the zumastor flavor config from?
2008-03-16 18:29 <pgquiles__> willn: now, that's a good question
2008-03-16 18:29 <pgquiles__> yesterday I was wondering that
2008-03-16 18:29 <willn> Each time I try one it fails with unset options
2008-03-16 18:29 <pgquiles__> I believed I had forked from -server but turns not I did not
2008-03-16 18:29 <pgquiles__> ... turns out I did not ...
2008-03-16 18:30 <pgquiles__> and I did not fork from lpia, ume, rt, xen or the config in zumastor svn, so heck, I don't know!
2008-03-16 18:30 <pgquiles__> -server should be it, IMO
2008-03-16 18:33 <willn> Still prompting. Le sigh.
2008-03-16 18:33 <pgquiles__> what's prompting?
2008-03-16 18:33 <willn> Config opts
2008-03-16 18:33 <willn> make silentoldconfig is not so silent
2008-03-16 18:34 <pgquiles__> yes, I remembed having that problem
2008-03-16 18:35 <pgquiles__> I ended up using diffs from configs and applying them
2008-03-16 18:36 <willn> hmm?
2008-03-16 18:37 <pgquiles__> according to http://zumastor.googlecode.com/svn/trunk/ddsnap/INSTALL you only need to enable 3 options (four if you include debug) in a config to have ddsnap support, so it's easier to take an existing config and apply a diff which enables those options
2008-03-16 18:38 <pgquiles__> the configs in the zumastor svn tree are not "special", "optimized" or anything like that, I was told a few months ago
2008-03-16 18:38 <pgquiles__> so I better stick to ubuntu configs as much as possible and enable ddsnap
2008-03-16 18:41 <willn> Thats what i'm trying to do
2008-03-16 18:41 <willn> I took the -generic i386 config, added the ddsnap as module option
2008-03-16 18:41 <pgquiles__> willn: btw, what versioning are you using for those kernels in the changelog? (for instance, I'm using 2.6.22-14.52~ppa1)
2008-03-16 18:42 <willn> That's what ive got at the momement
2008-03-16 18:42 <willn> ~ppa1 appended to upstream
2008-03-16 18:42 <pgquiles__> ok
2008-03-16 18:43 <pgquiles__> have you already built any ~ppa1 kernel?
2008-03-16 18:46 <willn> Nope. All fail on zumastor flavors, due to this config bullshit
2008-03-16 18:46 <pgquiles__> ok
2008-03-16 18:47 <willn> Next time we need a project that beings with 0 or a, so the build would fail sooner
2008-03-16 18:47 <pgquiles__> I don't understand :-?
2008-03-16 18:49 <willn> When you submit the build, it builds all of the flavors. Since zumastor is z, it builds near the lsat
2008-03-16 18:49 <pgquiles__> you can define the order in which flavours are built in rules/0-common-vars.mk, in all_custom_flavours = zumastor xenzumastor xen rt ume cell lpiacompat lpia
2008-03-16 18:49 <willn> Hmmmm. Looks like the Xen target modifies Kconfig with its patches
2008-03-16 18:49 <pgquiles__> you'll need to change this:
2008-03-16 18:49 <pgquiles__> rules.d/6-binary-custom.mk:     if [ $(target_flavour) == "xen" -o $(target_flavour) == "xenzumastor" ];  then \
2008-03-16 18:50 <pgquiles__> (add -o $(target_flavour) == "xenzumastor")
2008-03-16 18:51 <willn> should be rules.d/$arch.mk, custom_flavors iirc
2008-03-16 18:51 <pgquiles__> willn: both, actually
2008-03-16 18:51 <pgquiles__> common vars, i386 and amd64
2008-03-16 18:51 <pgquiles__> (i386 and amd64 == $arch)
2008-03-16 18:53 <pgquiles__> now, this tip is going to literally save you hours: when you build the kernel, it searches for symbols in the abi/ directory and searches for a directory named as the *last but one* entry in the changelog, so when you build 2.6.24-12.22 it looks for 2.6.24-12.21, and when you build 2.6.24-12.22~ppa1 it will look for 2.6.24-12.22, which does not exist
2008-03-16 18:53 <willn> Hmm. The readme shows only the arch.mk as required edit
2008-03-16 18:53 <willn> Unfortunately, docs on this process seem minimal
2008-03-16 18:54 <pgquiles__> I can't remember why I edit the three .mk, but probably it didn't work with just arch.mk
2008-03-16 18:59 <pgquiles__> for the abi thing, just copy or symlink abi/2.6.24-12.21 to abi/2.6.24-12.22 and you are done. Then, when you add ~ppa2, ~ppa3, etc you either copy/symlink again or remove the former ~ppaN from the changelog (bad practice, but it works :-) )
2008-03-16 18:59 <flipz> willn, ddsnap patches for 2.6.24.2, lvm3 patches for 2.6.24.3
2008-03-16 19:00 <willn> I just nabbed the 2.6.24.2 i386 config from svn, stuck that in the package
2008-03-16 19:01 <willn> no longer a ubuntu compat config, but at least it builds..
2008-03-16 19:01 <pgquiles__> willn: generic + the three kernel options in the INSTALL should build as well :-?
2008-03-16 19:03 <willn> Nope. Still a bunch of unset options
2008-03-16 19:03 <willn> hang on, ill post the dsc, diff
2008-03-16 19:04 <willn> http://compbrain.net/tmp/linux_2.6.24-12.20~ppa1.dsc http://compbrain.net/tmp/linux_2.6.24-12.20~ppa1.diff.gz
2008-03-16 19:05 <willn> You need the original tar.gz too, but you can find that from your favorite local ubuntu mirror
2008-03-16 19:06 <pgquiles__> willn: why are you using 12.20 instead of 12.22?
2008-03-16 19:07 <willn> I just started working on 22
2008-03-16 19:07 <willn> The 20 stuff has been a work in progress
2008-03-16 19:07 <pgquiles__> ok
2008-03-16 19:08 <willn> pgquiles__: If you stay up much longer you won't have to worry about getting up in the morning ;)
2008-03-16 19:08 <pgquiles__> willn: I've got one week of holidays :-)
2008-03-16 19:09 <pgquiles__> it's Falles, here: http://www.valenciavalencia.com/culture-guide/fallas/fallas.htm
2008-03-16 19:09 <willn> Ah, that works
2008-03-16 19:09 <pgquiles__> anyway, I'm a bit busy now, I'm moving
2008-03-16 19:09 <pgquiles__> I've been packaging and throwing stuff to the bin for three days :-/
2008-03-16 19:09 <willn> :X
2008-03-16 19:10 <pgquiles__> willn: your diff does not define the packages in debian/control or debian/control.stub, your packages won't get built
2008-03-16 19:11 <pgquiles__> ( -zumastor, I meant)
2008-03-16 19:16 <willn> hmm, That should have been in there
2008-03-16 19:18 <willn> Its in one of these things...
2008-03-16 19:18 <willn> The package gets built not packaged ;)
2008-03-16 19:20 <pgquiles__> indeed, my bad :-)
2008-03-16 19:21 <pgquiles__> worst part with kernel packages is how long they take to build :-/
2008-03-16 19:21 <willn> Yep
2008-03-16 19:21 <willn> My technique is to schedule the job, and get food :D
2008-03-16 19:21 <pgquiles__> if I find some time by the end of the week, I'd like to try building the kernel with tcc
2008-03-16 19:22 <willn> tcc?
2008-03-16 19:22 <pgquiles__> yeah, I do sort of that: schedule the job, go to bed
2008-03-16 19:22 <pgquiles__> http://fabrice.bellard.free.fr/tcc/
2008-03-16 19:22 <pgquiles__> I don't know if tcc can build a 2.6 kernel
2008-03-16 19:22 <willn> Interesting
2008-03-16 19:23 <willn> I've been saying speed/resources to the wind, just do it safely
2008-03-16 19:23 <flipz> willn, should we go with 2.6.24.2 or 2.6.24.3 in svn?
2008-03-16 19:24 <pgquiles__> and a tcc-built kernel might not be compatible with propietary modules because of ABI (it should be, because there are no mangling issues in plain C but something says to me it will fail...)
2008-03-16 19:24 <flipz> 
ACTION makes a note to be in valencia next march

2008-03-16 19:24 <willn> Ok, kernel build running on zumabuild2
2008-03-16 19:25 <pgquiles__> flipz: I escape from valencia every year by these days, kids act like they are bin laden pupils, fireworks all the time :-D
2008-03-16 19:25 <flipz> :-D
2008-03-16 19:26 <willn> I think i'm going to go walk the beach, find some dinner, then deal with the control file issues
2008-03-16 19:26 <flipz> some spanish fireworks people always came up to Siegen in the Rhuhr gebiet, what a show
2008-03-16 19:26 <flipz> best fireworks ever
2008-03-16 19:27 <pgquiles__> flipz: Chinese are also very good with fireworks
2008-03-16 19:27 <flipz> they are, but spanish do it a little dfferently
2008-03-16 19:28 <pgquiles__> anyways, this is my last Fallas in Valencia, I'm moving with my girlfriend 200 km away from here. In a couple of months, either I'm a happy husband or I'm looking for a job out of Spain :-D
2008-03-16 19:28 <flipz> berliners are pretty good at it too, when you have 1 million of them all together at the brandenburg gate on new year's
2008-03-16 19:28 <flipz> exciting
2008-03-16 19:29 <flipz> pgquiles__, what is your job there?
2008-03-16 19:29 <pgquiles__> flipz: I wonder that every day, too :-)
2008-03-16 19:29 <flipz> _supposed_ to be?
2008-03-16 19:30 <pgquiles__> flipz: I was hired to port military-grade software for ship control from windows to linux because the client is not allowed to run windows (go figure what country could it be...)
2008-03-16 19:30 <flipz> wow
2008-03-16 19:30 <pgquiles__> but I finished the task so early I've been doing essentially what I want everyday since
2008-03-16 19:30 <flipz> nice gig
2008-03-16 19:30 <pgquiles__> the frontier between pleasure and work is so thin I barely know when I'm working and when I'm playing :-)
2008-03-16 19:31 <flipz> real time control stuff for the ships, or more it oriented?
2008-03-16 19:32 <pgquiles__> I formerly worked as a sysadmin (SAN, NAS, etc), that's why I'm setting up zumastor now at the office (and we'll probably sell a solution based on Zumastor to some Navy)
2008-03-16 19:32 <pgquiles__> flipz: our software controls everything but the combat system
2008-03-16 19:32 <pgquiles__> flipz: stability, lights, engines, electricity, etc
2008-03-16 19:32 <flipz> nice
2008-03-16 19:32 <pgquiles__> yeah
2008-03-16 19:32 <flipz> stability is real time control
2008-03-16 19:32 <flipz> controlling weapons is scary
2008-03-16 19:33 <pgquiles__> yes, but that's a crappy DOS program which we don't develop
2008-03-16 19:33 <flipz> I would not want to do that, though I did spent some years controlling things nearly as dangerous
2008-03-16 19:33 <pgquiles__> we are only provided the binary and communicate through text files... that's scary :-)
2008-03-16 19:33 <flipz> indeed
2008-03-16 19:33 <flipz> so if the ship turns upside down  we know why
2008-03-16 19:34 <pgquiles__> two weeks ago a couple of big shots from the shipyard came to visit us because they'd like us to do some weapons stuff too
2008-03-16 19:34 <flipz> if you do, use "nonevil" bullets ;)
2008-03-16 19:34 <pgquiles__> "evil" is not a problem for our clients, believe me :-D
2008-03-16 19:34 <flipz> and use rtai (first choice) or rtlinux (nearly as good)
2008-03-16 19:35 <pgquiles__> vxworks is what they use now for weapons
2008-03-16 19:36 <pgquiles__> they they are indeed thinking of some real time linux
2008-03-16 19:46 <flipz> vxworks is not a bad choice either
2008-03-16 19:46 <flipz> rtai is as good or better I think
2008-03-16 19:46 <flipz> and free
2008-03-16 19:47 <flipz> if a robot is going to be shooting at me, at least make it not be running windows ;)
2008-03-16 19:48 <pgquiles__> flipz: the worst part about navy stuff is people in charge are militars, you know groucho marx' quote: "Military intelligence is a contradiction in terms" :-)
2008-03-16 19:49 <flipz> :)
2008-03-16 19:49 <flipz> good thing for your contract they don't hang on irc
2008-03-16 19:50 <pgquiles__> :-)
2008-03-16 19:51 <pgquiles__> I'd appreaciate they do not, I wouldn't like getting my balls fried :-)
2008-03-16 19:52 <pgquiles__> I don't really know the internals of the software, I just ported it to linux, ported the build system to cmake and implemented some win32 stuff on linux
2008-03-16 19:52 <pgquiles__> hey CIA, kidnapping me won't get you anywhere!!! sorry for disappointing you, but I'm not your guy!!! :-D
2008-03-16 20:04 <flipz> not at war with spain anyway, this week
2008-03-16 20:05 <flipz> ok, time to do some more work on the git tree
2008-03-16 20:05 <flipz> hopefully the dd git tree will be announced on tuesday or so
2008-03-16 20:06 <pgquiles__> nice
2008-03-16 20:06 <pgquiles__> re my work, I've not said anything you couldn't find by yourself just reading newspapers :-)
2008-03-16 20:06 <flipz> woohoo, ps3 outsold xbox again last month
2008-03-16 20:06 <flipz> so did ps2 for that matter,
2008-03-16 20:06 <pgquiles__> :-D
2008-03-16 20:07 <flipz> by a much bigger margin even
2008-03-16 20:07 <pgquiles__> last game I played was Indiana Jones and the Last Crusade
2008-03-16 20:07 <pgquiles__> back in '94, IIRC
2008-03-16 20:07 <flipz> oh, there are more recent once for the ps2
2008-03-16 20:07 <flipz> In think I may pick one up, just see how it plays
2008-03-16 20:07 <flipz> some ps2 games are really good
2008-03-16 20:08 <flipz> superman for the ps2 is surprisingly good
2008-03-16 20:08 <pgquiles__> I'm sure about that but I have a problem with videogames: I feel guilty while playing them, there are so many interesting things to do... man, if I was able, I wouldn't sleep just to try to catch on :-)
2008-03-16 20:09 <flipz> I don't usually play the games all the way through
2008-03-16 20:09 <flipz> only the best of them
2008-03-16 20:09 <flipz> saves a lot of time
2008-03-16 20:10 <pgquiles__> I watch TV shows in English and lie to myself thinking "I'm not wasting time, just learning English" :-D
2008-03-16 20:14 <pgquiles__> done with books, clothes and furniture, only small stuff is left
2008-03-16 20:23 <flipz> oh right, I don't have a tv, that gives me a little time for video games
2008-03-16 20:26 <pgquiles__> I don't either, P2P has been the best TV for some time now
2008-03-16 20:35 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-16 20:37 <willn> kernel still building, thats a plus
2008-03-16 20:39 <flipz> I guess I am old school, but having a blockbuster within walking distance works for me
2008-03-16 20:40 <flipz> picked up two used ps2 games last time I was there, superman and a game where you battle dragons by shooting rockets out of the back of a 4x4, surprisinly immersive
2008-03-16 20:41 <pgquiles__> :-)
2008-03-16 20:41 <tim_> I narrowly missed being in my own blockbuster on saturday
2008-03-16 20:41 <tim_> tornado in Atlanta, if you didn't see it on CNN
2008-03-16 20:41 <flipz> whoops
2008-03-16 20:41 <flipz> been there ;)
2008-03-16 20:41 <flipz> was in the peachtree during a tornado
2008-03-16 20:41 <tim_> we were about 500 meters away
2008-03-16 20:42 <flipz> clouds were going round & round
2008-03-16 20:42 <flipz> close
2008-03-16 20:42 <flipz> I wasn't really worried about the tornado taking out the peachtree
2008-03-16 20:42 <flipz> it took out a nearby trailer park instead
2008-03-16 20:42 <willn> flipz: blockbuster? really?
2008-03-16 20:42 <tim_> the loft building that was hit the hardest was built by my namesake's grandfather in the 1870's
2008-03-16 20:43 <flipz> while we sat in the bar at the top watching everything spin
2008-03-16 20:43 <tim_> tequila does that
2008-03-16 20:43 <flipz> are you back now?
2008-03-16 20:43 <tim_> yup
2008-03-16 20:43 <flipz> mint juleps as I recall
2008-03-16 20:43 <tim_> how southern of you
2008-03-16 20:44 <flipz> and bacardi+coke
2008-03-16 20:44 <tim_> why would you mess up a coke with bacardi?
2008-03-16 20:44 <tim_> ;-)
2008-03-16 20:44 <flipz> :)
2008-03-16 20:44 <flipz> what is the correct rum?
2008-03-16 20:44 <tim_> deserves a good rum
2008-03-16 20:45 <tim_> uh, you know, a good one :-P
2008-03-16 20:45 <flipz> guess that would let out capn morgan
2008-03-16 20:45 <tim_> heh
2008-03-16 20:45 <tim_> i can't remember a good rum
2008-03-16 20:46 <flipz> straight coke then, for the hard core
2008-03-16 20:46 <tim_> damn straight
2008-03-16 20:46 <flipz> you back?
2008-03-16 20:46 <tim_> y
2008-03-16 20:46 <flipz> congrats on staying ahead of the twister
2008-03-16 20:49 <pgquiles__> 
ACTION notices Intel publishes some really interesting books on multicore software development

2008-03-16 21:04 <pgquiles__> I just found the SparcStation5-keyboard I've been searching for since 2005 :-D
2008-03-16 21:12 <flipz> :-)
2008-03-16 21:12 <flipz> I think there's one sitting beside my desk at goog
2008-03-16 21:13 <flipz> got a sparcstation attached to it though
2008-03-16 21:13 <pgquiles__> I have a couple more sparcs so missing one keyboard was not a big issue
2008-03-16 21:30 <willn> I can't find a place in my heart for sparc arch
2008-03-16 21:32 <tim_> here's a zumastor accelerator:  http://www.news.com/8301-10784_3-9894733-7.html?part=rss&subj=news&tag=2547-1_3-0-5
2008-03-16 21:36 <pgquiles__> bedtime, more (real life) packaging tomorrow
2008-03-16 21:37 <flipz> bye
2008-03-16 23:33 <flipz> slight increase from 17 pages of google hits on zumastor to 18 over the weekend
irc.oftc.net #zumastor log beginning Mon Mar 17 00:00:01 PDT 2008
2008-03-17 02:28 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-17 02:32 <flipz> 2008-03-17_08:39:13.79151 [4790] Connection from 127.0.0.1:2263
2008-03-17 02:32 <flipz> 2008-03-17_08:39:13.79156 [4790] Extended attributes (16 bytes) exist <host=localhost>
2008-03-17 02:32 <flipz> 2008-03-17_08:39:13.79191 [4790] Request upload-pack for '/git/ddtree'
2008-03-17 02:32 <flipz> 2008-03-17_08:39:13.79249 fatal: packfile ./objects/pack/pack-d37a0c64e9ce1c8b29ad9c02a39636ca9c609c31.pack cannot be mapped.
2008-03-17 02:32 <flipz> 2008-03-17_08:39:13.79259 [4790] Disconnected (with error)
2008-03-17 02:33 <flipz> git trying to service a clone request
2008-03-17 02:33 <flipz> they tell me "maybe it means out of memory"
2008-03-17 02:33 <flipz> I have 1 gig and the pack file it is trying to upload is 176 MB
2008-03-17 02:33 <flipz> so... bad git
2008-03-17 02:33 <flipz> bad bad
2008-03-17 02:33 <flipz> and our error messages are much better
2008-03-17 02:54 <shapor> weird
2008-03-17 03:05 <flipz> I'm looking at mercurial
2008-03-17 03:05 <flipz> git is violently crapacious in many respects
2008-03-17 03:06 <flipz> but it will have to be git if it is to be a git tree...
2008-03-17 03:06 <flipz> 
ACTION makes the sign of the beast in linus's direction

2008-03-17 03:08 <flipz> I think I might offer a mecurial tree to start
2008-03-17 03:08 <flipz> people will have to go: apt-get install mercurial; hg clone <url>
2008-03-17 03:09 <flipz> but then...
2008-03-17 03:09 <flipz> tracking linus's tree will be a pain
2008-03-17 03:09 <flipz> so no
2008-03-17 04:03 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor
2008-03-17 07:31 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-17 10:29 <willn> 
ACTION fights buids

2008-03-17 11:35 <willn> make[4]: *** No rule to make target `drivers/md/dm-ddsnap.c', needed by `drivers/md/dm-ddsnap.o'.  Stop.
2008-03-17 11:35 <willn> I can't win.
2008-03-17 11:59 <jiayingz> willn, I think you need to run ddsnap/scripts/genpatch.sh that auto generates the dm-ddsnap patch
2008-03-17 12:37 <willn> thats, uh, not going to work
2008-03-17 12:46 <shapor> willn: why not?
2008-03-17 12:50 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-17 12:54 <willn> If that needs to be run every build, its not going to work inside the safe build container
2008-03-17 12:54 <willn> so, I can generate a static patch and have that applied
2008-03-17 13:36 <cbsmith> Looking at the Zumastor code, it looks like the only aio behavior we really need is "submit write" and "write completed/failed". I couldn't find cases for cancel or even "block until X writes complete". Does that sound right?
2008-03-17 14:06 <willn> Ok, 2.6.24 kernels in about an hour
2008-03-17 14:35 -!- dld(~dld@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-17 15:01 <willn> <_< It's drake!
2008-03-17 16:03 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-17 16:04 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has left #zumastor
2008-03-17 19:03 <willn> I'm determined to have a working kernel by the EOB tommorow
2008-03-17 19:03 <willn> I have hardy-only packages
2008-03-17 19:14 <flipz> likewise, a working ddtree here before the cock crows
2008-03-17 19:54 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-17 20:01 <willn> Its quite annoying when the build bails in a different place every time I give it a go
2008-03-17 20:03 <cbsmith> 
ACTION is working on getting the aio library all good.

2008-03-17 20:03 <cbsmith> It actually looks pretty, which is shocking me.
2008-03-17 20:03 <willn> This is why you automate builds, so when they break you don't look back on a lot of wasted prep time
2008-03-17 20:07 <cbsmith> hehe
2008-03-17 20:12 <willn> I don't get it. How is it failing to build mmc/mss stuff when I don't even touch it
2008-03-17 20:13 <willn> /tmp/tmp.lFxsk29272/linux-2.6.24/debian/build/custom-source-zumastor/drivers/mmc/mss/mss_core.c:592: warning: passing argument 1 of 'schedule_work' from incompatible pointer type
2008-03-17 20:20 <flipz> "Shuttle Launches $199 KPC" http://news.digitaltrends.com/news/story/16063/shuttle_launches_199_kpc
2008-03-17 20:51 <willn> I'm thinking I botched the config
2008-03-17 20:55 <willn> It's failed with some new MMC options as far as I can tell
2008-03-17 20:56 <shapor> flipz wow those would be great dedicated servers for hosting companies
2008-03-17 20:57 <flipz> at $200 I guess I need to order one tonight
2008-03-17 20:57 <willn> You cannot buy them yet
2008-03-17 20:57 <flipz> :(
2008-03-17 20:58 <flipz> I can order though
2008-03-17 20:58 <flipz> I think
2008-03-17 20:58 <willn> The first KPCs will be available through Newegg, although Everex’s link to buy the product from the site has not yet gone live at the time of writing.
2008-03-17 20:58 <flipz> daniel@moonbase:~$ time git clone /var/cache/git/ddtree
2008-03-17 20:58 <flipz> remote: Generating pack...
2008-03-17 20:58 <flipz> remote: Done counting 725622 objects.
2008-03-17 20:58 <flipz> remote: Deltifying 725622 objects.
2008-03-17 20:58 <flipz> remote:  100% (725622/725622) done
2008-03-17 20:58 <flipz> Indexing 725622 objects.
2008-03-17 20:58 <flipz> remote: Total 725622, written 725622 (delta 602290), reused 725621 (delta 602289)
2008-03-17 20:58 <flipz>  100% (725622/725622) done
2008-03-17 20:58 <flipz> Resolving 602290 deltas.
2008-03-17 20:58 <flipz>  100% (602290/602290) done
2008-03-17 20:58 <flipz> Checking files out...
2008-03-17 20:58 <flipz>  100% (23787/23787) done
2008-03-17 20:58 <flipz> real    6m4.210s
2008-03-17 20:58 <flipz> user    5m20.956s
2008-03-17 20:58 <shapor> preorder and you get 2gb flash drive and 160gb hdd for free
2008-03-17 20:58 <flipz> sys     0m7.864s
2008-03-17 20:59 <flipz> time to clone a git repo sucks bumm
2008-03-17 20:59 <flipz> but it did it
2008-03-17 20:59 <flipz> just a moment while I expose it
2008-03-17 21:01 <flipz> ok, just try this http://phunq.net/
2008-03-17 21:01 <flipz> see anything?
2008-03-17 21:02 <flipz> shapor, ping
2008-03-17 21:02 <flipz> I am preordering
2008-03-17 21:02 <flipz> right now
2008-03-17 21:02 <flipz> good enough for me
2008-03-17 21:02 <flipz> shapor, what do you see @ phunq.net?
2008-03-17 21:03 <shapor> flipz pong
2008-03-17 21:03 <flipz> phunq.net?
2008-03-17 21:03 <shapor> hanging browser..
2008-03-17 21:04 <shapor> dns is working
2008-03-17 21:04 <flipz> good?
2008-03-17 21:04 <shapor> works now
2008-03-17 21:04 <flipz> :)
2008-03-17 21:04 <shapor> looks like some http git output minus the css
2008-03-17 21:04 <flipz> ok, just have to figure out how to make it phunq.net/ddtree instead of root
2008-03-17 21:05 <flipz> everybody braindamaged, yes
2008-03-17 21:05 <flipz> configuration hell
2008-03-17 21:05 <flipz> daemon hell
2008-03-17 21:05 <flipz> git hell
2008-03-17 21:05 <shapor> git clone git://phunq.net/ddtree is hanging now
2008-03-17 21:05 <flipz> let/s see what the server log says
2008-03-17 21:05 <flipz> try git clone http://phunq.net/ddtree if that does not work
2008-03-17 21:06 <shapor> on [pid 11145] connect(3, {sa_family=AF_INET, sin_port=htons(9418), sin_addr=inet_addr("64.81.85.152")}, 16
2008-03-17 21:06 <shapor> firewall issue?
2008-03-17 21:06 <flipz> git-daemon issue I think
2008-03-17 21:06 <shapor> netstat -nlpt | grep 9418
2008-03-17 21:07 <flipz> handy
2008-03-17 21:07 <flipz> want to come in and poke around?
2008-03-17 21:07 <flipz> well
2008-03-17 21:07 <flipz> I can post it if git web works
2008-03-17 21:07 <shapor> no!
2008-03-17 21:07 <flipz> and just plead ignorance about the rest of it
2008-03-17 21:07 <shapor> thats so lame
2008-03-17 21:07 <flipz> no to come in?
2008-03-17 21:07 <flipz> I am lame
2008-03-17 21:07 <flipz> :p
2008-03-17 21:08 <shapor> no to posting without git:// working!
2008-03-17 21:08 <flipz> ok
2008-03-17 21:08 <flipz> let me set up a shapor accound
2008-03-17 21:08 <flipz> let me set up a shapor account
2008-03-17 21:08 <shapor> might as well make the subject: "i'm lame, please please flame me"
2008-03-17 21:08 <shapor> :P
2008-03-17 21:08 <flipz> why not
2008-03-17 21:08 <flipz> I was thinking, tell me how to do this
2008-03-17 21:08 <flipz> because it is all too weird for me
2008-03-17 21:09 <flipz> 
ACTION thinks fondly of hg

2008-03-17 21:09 <shapor> well the fact it is hanging and not getting ECONNREFUSED makes me think its a networking issue
2008-03-17 21:10 <shapor> i just set up a git-daemon the other day, i can help if you want
2008-03-17 21:12 <shapor> descriptionUnnamed repository; edit this file to name it for gitweb.
2008-03-17 21:12 <shapor> nice :)
2008-03-17 21:46 <flipz> http://phunq.net/ddtree
2008-03-17 21:56 <flipz> moved it to the right place and fixed the description
2008-03-17 22:47 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-17 22:49 <flipz> hi maze
2008-03-17 22:54 <MaZe> hey
2008-03-17 22:55 <flipz> http://phunq.net/ddtree <- new device driver git tree
2008-03-17 22:55 <flipz> main purpose is to hold the code for lvm3
2008-03-17 22:56 <flipz> having a little trouble with the penguin just now ;)
2008-03-17 22:56 <MaZe> I'm continuously fighting my stupid wireless driver
2008-03-17 22:58 <flipz> on what machine?
2008-03-17 23:01 <MaZe> on my laptop
2008-03-17 23:03 <MaZe> basically two things broken with it: the wireless and the video - both are bblobs, the video I can mostly live with, as long as I don't resume the laptop while connected to an external monitor and don't use 3D stuff it mostly works, but the wireless just randomly stop working, and then requires a suspend to ram to basically reset the hardware (either the card itself or more likely the pci bridge which gets hosed with mabort+)
2008-03-17 23:03 <flipz> linux I presume?
2008-03-17 23:04 <flipz> what chipset?
2008-03-17 23:04 <flipz> what wireless driver?
2008-03-17 23:05 <MaZe> it's an atheros 5418
2008-03-17 23:05 <MaZe> not yet supported by ath5k, so using madwifi - binary blob hell
2008-03-17 23:06 <flipz> I thought madwifi went all open
2008-03-17 23:06 <MaZe> tons of people have reported the issue, and they don't know how to fix it...
2008-03-17 23:06 <MaZe> ath5k is open ;-)
2008-03-17 23:06 <flipz> I see
2008-03-17 23:06 <flipz> patience?
2008-03-17 23:06 <MaZe> or use a cable...
2008-03-17 23:07 <flipz> what brand is the laptop just out of interest?
2008-03-17 23:07 <flipz> (and I will remember not to get it)
2008-03-17 23:07 <MaZe> macbook pro
2008-03-17 23:08 <MaZe> mind you, mac os x on it also has issues...
2008-03-17 23:08 <MaZe> mostly leopard still has bugs and has a tendency to lock up
2008-03-17 23:11 <flipz> brave you for linuxing it
2008-03-17 23:11 <flipz> I need to give my linux thinkpad some lovin
2008-03-17 23:11 <flipz> question of time available
2008-03-17 23:24 <flipz> http://phunq.net/ddtree <- marginally presentable now
2008-03-17 23:24 <flipz> time to check in some groovy code
irc.oftc.net #zumastor log beginning Tue Mar 18 00:00:01 PDT 2008
2008-03-18 00:31 <flipz> git-clone http://phunq.net/dd/.git <- try it :-)
2008-03-18 00:35 <flipz> did I ever say that the concept of pack files in git is idiotic?
2008-03-18 00:35 <flipz> if not, I am saying it now
2008-03-18 00:41 <flipz> git is sloooooow at cloning
2008-03-18 00:44 <willn> yar
2008-03-18 00:44 <willn> Builds running again with better config..
2008-03-18 01:00 <flipz> ok, here is the new spelling: git-clone http://phunq.net/git/ddtree
2008-03-18 01:01 <flipz> willn, could you try that?
2008-03-18 01:01 <flipz> apt-get install git && git-clone http://phunq.net/git/ddtree
2008-03-18 01:08 <willn> is that supposed to redirect
2008-03-18 01:08 <willn> to moonbase.com
2008-03-18 01:08 <flipz> no
2008-03-18 01:08 <flipz> bug
2008-03-18 01:08 <flipz> I wonder why it does that
2008-03-18 01:08 <flipz> let's see what's in the httpd.conf
2008-03-18 01:10 <willn> 
ACTION heads to sleep to wake in 6

2008-03-18 01:10 <flipz> see you
2008-03-18 01:10 <flipz> may have fixed it
2008-03-18 01:10 <flipz> not sure
2008-03-18 01:49 -!- erwan_taf(~erwan@LAubervilliers-151-13-63-69.w217-128.abo.wanadoo.fr) has joined #zumastor
2008-03-18 02:34 <flipz> ok, let's try to pull a stable kernel tree into this shiny new dd git tree
2008-03-18 02:40 -!- erwan__taf(~erwan@LAubervilliers-151-13-63-69.w217-128.abo.wanadoo.fr) has joined #zumastor
2008-03-18 02:48 -!- erwan_taf(~erwan@visage.seanodes.com) has joined #zumastor
2008-03-18 03:07 <flipz> git-checkout -b dd v2.6.24 <- I think this actually creates the dd tree, based off linus's tagged 2.6.24
2008-03-18 03:08 <flipz> now to apply some patches to dd
2008-03-18 03:52 <flipz> ok, ddlink and ddsetup are in ddtree, that is enough for tonight (today...)
2008-03-18 03:53 <flipz> tomorrow... the bio stacking patches I think, that should be enough to announce
2008-03-18 03:53 <flipz> make throw in bio.throttle
2008-03-18 03:53 <flipz> s/make/maybe/
2008-03-18 03:54 <flipz> wow, sounds like a dozen fire trucks racing through sleepy santa monica in the wee hours
2008-03-18 03:54 <flipz> what a circus
2008-03-18 04:13 <flipz> http://zumastor.org/soc.html <- we need to add some more here
2008-03-18 04:13 <flipz> now that we are officially summer-of-codable
2008-03-18 05:02 -!- pgquiles(~pgquiles@248.Red-83-35-113.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-18 05:32 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-18 06:05 -!- _TuXeD_(~jlhl@176.pool80-102-170.dynamic.orange.es) has joined #zumastor
2008-03-18 06:05 <_TuXeD_> Hi
2008-03-18 06:05 <_TuXeD_> Hi pgquiles
2008-03-18 06:13 <pgquiles> hi _TuXeD_
2008-03-18 06:13 <_TuXeD_> how are you?
2008-03-18 08:01 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-18 09:10 <willn> hah!
2008-03-18 09:10 <willn> It worked overnight
2008-03-18 09:10 <willn> Now to try the full build.
2008-03-18 09:22 <willn> oh bloody hell
2008-03-18 09:23 <willn> /tmp/tmp.zBrPu14560/linux-2.6.24-12.22~zumappa/kernel/power/snapshot.c: In function 'init_header_complete':
2008-03-18 09:23 <willn> /tmp/tmp.zBrPu14560/linux-2.6.24-12.22~zumappa/kernel/power/snapshot.c:1247: error: expected expression before ';' token
2008-03-18 09:23 <willn> /tmp/tmp.zBrPu14560/linux-2.6.24-12.22~zumappa/kernel/power/snapshot.c: In function 'check_image_kernel':
2008-03-18 10:09 -!- |TP|Grunt(~grunt@c8u.is.edu.ro) has joined #zumastor
2008-03-18 10:22 <pgquiles> willn: how are you doing with the ddsnap-aware kernel?
2008-03-18 11:18 <willn> pgquiles: Heh. I've been able to use the ubuntu package to build a zumastor binary by hand (fakeroot debian/rules custom-binary-zumastor)
2008-03-18 11:18 <willn> but the overall package fails compilation on some swsusp functions
2008-03-18 12:01 <willn> I think it may be working now..
2008-03-18 12:31 <flipz> biggest deficiency of git is, it has no concept of "a patch"
2008-03-18 12:32 <flipz> much less "a patch set"
2008-03-18 12:32 <flipz> so submitters have to keep remaking their patches, over and over again
2008-03-18 12:32 <flipz> git is for maintainers, not submitters
2008-03-18 12:32 <willn> I should use ccache for these builds, since essentially i'm building the same code ~8 times
2008-03-18 12:32 <flipz> yes
2008-03-18 12:35 <willn> Once this build completes, i've got to find a kernel version number to use that wont trigger the bug in the swsusp module
2008-03-18 13:00 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-18 14:09 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-18 14:10 <willn> hmm, was just giong to ask pgquiles a question
2008-03-18 15:14 <flipz> they do sleep in spain from time to time
2008-03-18 15:14 <flipz> ...or maybe the cia...
2008-03-18 15:25 <flipz> hmm.
2008-03-18 15:25 <flipz> Software error:
2008-03-18 15:25 <flipz> /var/cache/git/ddtree/include/asm-um/arch is encountered a second time at /usr/share/perl/5.8/File/Find.pm line 540.
2008-03-18 15:25 <flipz> For help, please send mail to the webmaster (webmaster@localhost), giving this error message and the time and date of the error. 
2008-03-18 15:25 <flipz> shapor, got any ideas?
2008-03-18 15:26 <flipz> also I guess I should do something about that webmaster@localhost, funny as it may be
2008-03-18 15:27 <flipz> fixed
2008-03-18 15:27 <flipz> and I will filter webmaster to dev/null ;)
2008-03-18 15:32 <flipz> answer: gitweb does not know how to serve a tree that has been built.  Stupid stupid gitweb.
2008-03-18 15:37 <flipz> http://phunq.net/ddtree <- relatively presentable now
2008-03-18 16:09 <flipz> why does this produce surprising results? echo c b a | sort
2008-03-18 16:09 <flipz> where c b and a are not actually files
2008-03-18 16:13 <flipz> this works: echo c b a | xargs -n1 echo | sort
2008-03-18 16:13 <flipz> is this braindamage or what?
2008-03-18 16:54 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-18 16:54 <pgquiles> is anything wrong with zumastor.org? it's not loading for me :-?
2008-03-18 16:58 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-18 17:00 -!- pgquiles(~pgquiles@30.Red-81-39-235.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-18 19:20 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-18 19:26 -!- shapor(~shapor@216.75.12.152) has joined #zumastor
2008-03-18 19:44 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-18 20:18 <flipz> great, phunq.net is back
2008-03-18 20:18 <flipz> can release ddtree now
2008-03-18 20:44 <flipz> shapor, wb
2008-03-18 21:02 <shapor> flipz pong
2008-03-18 21:03 <flipz> I see you got your server back
2008-03-18 21:03 <shapor> my provider, cari.net, had their whole network offlined
2008-03-18 21:03 <shapor> http://www.cari.net/network-status.html?netstat_month=3&netstat_year=2008&netstat_day=false&netstat_start=0
2008-03-18 21:04 <shapor> zumastor.org was also down
2008-03-18 21:04 <shapor> few hours
2008-03-18 21:04 <shapor> i should move the secondary dns elsewhere
2008-03-18 21:04 <shapor> single point of failure
2008-03-18 21:04 <flipz> 
ACTION points shapor in the general direction of query chat =>

2008-03-18 21:09 <flipz> shapor, what is the url for zumastor.org stats again?
2008-03-18 21:12 <shapor> wow pgquiles blog post really got a lot of hits
2008-03-18 21:13 <shapor> 8% of our traffic this month from espana
2008-03-18 21:14 <flipz> :)
2008-03-18 21:14 <flipz> gratias
2008-03-18 21:14 <shapor> lots of hits yesterday from the summer of code sitre
2008-03-18 21:14 <shapor> site*
2008-03-18 21:14 <flipz> let's check the googlebot index
2008-03-18 21:14 <shapor> http://code.google.com/soc/2008/zumastor/about.html
2008-03-18 21:15 <shapor> ?
2008-03-18 21:16 <flipz> 15 pages of google search results, down one from a couple days ago
2008-03-18 21:16 <flipz> interesting
2008-03-18 21:16 <flipz> well, given that lkml has been down all day and is still down, I guess my ddtree announcement just got not so urgent
2008-03-18 21:17 <flipz> I will consider putting in some PS3 time now
2008-03-18 21:22 <flipz> "The purpose of this new kernel git tree is to help break out of the rut that has been Linux storage for the last several years, to provide a congenial forum for development of forward looking storage features such as replication and clustering, and to provide kernels to people who consider it important that storage servers should run reliably under heavy load without deadlocking." <= how does it sound?
2008-03-18 21:24 <flipz> 
ACTION deletes the "rut" clause

2008-03-18 21:52 <flipz> ok, ddtree post is all written and just waiting for lkml to come back from the dead
2008-03-18 22:58 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
irc.oftc.net #zumastor log beginning Wed Mar 19 00:00:01 PDT 2008
2008-03-19 02:30 -!- acesta(~grunt@c8u.is.edu.ro) has joined #zumastor
2008-03-19 03:17 <|TP|Grunt> damn it
2008-03-19 04:40 -!- MaZe(~MaZe@nike.tcs.uj.edu.pl) has joined #zumastor
2008-03-19 05:34 -!- acesta(~grunt@c8u.is.edu.ro) has joined #zumastor
2008-03-19 05:50 -!- erwan_taf(~erwan@visage.seanodes.com) has joined #zumastor
2008-03-19 05:51 -!- MaZe(~MaZe@nike.tcs.uj.edu.pl) has joined #zumastor
2008-03-19 08:33 -!- pgquiles_(~pgquiles@240.Red-81-32-36.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-19 09:17 -!- pgquiles_(~pgquiles@146.red-81-36-194.dynamicip.rima-tde.net) has joined #zumastor
2008-03-19 10:26 <willn> 
ACTION works on the amd64 kernel

2008-03-19 11:06 <flipz> lkml still dead -> no ddtree announcement this morning
2008-03-19 13:59 <flipz> got a response to ddtree announce... nicest possible... almost like we planned it
2008-03-19 14:09 <flipz> ok, improved git workflow coming up... I now have /src/ddtree (hacking tree for me) and /var/cache/git/ddtree (public tree for releasing code)
2008-03-19 14:11 <flipz> so whenever I have something reasonable, I pull into /var/cache/git, thus bypassing various bogus commits... oh wait
2008-03-19 14:11 <flipz> you can't bypass bogus comits in git :-(
2008-03-19 14:11 <flipz> uless you hack git... :-)
2008-03-19 14:26 <flipz> cloning git locally, only 203 MB, takes 7 minutes, no network involved
2008-03-19 14:27 <flipz> :p
2008-03-19 14:27 <flipz> on the other hand, only 203 MB is an impressive repository size considering it is the entire history for most of the life of 2.6
2008-03-19 14:37 <willn> flipz: Did you get a chance to look at the option parsing mess in bin/zumastor
2008-03-19 14:39 <flipz> did not
2008-03-19 14:39 <flipz> why not just clean it up?
2008-03-19 14:39 <flipz> abstract it
2008-03-19 14:39 <flipz> bash style
2008-03-19 14:39 <flipz> I guess I wrote the original messy parser
2008-03-19 14:45 <flipz> daniel@moonbase:/src/ddtree$ svn diff
2008-03-19 14:45 <flipz> svn: '.' is not a working copy
2008-03-19 14:45 <flipz> svn: Can't open file '.svn/entries': No such file or directory
2008-03-19 14:45 <flipz> :)
2008-03-19 14:45 <flipz> svn getting snippy about finding itself in a git repository
2008-03-19 15:03 <flipz> http://thread.gmane.org/gmane.linux.kernel/655209 <- warm n fuzzy re ddtree
2008-03-19 15:08 <flipz> ah, there is a mistake in 2.6.24.2 bio.throttle.patch
2008-03-19 15:08 <flipz> the bio.c chunk is missing
2008-03-19 15:09 <flipz> so .24. won't run very long at all before just sitting there stupidly waiting for a throttle release that will never come
2008-03-19 15:11 <willn> Sounds fun
2008-03-19 15:12 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-19 15:21 <willn> 
ACTION has a working amd64 config

2008-03-19 15:21 <willn> flipz: Can you fix the 2.6.24 patch
2008-03-19 15:21 <flipz> fixing right now
2008-03-19 15:21 <willn> gratzi
2008-03-19 15:21 <flipz> np
2008-03-19 15:24 <flipz> I can see why nobody cares about how badly page cache is broken these days
2008-03-19 15:24 <flipz> git diff is fast, every core dev uses it
2008-03-19 15:24 <flipz> so nobody really notices the cache braindamage
2008-03-19 15:24 <flipz> s/every/almost every/
2008-03-19 15:25 <flipz> and those who don't are running 16 GB dev boxes
2008-03-19 15:31 <flipz> willn, you pinged drake about it?
2008-03-19 15:33 <willn> flipz: The patch? Drake is in Mountain View for the rest of the week
2008-03-19 15:33 <flipz> ah
2008-03-19 15:34 <flipz> ignorance is bliss
2008-03-19 15:34 <flipz> wrong quote
2008-03-19 15:34 <willn> Until it catches up with you
2008-03-19 15:34 <flipz> "what you don't know will not hurt you"
2008-03-19 15:34 <flipz> "if a patch fails in the forest did it really oops?"
2008-03-19 15:36 <flipz> the bio.throttle patch is getting shorted, by using bi_max_vecs as the counting unit instead of a driver-defined metric
2008-03-19 15:36 <flipz> also by removing jiaying's intrusion for the ddsnap proc interface ;)
2008-03-19 15:36 <flipz> and giving that its own patch
2008-03-19 15:37 <willn> I should probably run to the bank before it closes, eh/
2008-03-19 15:38 <flipz> 
ACTION thought only canucks say eh

2008-03-19 15:38 <flipz> oh, but that was eh/
2008-03-19 15:38 <willn> 
ACTION blames the internets

2008-03-19 15:46 <flipz> basic kernel boots and runs
2008-03-19 15:46 <flipz> time to check in the new patch, it can only be less broken than now
2008-03-19 15:51 <willn> 
ACTION blinks back from the bank

2008-03-19 16:11 <flipz> wow, what git does well it does really really really well
2008-03-19 16:11 <flipz> just pulled from one repo on this pc to another for public viewing
2008-03-19 16:33 <flipz> scramjet time
2008-03-19 16:34 <flipz> http://lwn.net/Articles/274079/ <- lkml outage
2008-03-19 18:27 <willn> So, we're building locally
2008-03-19 18:27 <willn> I submitted to launchpad PPA, but they have some quirks i'm dealing with
2008-03-19 20:18 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-19 20:28 -!- phoenix24(~sch@59.178.116.70) has joined #zumastor
2008-03-19 20:28 <phoenix24> Hi channel!
2008-03-19 20:29 <phoenix24> I just threw in a some questions at the mailing list, where Dan suggested me to hang out on irc for a prompt response.
2008-03-19 20:30 <phoenix24> !zumalog
2008-03-19 20:34 -!- phoenix24_(~gopi_daii@59.178.116.70) has joined #zumastor
2008-03-19 20:42 <willn> phoenix24: Our IRC logs are at http://zumastor.org/irclogs
2008-03-19 20:43 <phoenix24> thanks!
2008-03-19 20:48 <willn> It'd be swell if launchpad had an XML feed of the PPA status or something like it
2008-03-19 20:50 <phoenix24> hmm..
2008-03-19 20:51 <phoenix24> I have a question on, Metadata B*Tree leaf nodes.
2008-03-19 20:51 <phoenix24> it contains
2008-03-19 20:51 <phoenix24> a list of logical chunk entries and each logical chunk entry then 
2008-03-19 20:51 <phoenix24> containting one or more physical address
2008-03-19 20:51 <phoenix24> Does this list of physical address (referring of one or more snapshots), correspond to the different versions of the same origin chunk block ?
2008-03-19 20:53 <willn> flipz: or shapor can field that one, i'm not sure
2008-03-19 21:01 -!- phoenix24(~sch@59.178.100.197) has joined #zumastor
2008-03-19 21:15 <flipz> phoenix24, the answer is yes
2008-03-19 21:15 <phoenix24> flipz: for both ?
2008-03-19 21:15 <flipz> each physical address corresponds to a different version of the corresponding origin logical chunk
2008-03-19 21:15 <phoenix24> ok
2008-03-19 21:15 <flipz> and the other question?
2008-03-19 21:16 <phoenix24> How does the difference between two successive directory pointers gives the number of distinct exceptions for a given chunk ?
2008-03-19 21:17 <phoenix24> 
ACTION is talking about the leaf-nodes of the Metadata B*Tree.

2008-03-19 21:18 <flipz> just subtract
2008-03-19 21:18 <flipz> they are ordered by logical address
2008-03-19 21:18 <flipz> deep questions
2008-03-19 21:20 <phoenix24> hmm.. because, leafnodes already contain the exceptions(or sharing) data with them.
2008-03-19 21:20 <phoenix24> thus, substraction directly helps.
2008-03-19 21:21 <phoenix24> 
ACTION is a GSoC student, understanding Zumastor design. To decide on a Zumastor Project idea.

2008-03-19 21:27 <phoenix24> flipz: What is "bio" request ?
2008-03-19 21:28 <flipz> which soc project?
2008-03-19 21:28 <flipz> ours?
2008-03-19 21:29 <flipz> bio -> block I/O -> fill out a struct bio including pointers to pages of data and submit go submit_bio -> generic_make_request
2008-03-19 21:41 <phoenix24> flipz: yes, "Remove 64 snapshot limit" / "Improve Support for Volume Resizing"
2008-03-19 21:55 <willn> flipz: http://launchpadlibrarian.net/12757686/buildlog_ubuntu-gutsy-amd64.linux_2.6.24-12.22~zumappa1475_FAILEDTOBUILD.txt.gz
2008-03-19 21:56 <willn> :(
2008-03-19 22:20 <flipz> phoenix24, feel free to post your ideas to the mailing list
2008-03-19 22:20 <flipz> 
ACTION looks there now

2008-03-19 23:02 <flipz> willn, let me check it
2008-03-19 23:10 -!- natalie(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Thu Mar 20 00:00:01 PDT 2008
2008-03-20 00:20 <flipz> willn, whoops, fixed
2008-03-20 00:21 <flipz> the patch from my git repo accidentally picked up some context from the bio allocation changes
2008-03-20 00:21 <flipz> well, we need to start pulling our patch sets from a git repo I think
2008-03-20 00:21 <flipz> but for not I just fixed the context and rebased bio.throttle
2008-03-20 01:04 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-20 02:08 -!- flipz(~phillips@phunq.net) has joined #zumastor
2008-03-20 02:12 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-20 08:30 -!- phoenix24(~phoenix24@59.178.76.1) has joined #zumastor
2008-03-20 09:45 <willn> 
ACTION kicks the builder with a new patchset

2008-03-20 09:50 -!- pgquiles_(~pgquiles@56.Red-81-36-253.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-20 09:53 -!- pgquiles_(~pgquiles@56.Red-81-36-253.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-20 10:12 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-20 10:27 <willn> Looks like build times are ~5 hours for amd64/i386 kernels
2008-03-20 10:32 <flipz> snappy
2008-03-20 10:43 <willn> in other news, my local amd64 2.6.24 build finished, which is a first for us
2008-03-20 10:57 <flipz> in other news, dank and I both got mentions on linuxtoday today
2008-03-20 10:58 <flipz> by coincidence
2008-03-20 10:58 <flipz> willn, big boquets re amd64
2008-03-20 11:00 <shapor> willn: awesome :)
2008-03-20 11:13 <willn> I also just enabled the lpia kernel builds. Now we need to find someone with an OLPC to do testing
2008-03-20 11:20 <willn> We're using 6 PPA build machines at the moment :x
2008-03-20 11:30 <willn> Woops. AMD64 kernel not booting
2008-03-20 11:30 <willn> 'Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)'
2008-03-20 11:32 <willn> I'm thinking lame config issue
2008-03-20 11:34 <willn> ooh, another fun panic
2008-03-20 11:52 <willn> Can't really capture the call trace
2008-03-20 12:07 <flipz> I'll come over and eyeball it
2008-03-20 12:07 <flipz> digicam also works well
2008-03-20 12:07 <willn> It's on the sun box
2008-03-20 12:07 <willn> hangon
2008-03-20 12:09 <willn> http://compbrain.net/tmp/oops.png
2008-03-20 12:09 <flipz> I think I will make a ddsnap branch in ddtree and try automating the backport to 2.6.18 somewhat
2008-03-20 12:10 <flipz> bio_endio, hmm, suggests I am the culprit
2008-03-20 12:10 <flipz> but we shall see
2008-03-20 12:11 <flipz> sure would be nice to have a line number in bio_endio
2008-03-20 12:12 <flipz> can figure out it laboriously by comparing the code dump to asm output
2008-03-20 12:14 <flipz> ah, this could actually be a buggy _caller_ of bio_endio
2008-03-20 12:15 <flipz> I can insert a BUG_ON to determine that
2008-03-20 12:15 <flipz> after lunch
2008-03-20 12:55 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-20 12:55 <dank> And now ./   http://linux.slashdot.org/linux/08/03/20/1841201.shtml
2008-03-20 13:04 <willn> flipz: feel free to ping if you want to poke at the machine
2008-03-20 13:14 -!- phoenix24(~phoenix24@59.178.102.76) has joined #zumastor
2008-03-20 13:20 <dank> the ramback story on /. is up to 58 comments, mostly jokes about Vista or Eclipse's memory consumption....
2008-03-20 13:20 <dank> http://linux.slashdot.org/comments.pl?sid=494546&cid=22810790 could use a reply.
2008-03-20 13:23 <willn> 
ACTION is without a GUI for about an hour

2008-03-20 13:47 -!- nataliep_(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-20 14:32 <flipz> nice
2008-03-20 14:32 <flipz> now has two comments from me
2008-03-20 14:32 <flipz> one correcting the spelling of my name
2008-03-20 15:22 <willn> just as pgquiles_ predicted, abi failure
2008-03-20 15:23 <flipz> which abi?
2008-03-20 15:34 <willn> building abis
2008-03-20 15:34 <willn> see scrollback from 3 days ago
2008-03-20 15:34 <willn> heh
2008-03-20 15:34 <flipz> heh
2008-03-20 15:34 <flipz> cut n paste the scrollback?
2008-03-20 15:35 <shapor> zumastor.org/irclogs :)
2008-03-20 15:35 <flipz> nuther topic... I am really pleased with my git experience so far
2008-03-20 15:36 <flipz> it really takes a lot of the tension out of juggling patches
2008-03-20 15:36 <flipz> mercurial might be even better, that is next week's research project
2008-03-20 15:36 <willn> flipz: I'd love to paste, unfortunately i'm without an X session
2008-03-20 15:36 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-20 15:44 -!- phoenix24(~phoenix24@59.178.126.100) has joined #zumastor
2008-03-20 16:16 <phoenix24> What functionality is encapsulated by ddraid ?
2008-03-20 16:18 <dank> That's not quite part of zumastor at the moment - it provides a distributed raid device, so you can use disks from multiple machines 
2008-03-20 16:19 <phoenix24> I wonder, why does the code base already contain .. debian packaging information ?
2008-03-20 16:19 <phoenix24> ddraid, sounds cool!
2008-03-20 16:20 <phoenix24> are there any design docs for it ?
2008-03-20 16:21 <shapor> there is a paper linked from the zumastor.org main page i believe
2008-03-20 16:21 <shapor> "Cluster RAID"
2008-03-20 16:23 <dank> phoenix, we want it to be easy for people to try out zumastor.  (AFAIR we don't package ddraid... do we?)
2008-03-20 16:24 <shapor> no ddraid was never packaged
2008-03-20 16:24 <shapor> a good SoC project would be getting ddraid up to date
2008-03-20 16:25 <dank> yes
2008-03-20 16:25 <shapor> ported to a recent kernel
2008-03-20 16:25 <shapor> and some testing, perhaps benchmarking
2008-03-20 16:25 <shapor> thats how ddsnap came back to life
2008-03-20 16:25 <shapor> interns :)
2008-03-20 16:26 <shapor> many of the lessons learned with ddsnap can be applied to ddraid
2008-03-20 16:26 <dank> need a better goal than just "revive"
2008-03-20 16:44 <dank> ok, added an issue and an idea on soc page
2008-03-20 17:52 <phoenix24> the "Cluster RAID" project, does it require access to a RAID Cluster to be satisfactorily do the job ?
2008-03-20 18:35 <willn> Hooray, our build completed for lpia
2008-03-20 18:35 <willn> phoenix24: You probably need to be able to access or simulate >1 machine(s) with >1 disk(s)
2008-03-20 18:36 <phoenix24> ah!
2008-03-20 18:36 <willn> Vmware would probably suffice at the early phases
2008-03-20 18:37 <phoenix24> ok
2008-03-20 18:38 <phoenix24> I was thinking over, "removing the 64bit snapshot limit"
2008-03-20 18:39 <phoenix24> and the best approach is to use extents instead of bitmap.
2008-03-20 18:40 <phoenix24> the extent, could be an array/struct of bitmaps
2008-03-20 18:41 <phoenix24> I may be wrong somewhere, but as far as I've browsed the code.. this should solve the problem.
2008-03-20 18:42 <phoenix24> but then, what limits need to be exposed over ... extents ?
2008-03-20 18:42 <phoenix24> and is this approach viable ?
2008-03-20 20:06 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-20 21:29 <willn> kernels still failing with abi issues
2008-03-20 22:11 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-20 23:09 <willn> Has anyone looked at our failing tests?
2008-03-20 23:31 -!- flips(~phillips@phunq.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Fri Mar 21 00:00:01 PDT 2008
2008-03-21 01:26 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-21 01:38 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-21 03:50 <pgquiles_> back here for a few minutes
2008-03-21 03:50 <pgquiles_> willn: still having abi failures?
2008-03-21 08:15 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-21 08:40 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has left #zumastor
2008-03-21 09:16 <willn> pgquiles_: no, looks like everything is built!
2008-03-21 09:23 <willn> flips: same kernel panic on non-sun hardware
2008-03-21 09:23 <willn> (amd64)
2008-03-21 09:23 <willn> I'm going to try i386.
2008-03-21 09:53 <willn> flips: Panic on i386, just a bit different output. I've taken digicam shots of the i386 panic
2008-03-21 10:26 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-21 10:34 <flips> willn, ok, looking into it
2008-03-21 10:37 <willn> Did you get my pm with the photo?
2008-03-21 10:39 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-21 10:48 <flips> willn, spotted the problem
2008-03-21 10:52 <willn> I await your patch.
2008-03-21 11:08 <flips> willn, checked in
2008-03-21 11:14 <willn> 
ACTION starts testing

2008-03-21 11:54 <willn> 
ACTION ugs at quemu's blazing speed

2008-03-21 11:59 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-21 12:11 <willn> Its a race, will the kernel build finish before the ubuntu install
2008-03-21 12:37 <willn> for the record: the kernel build seemes to take half the time as a qemu ubuntu install
2008-03-21 13:09 <pgquiles_> willn: why have you built hardy's kernel for gutsy?
2008-03-21 13:17 <willn> The machine I built the source packages on is gutsy, so the changelog gets set there
2008-03-21 13:17 <willn> Once the kernels are not broken I'm going to resolve all the unclean things
2008-03-21 13:19 <pgquiles_> I don't really understand the first sentence but the second one makes me feel fine :-)
2008-03-21 13:19 <willn> When you do a dch -i, it will set the release to the release of the machine you run dch on
2008-03-21 13:20 <willn> Since the machine I did the build on is running gutsy, the changelog entry ends up reading gutsy instead of the proper hardy
2008-03-21 13:22 <pgquiles_> doesn't -D hardy work there?
2008-03-21 13:23 <willn> I neglected to include it :x, i'll put it in my change queue
2008-03-21 13:23 <pgquiles_> :-)
2008-03-21 13:23 <willn> I'm also going to start work on the 2.6.22 builder
2008-03-21 13:23 <pgquiles_> great
2008-03-21 13:24 <pgquiles_> packages for 2.6.22-14.52 are already built in my ppa, in case you want to take a look
2008-03-21 13:24 <willn> Yep, I've been using your work as an example
2008-03-21 13:25 <pgquiles_> what's the deadline to propose ideas for SoC?
2008-03-21 13:26 <pgquiles_> splitting ddsnap in libddsnap + very simple ddsnap tool could be a project
2008-03-21 13:26 <pgquiles_> as that shouldn't be difficult, maybe implementing zumastor on top of that libddsnap may be part of the same project
2008-03-21 13:35 <willn> ugh, debchange goes interactive because it does not know about hardy
2008-03-21 13:38 <willn> such an ugly hack.
2008-03-21 13:40 <pgquiles_> change line 343 in /usr/bin/dch :-)
2008-03-21 13:41 <pgquiles_> interesting, there's a /usr/bin/checkbashisms in Ubuntu
2008-03-21 13:59 <willn> Hah! The kernel works!
2008-03-21 14:11 <willn> Pushed to ppa for building.
2008-03-21 14:16 <willn> flips: Thats for that patch.
2008-03-21 14:28 <willn> Sometimes I say thanks instead of thats...
2008-03-21 15:34 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-21 17:34 <willn> Looks like our tester is completely broken.
2008-03-21 18:20 <pgquiles_> willn: I know where my config file for 2.6.22 in Gutsy came from
2008-03-21 18:21 <pgquiles_> I undid what splitconfig.pl does
2008-03-21 18:23 <pgquiles_> i. e. I combined debian/config/i386/config, debian/config/i386/config.386 and debian/config/i386/config.server to produce binary-custom.d/zumastor/config.i386
2008-03-21 18:29 <pgquiles_> no, that's wrong: binary-custom.d/zumastor/config.i386 = debian/config/i386/config + debian/config/i386/config.server
2008-03-21 19:01 <willn> oh, hmm
2008-03-21 19:02 <willn> 
ACTION downloads the latest ppa built package for amd64

2008-03-21 19:02 <willn> mirrors.kernel.org is having issues with packages not being there when the index says they are
2008-03-21 19:11 <flips> time to rotate my scramjet wheels
2008-03-21 19:35 <flips> ...mischief managed
2008-03-21 19:35 <flips> I used a klein bottle rotation pattern this time
2008-03-21 19:35 <flips> which ought to be a big help with higher dimensional tricks
2008-03-21 19:40 <willn> flips: So
2008-03-21 19:40 <willn> 2.6.24 boots, and mostly works
2008-03-21 19:40 <flips> except for?
2008-03-21 19:40 <willn> but ddsnap devices don't seem to work
2008-03-21 19:40 <willn> If I do a zumastor define volume with loop devices for volume and snapstore, then mkfs.ext3, the journal write never finishes
2008-03-21 19:41 <flips> hmm
2008-03-21 19:41 <flips> get you get a SysRq-t trace?
2008-03-21 19:41 <willn> I will once I fix a few other issues
2008-03-21 19:41 <flips> Alt-SysRq-t
2008-03-21 19:41 <flips> ok
2008-03-21 19:41 <willn> (like no net)
2008-03-21 19:41 <willn> heh
2008-03-21 19:42 <flips> yeah
2008-03-21 19:48 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-21 19:59 <willn> pgquiles_: Perfect, re: configs
2008-03-21 19:59 <willn> I'll update things accordingly
2008-03-21 23:58 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Sat Mar 22 00:00:01 PDT 2008
2008-03-22 01:44 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-22 01:45 -!- shapor_(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-22 01:48 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-22 01:59 -!- flips(~phillips@phunq.net) has joined #zumastor
2008-03-22 02:31 -!- pgquiles_(~pgquiles@56.Red-81-36-253.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-22 03:19 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-22 04:30 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 05:27 -!- pgquiles_(~pgquiles@212.Red-83-38-37.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-22 08:23 -!- knweiss(~karsten@p3EE03530.dip0.t-ipconnect.de) has joined #zumastor
2008-03-22 09:14 -!- DavidSev(~David@cl-179.lon-02.gb.sixxs.net) has joined #zumastor
2008-03-22 09:16 <DavidSev> Hi.  I was looking at your gsoc ideas page, and was wondering who I should speak to about the "Easy way for users to access snapshots e.g. in Nautilus" one.
2008-03-22 09:19 <pgquiles_> DavidSev: daniel kegel (dank) or daniel philips (flipz) are the most appropiate persons, they are usually here (american time :-) )
2008-03-22 09:19 <pgquiles_> DavidSev: what do you want to do?
2008-03-22 09:19 <pgquiles_> you may also want to send an e-mail to the mailing list
2008-03-22 09:20 <DavidSev> Ok.
2008-03-22 09:20 <DavidSev> Thanks
2008-03-22 09:21 <pgquiles_> I think I also contributed some comments to the issue tracker for that bug
2008-03-22 09:23 <DavidSev> Do you count tcl/tk as "universally deployed"?
2008-03-22 09:27 <pgquiles_> is it installed by default by the most common linux distributions?
2008-03-22 09:27 <DavidSev> Don't think so
2008-03-22 09:31 <pgquiles_> then, imho, you've answered yourself :-)
2008-03-22 09:31 <DavidSev> Yeah, oh well.
2008-03-22 09:31 <pgquiles_> however, ask flipz or dank
2008-03-22 09:31 <pgquiles_> I'm not a zumastor developer, just a user
2008-03-22 09:31 <DavidSev> Ok.
2008-03-22 09:36 -!- mobin(mobin@116.68.121.187) has joined #zumastor
2008-03-22 09:48 -!- DavidSev_(david@lolcat.mercenariesguild.net) has joined #zumastor
2008-03-22 10:36 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has left #zumastor
2008-03-22 10:37 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 10:56 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has left #zumastor
2008-03-22 11:58 <willn> 
ACTION blinks

2008-03-22 12:01 <willn> pgquiles_: End of today the kernel builder should be done. I've changed it around quite a bit so it should be easy to create a flavor for anything
2008-03-22 12:04 <pgquiles_> willn: awesome
2008-03-22 12:04 <pgquiles_> willn: as soon as 2.6.22-14 packages for gutsy are available from the zumastor-team PPA, I will remove them from my PPA
2008-03-22 12:05 <willn> The packages for 2.6.24 boot and all, but -- their not very functional.
2008-03-22 12:05 <willn> (most device drivers are not there and ddsnap seems wedged)
2008-03-22 12:05 <pgquiles_> oh
2008-03-22 12:05 <pgquiles_> what config are you using?
2008-03-22 12:06 <willn> I was using drakes -- i've switched to the packaged configs as you've described
2008-03-22 12:06 <willn> cat config config.foo > config.new
2008-03-22 12:06 <pgquiles_> that should work fine
2008-03-22 12:06 <pgquiles_> at least it has been working fine for me on gutsy for months
2008-03-22 12:12 <pgquiles_> willn: are you changing the signature in the config file?
2008-03-22 12:13 <pgquiles_> (CONFIG_VERSION_SIGNATURE)
2008-03-22 12:14 <pgquiles_> yesterday I noticed I'm not changing it for gutsy (it's "unofficial") but it should be changed to "-zumastor" and "-xenzumastor"
2008-03-22 12:23 -!- mobin(mobin@116.68.121.187) has left #zumastor
2008-03-22 12:32 <flips> ramback article just squeaked past 400 comments
2008-03-22 12:33 <willn> pgquiles_: For the current 2.6.24 packages, yes
2008-03-22 12:37 <flips> http://wam.umd.edu/~gjm/img/screenshot.png <- 1001 tabs
2008-03-22 12:40 <shapor> haha
2008-03-22 13:14 <flips> ls
2008-03-22 13:14 <flips> -  a  a-  a--  a-a  a--a
2008-03-22 13:15 <flips> is ls sorting fubar?
2008-03-22 13:15 <flips> why does a--a sort after a-a, when - is before a?
2008-03-22 13:40 <shapor> mine doens't sort that way
2008-03-22 13:41 <shapor> are you sure you're sorting? or is that just the order they are in the directory
2008-03-22 13:42 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 13:44 <MaZe> LC_ALL=C ls
2008-03-22 13:48 <flips> hmm
2008-03-22 13:48 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 13:48 <flips> sure I'm sorting
2008-03-22 13:49 <flips> ls | sort
2008-03-22 13:49 <flips> -
2008-03-22 13:49 <flips> a
2008-03-22 13:49 <flips> a-
2008-03-22 13:49 <flips> a--
2008-03-22 13:49 <flips> a-a
2008-03-22 13:49 <flips> a--a
2008-03-22 13:49 <flips> a-a-
2008-03-22 13:49 <flips> a-aa
2008-03-22 13:49 <flips> but I get correct results with LC_ALL=C
2008-03-22 13:50 <flips> ls | LC_ALL=C sort
2008-03-22 13:50 <flips> -
2008-03-22 13:50 <flips> a
2008-03-22 13:50 <flips> a-
2008-03-22 13:50 <flips> a--
2008-03-22 13:50 <flips> a--a
2008-03-22 13:50 <flips> a-a
2008-03-22 13:50 <flips> a-a-
2008-03-22 13:50 <flips> a-aa
2008-03-22 13:50 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 13:51 <MaZe> so it's a matter of locale settings - standard stuff
2008-03-22 13:51 <MaZe> it's sorting in local language dictionary order...
2008-03-22 13:51 <MaZe> [which is usually not what you want, but thus have benefits...]
2008-03-22 13:51 <flips> I wonder why it is braindamaged by default
2008-03-22 13:51 <MaZe> [for example LC_ALL=C gives: A B C ... X Y Z Ą Ź Ć Ż Ś Ń Ó Ę Ł   for polish, but LC_ALL=pl gives A Ą B C Ć D ... X Y Z Ź Ż]
2008-03-22 13:52 <MaZe> and the second is the 'desired' order for Polish non-programmers
2008-03-22 13:52 <MaZe> for programmers it's annoying of course ;-)
2008-03-22 13:52 <flips> I can't think of any logic that could make the LC_ALL="" result correct
2008-03-22 13:53 <MaZe> that's because language sort orders are brain-dead
2008-03-22 13:53 <MaZe> you should see the French language rules
2008-03-22 13:53 <flips> buggy even
2008-03-22 13:53 <MaZe> some stuff is forward sorted, some letters are ignored, accents are sorted from the back of the word... etc
2008-03-22 13:54 <flips> I would l think that if no environment variable is set then the default would be ascii lexical order
2008-03-22 13:54 <flips> but this does not appear to be the case
2008-03-22 13:55 <MaZe> my guess the sort order is sorting on alphabetic characters first with any non-alphabetical sequence of characters replaced with a single space, and then some sort of tie resolution is done
2008-03-22 13:55 <MaZe> type in locale
2008-03-22 13:55 <flips> see "braindamage" above
2008-03-22 13:55 <MaZe> LC_ALL is actually a super variable
2008-03-22 13:55 <MaZe> it overrides a bunch of other variables
2008-03-22 13:55 <flips> en_CA
2008-03-22 13:55 <MaZe> I believe LC_COLLATE is the one you actually want
2008-03-22 13:55 <flips> I can state with some authority that that is not how we sort things in _CA
2008-03-22 13:56 <flips> thankyou, that was informative and sad
2008-03-22 13:56 <MaZe> LC_COLLATE=C ls
2008-03-22 13:56 <flips> eys
2008-03-22 13:56 <flips> yes
2008-03-22 13:56 <MaZe> I wouldn't be so sure ;-)
2008-03-22 13:56 <MaZe> I'm pretty sure that the en_CA sort order is indeed not the ASCII one
2008-03-22 13:57 <flips> ascii or not, it is still a bug
2008-03-22 13:57 <flips> committee got involved?
2008-03-22 13:57 <MaZe> since you'd ignore hyphens and apostrophes in a dictionary I believe
2008-03-22 13:57 <MaZe> I believe it's supposed to be user friendly...
2008-03-22 13:58 <MaZe> it sorts correctly for normal words - but fails spectacularly for stuff which isn't a valid word form
2008-03-22 13:58 <flips> foolish
2008-03-22 13:59 <MaZe> and there's other stuff in some languages, like CH being considered a separate letter between C and D, so you have CA..CG,CI..CZ,CH,DA..DZ,etc
2008-03-22 13:59 <flips> haha
2008-03-22 13:59 <flips> no wonder the world winds down over time
2008-03-22 13:59 <MaZe> lol
2008-03-22 14:00 <flips> git adds its own idiocy to the pile
2008-03-22 14:00 <MaZe> hehe
2008-03-22 14:00 <flips> just getting deep enough to feel the full effect of the warts
2008-03-22 14:01 <flips> I guess I will end up using it mainly as a communication vehicle, and go to something reasonable like mercurial for real work
2008-03-22 14:01 <flips> git does not add a file to a commit just because you changed it...
2008-03-22 14:01 <MaZe> haven't used it much, but when I first read an article about it my first thought was wow, this is almost like something I'd been thinking about (never actually started coding anything though...) - but it's come a long long way since then (and it's still an infant...)
2008-03-22 14:02 <flips> regardless of whether it is in the ignore list or not
2008-03-22 14:02 <MaZe> well neither does perforce
2008-03-22 14:02 <flips> that makes it right?
2008-03-22 14:02 <MaZe> no, just pointing that out
2008-03-22 14:02 <flips> so maybe linus should think twice before calling svn developers stupid
2008-03-22 14:02 <MaZe> you just have to have appropriate habits (and the right aliases)
2008-03-22 14:02 <flips> he got a bit o that hizself
2008-03-22 14:11 <flips> git notion of workflow is badly broken
2008-03-22 14:11 <flips> the index is handled all wrong
2008-03-22 14:11 <flips> will keep git relegated to a lkml only thing
2008-03-22 14:11 <flips> too bad about that
2008-03-22 14:12 <flips> nearly great
2008-03-22 14:14 <flips> git-branch creates a new branch but does not switch to it :p
2008-03-22 14:43 -!- charlesn1(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-22 16:40 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-22 17:10 <flips> sigh
2008-03-22 17:11 <flips> a whole day gone to fighting with git and debian sillyness
2008-03-22 17:12 <willn> flips: There is always windows.
2008-03-22 17:12 <flips> eek
2008-03-22 17:28 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-22 18:35 -!- willn-colo(wan@209.67.252.126) has joined #zumastor
2008-03-22 18:38 <willn> willn-colo: rss-watch zumastor-commits
2008-03-22 18:38 <willn-colo> watcher thread started
2008-03-22 18:39 <willn> willn-colo: log-on
2008-03-22 18:39 <willn-colo> logging enabled in #zumastor
2008-03-22 18:39 <willn> flips: Shall I get you that sysrq output?
2008-03-22 19:19 <willn> I would love to test i386, but the disk encryption module wasnt built...
2008-03-22 19:20 <willn> 
ACTION fixes the new builer a bit more and starts a DIY build

2008-03-22 19:38 <willn-colo> zumastor-commits: [zumastor commit] r1489 - in trunk: ddsnap/scripts kernel/config kernel/ubuntu-package
2008-03-22 19:43 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1489 (source)
2008-03-22 19:44 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1489 (source)
2008-03-22 20:30 <willn> flips: uploading..
2008-03-22 20:37 <willn> flips: http://www.ccs.neu.edu/home/wan/mind-sysrq-hang-mkfs.ext3.txt.gz
2008-03-22 20:39 <flips> thanks willn
2008-03-22 20:41 <willn> sorry for the delay
2008-03-22 20:46 <flips> blluetoothd?
2008-03-22 20:46 <flips> well we should start by removing hald I think
2008-03-22 20:47 <willn> this is on my laptop. 
2008-03-22 20:47 <flips> oh :)
2008-03-22 20:47 <flips> don't actually see a smoking gun
2008-03-22 20:47 <willn> i'm testing with what i've got infront of me ;)
2008-03-22 20:47 <flips> it makes zero progress, or just very slow?
2008-03-22 20:48 <willn> its a 200MB loop device
2008-03-22 20:48 <willn> and it doesnt finish in 10 minutes
2008-03-22 20:48 <flips> loop has issues
2008-03-22 20:48 <flips> would be more interesting if it was a real partition
2008-03-22 20:49 <willn> I can try getting that online. Since we use loop devices in our manual
2008-03-22 20:49 <flips> should include a warning "don't expect this to work"
2008-03-22 20:50 <flips> silly number of tasks
2008-03-22 20:50 <willn> Yea
2008-03-22 20:50 <flips> what is the world coming to
2008-03-22 20:50 <willn> its a full on gdm login screen boot
2008-03-22 20:51 <flips> console-kit-d?
2008-03-22 20:52 <flips> seems the task trace was too big for the buffer
2008-03-22 20:52 <willn> http://www.freedesktop.org/wiki/Software/ConsoleKit
2008-03-22 20:53 <willn> I probably need to do the test on a cleaner system, but its the same behavior as i was getting before
2008-03-22 20:53 <willn> (on a cleaner system)
2008-03-22 20:54 <flips> from the inside, console kit looks completely gross
2008-03-22 20:54 <flips> loop mount is not good for this
2008-03-22 20:54 <flips> marginally interesting that it doesn't work
2008-03-22 20:54 <willn> if its not something we can say works, we shouldnt put it in our quick demo ;)
2008-03-22 21:18 <flips> true
2008-03-22 21:20 -!- mobin(mobin@116.68.121.187) has joined #zumastor
2008-03-22 21:29 -!- phoenix24(wzkf@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-22 21:37 <phoenix24> Hi Everyone!
2008-03-22 21:39 <phoenix24> I think the Zumastor codebase, could use a lot of *good programming practices*
2008-03-22 21:40 <phoenix24> is it possible, that patches for code cleanup would be accepted ?
2008-03-22 22:01 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-22 22:04 <shapor> phoenix24 post your ideas for cleanup, perhaps with an example of a small section cleaned up
2008-03-22 22:30 -!- phoenix24(gnodb@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-22 23:03 <flips> phoenix24, very possible
irc.oftc.net #zumastor log beginning Sun Mar 23 00:00:01 PDT 2008
2008-03-23 00:54 -!- MaZe(~MaZe@gaia.ifj.edu.pl) has joined #zumastor
2008-03-23 02:03 -!- flips(~phillips@phunq.net) has joined #zumastor
2008-03-23 09:02 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-23 09:40 <willn> 
ACTION goes out for brunch

2008-03-23 10:11 -!- phoenix24(iavxc@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 10:26 <phoenix24> Hi
2008-03-23 11:14 -!- phoenix24(uewlw@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 11:14 <phoenix24> I've a confusion over the "struct disksuper", What is it used for ?
2008-03-23 12:19 -!- phoenix24(wph@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 12:24 <shapor> phoenix24: thats the on-disk superblock
2008-03-23 12:25 <shapor> struct superblock has some additional values which are not stored persistently on disk
2008-03-23 12:25 <phoenix24> ok! got it.
2008-03-23 12:26 <shapor> i suppose that could use a comment
2008-03-23 12:26 <phoenix24> yeah! a lot of comments could be placed in.
2008-03-23 12:28 <phoenix24> currently only 64 simultaneous are possible, due to a "u64 snapmask."
2008-03-23 12:28 <shapor> i'm looking at your diagram
2008-03-23 12:28 <shapor> yes, that is true
2008-03-23 12:28 <shapor> making it larger would waste a lot of space
2008-03-23 12:28 <shapor> in fact, just storing it as u64 wastes a lot of space
2008-03-23 12:28 <phoenix24> a possible way to remove this limit is.. to replace "u64 snapmask" by "struct snapmask" or "array snapmask"
2008-03-23 12:29 <shapor> since many data chunks share the same snapmask
2008-03-23 12:29 <phoenix24> but, its important to have more simultaneous snapshots ?
2008-03-23 12:29 <shapor> something to note is that the whole btree really needs to fit in ram
2008-03-23 12:30 <shapor> to make it efficient
2008-03-23 12:30 <phoenix24> on the btree, is the a 2-4 .. 129-513 btree ?
2008-03-23 12:32 <phoenix24> on snapmask, one could replace "u64 snapmask" for each chunk with.. snapmask for the entire logical entry.
2008-03-23 12:32 <shapor> so we've talked about the idea of using run-length encoding to both support more (possibly unlimited) snapshots and reducing our metadata (memory) footprint at the same time
2008-03-23 12:32 <shapor> another method would be to add extent support to the btree
2008-03-23 12:32 <shapor> i think thats what you're getting at
2008-03-23 12:32 <phoenix24> yes
2008-03-23 12:33 <shapor> it will be tricky to get that right
2008-03-23 12:33 <phoenix24> but I haven't completely got the idea of adding extents support
2008-03-23 12:33 <shapor> will need some really good unit test to ensure correctness
2008-03-23 12:34 <phoenix24> 
ACTION thought to discuss it .. before actually writing up a project proposal.

2008-03-23 12:35 <shapor> have you looked at doc/desimplnotes.html ?
2008-03-23 12:36 <phoenix24> partially, it does not talk much on Metadata or the btree 
2008-03-23 12:37 <shapor> yeah i'm looking at it now
2008-03-23 12:37 <shapor> we had some more documention
2008-03-23 12:37 <shapor> frank made a diagram similar to page 2 of yours
2008-03-23 12:37 <shapor> i dont think it made it in to the repo though :(
2008-03-23 12:38 <phoenix24> is frank around ?
2008-03-23 12:38 <shapor> doubtful
2008-03-23 12:38 <shapor> not today
2008-03-23 12:38 <phoenix24> extents are contigous blocks of data.
2008-03-23 12:39 <flips> we could have an encoding thread on the mailing list
2008-03-23 12:39 <flips> it's a knotty problem
2008-03-23 12:39 <phoenix24> encoding thread ?
2008-03-23 12:39 <flips> how to encode the snapshot leaves better
2008-03-23 12:40 <flips> good research project
2008-03-23 12:40 <phoenix24> yeah!
2008-03-23 12:42 <phoenix24> which other Filesystems should I look at, to understand the extents support ?
2008-03-23 12:42 <flips> btrfs would be a good start
2008-03-23 12:42 <phoenix24> it seems fairly simple to me as of now :(
2008-03-23 12:42 <flips> but not that ddsnap has a harder extents problem to solve than btrfs, xfs, jfs or reiserfs
2008-03-23 12:43 <shapor> phoenix24: I just checked in Frank's btree diagram in to the doc/ directory
2008-03-23 12:43 <flips> s/not/note/
2008-03-23 12:43 <shapor> I had it in my email
2008-03-23 12:43 <shapor> BTree.pdf
2008-03-23 12:43 <flips> because ddsnap has multiple overlapping extents... fun to edit. Not.
2008-03-23 12:44 <phoenix24> flips: could you send the BTree.pdf to mailing list ?
2008-03-23 12:45 <flips> let's find the url
2008-03-23 12:45 <willn-colo> zumastor-commits: [zumastor commit] r1490 - trunk/doc
2008-03-23 12:48 <phoenix24> Is it possible to create a new, svn branch called, "soc".. where students could poolin or put experimental code ?
2008-03-23 12:48 <phoenix24> as I wanted to do some code cleanup ?
2008-03-23 12:49 <phoenix24> or maybe svn branch called "soc-experimental" :P
2008-03-23 12:50 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1490 (source)
2008-03-23 12:50 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1490 (source)
2008-03-23 12:53 <shapor> phoenix24: http://code.google.com/p/zumastor/source/browse/trunk/doc/BTree.pdf
2008-03-23 12:53 <phoenix24> got it!
2008-03-23 12:54 <flips> firefox refuses to display it, says it is too large
2008-03-23 12:54 <phoenix24> ah! there's a Cleanup_roadmap.txt too :)
2008-03-23 12:54 <phoenix24> yeah
2008-03-23 12:54 <shapor> flips: have you looked at Cleanup_roadmap.txt
2008-03-23 12:54 <flips> no, code.google.com
2008-03-23 12:54 <flips> fricking pos
2008-03-23 12:55 <shapor> i really dont like the new svn browsing interface on code.google.com
2008-03-23 12:55 <shapor> the old wwwsvn or whatever was much nicer
2008-03-23 12:55 <shapor> the new one seems to ignore the mime-type propset too :(
2008-03-23 12:56 <phoenix24> doesn't open the branches/sud-dirs for me
2008-03-23 12:56 <flips> I didn't even notice the change
2008-03-23 12:56 <shapor> yeah that link i posted is no good
2008-03-23 12:56 <shapor> wtf
2008-03-23 12:56 <flips> :-/
2008-03-23 12:56 <flips> mie-type propset?
2008-03-23 12:56 <shapor> "svn propset svn:mime-type text/html desimplnotes.html"
2008-03-23 12:57 <shapor> that was how we set properties on files in svn
2008-03-23 12:57 <shapor> so the old wwwsvn could set the mim tpy eproperly
2008-03-23 12:57 <shapor> so they display pretty in the browser
2008-03-23 12:57 <flips> it's quite sluggish
2008-03-23 12:57 <shapor> well this blows, i'll check it in to the www directory so it appears on zumastor.org
2008-03-23 12:57 <flips> try this: http://phunq.net/ddtree
2008-03-23 12:58 <flips> my nice penquin disappeared when I upgraded gitweb
2008-03-23 13:00 <shapor> it overlapped with the text anyway
2008-03-23 13:00 <shapor> was it a floating div or something?
2008-03-23 13:00 <flips> only on the summary page
2008-03-23 13:00 <flips> or on a small display
2008-03-23 13:01 <flips> html is so lame
2008-03-23 13:01 <shapor> man its way too nice out to be indoors right now
2008-03-23 13:01 <flips> true
2008-03-23 13:01 <flips> got to try out the m$ frisbees with dana
2008-03-23 13:02 <shapor> watch out i think those are calibrated to incapacitate open sourcers
2008-03-23 13:02 <shapor> i tossed mine around and razors deployed
2008-03-23 13:04 <flips> cleverrrr
2008-03-23 13:04 <shapor> i just moved BTree.pdf from trunk/doc/ to www/
2008-03-23 13:04 <shapor> so it will be served properly from www.zumastor.org/BTree.pdf
2008-03-23 13:04 <shapor> as soon as the cron job runs :)
2008-03-23 13:05 <shapor> and there it is
2008-03-23 13:05 <willn-colo> zumastor-commits: [zumastor commit] r1491 - trunk/doc www
2008-03-23 13:06 <flips> I need to do something about that rchunk braindamage
2008-03-23 13:06 <shapor> hrm I thought we were going to release 0.7 on Thursday
2008-03-23 13:06 <shapor> the site still says 0.6 is current
2008-03-23 13:06 <flips> did we release?
2008-03-23 13:06 <shapor> I dont think so?
2008-03-23 13:07 <shapor> the final changes have been in for a while 
2008-03-23 13:07 <flips> ping dank?
2008-03-23 13:07 <shapor> I think its a release when we call it a release
2008-03-23 13:07 <phoenix24> is willn-col the official commit bot ?
2008-03-23 13:07 <shapor> it looks like its willn's bot
2008-03-23 13:07 <shapor> zumalog is the irc logging bot
2008-03-23 13:08 <shapor> I hadn't noticed willn-colo before
2008-03-23 13:08 <shapor> mustbe something new
2008-03-23 13:08 <flips> :)
2008-03-23 13:08 <phoenix24> :)
2008-03-23 13:08 <flips> 
ACTION likes new shiny things

2008-03-23 13:09 <shapor> we should check the irc bots in to the repo
2008-03-23 13:09 <flips> it /doc?
2008-03-23 13:09 <shapor> mine is just a daemontols invokation of ii
2008-03-23 13:09 <flips> in /doc?
2008-03-23 13:09 <shapor> few lines of shell
2008-03-23 13:09 <flips> nice demo
2008-03-23 13:10 <flips> my change to gitweb.cgi didn't show on the page for some reason
2008-03-23 13:10 <shapor> the irclogs shell script cgi prints its source out at the bottom :)
2008-03-23 13:10 <shapor> the script is indeed published too, just not in svn http://www.zumastor.org/irclogs/run
2008-03-23 13:11 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1491 (source)
2008-03-23 13:11 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1491 (source)
2008-03-23 13:19 -!- phoenix24(uig@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 13:42 <phoenix24> shapor: Is the diagram correct ?
2008-03-23 13:42 <flips> looks good to me
2008-03-23 13:46 <phoenix24> flips: there should be more details into the Metadata Btree drawing.
2008-03-23 13:47 <flips> feel free :-)
2008-03-23 13:47 <willn> shapor: Yea, I'm trialling out the willn-colo bot. Its subscribed to the groups rss feed
2008-03-23 13:47 <flips> what details are missing?
2008-03-23 13:47 <phoenix24> the order of the btree.
2008-03-23 13:48 <flips> stored in the root
2008-03-23 13:48 <flips> superblock actually
2008-03-23 13:48 <flips> hmm
2008-03-23 13:48 <flips> no wait
2008-03-23 13:48 <flips> the root has a slightly different format, true
2008-03-23 13:58 <willn> supposedly this bot has all sorts of features. its logging now too.
2008-03-23 14:17 <DavidSev> Hi.  For finding old versions of files, is the best way to look in /var/run/zumastor/snapshot/* and see where it changes, or is there a better and more effecent way to do it?
2008-03-23 14:19 <willn> faik, thats the only way currently. 
2008-03-23 14:20 <DavidSev> ok, thanks.
2008-03-23 14:24 <DavidSev> If I disable Zumastor, (/etc/init.d/Zumastor stop or whatever) does it fully clean up, or does its kernal bits still eat a load of memory etc?
2008-03-23 14:29 -!- phoenix24(nnit@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 14:58 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-23 15:26 <willn> flips: ramback made it to reddit
2008-03-23 15:26 <flips> leet
2008-03-23 15:26 <flips> woot
2008-03-23 15:26 <flips> what's reddit?
2008-03-23 15:27 <flips> http://en.wikipedia.org/wiki/Reddit never mind
2008-03-23 15:28 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-23 15:28 <willn> It and digg are basically the same place now
2008-03-23 15:30 <flips> pcworld
2008-03-23 15:30 <flips> http://www.pcworld.idg.com.au/index.php/id;1829360654;pp;1
2008-03-23 15:30 -!- phoenix24(zrhlz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 15:30 <flips> linked from linuxworld
2008-03-23 15:31 <flips> jon is getting a lot of links out of this
2008-03-23 15:31 <willn> I bet violin is enjoying the attention
2008-03-23 15:32 <willn> Let me draw that graph... _____^
2008-03-23 15:32 <flips> about right
2008-03-23 15:50 <flips> http://phunq.net/ddtree <- new, improved
2008-03-23 16:02 <phoenix24> looks cool!
2008-03-23 18:44 -!- phoenix24(tptwa@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 18:49 <shapor> DavidSevDavidSevDavidSevsadfasdf
2008-03-23 18:50 <phoenix24> eew!
2008-03-23 18:50 <shapor> !oops
2008-03-23 18:50 <phoenix24> shapor, what's up ?
2008-03-23 18:50 <shapor> DavidSev: it should be fully cleaned up (after zumastor stop)
2008-03-23 18:51 <shapor> phoenix24 just stopping in online between rollerblading and dinner
2008-03-23 18:51 <shapor> :)
2008-03-23 18:51 <shapor> and having a keyboard mailfunction at the same time ;)
2008-03-23 18:51 <phoenix24> :D  :P
2008-03-23 18:55 <willn> !hello to shapor
2008-03-23 18:55 <willn-colo> hello willn
2008-03-23 18:55 <willn> well, that failed
2008-03-23 18:58 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-23 19:02 <phoenix24> Is there any utility, that can generate call graphs from the source-code ?
2008-03-23 20:50 -!- phoenix24(sttme@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 21:08 <flips> phoenix24, there is but I have not used it
2008-03-23 21:08 <flips> I forget the name
2008-03-23 21:09 <phoenix24> cbrowser ?
2008-03-23 21:10 <phoenix24> Its not really good, I just tried it.
2008-03-23 21:11 <phoenix24> flips, what are the defining points for, Zumastor Volume store creation?
2008-03-23 21:12 <flips> defining points?
2008-03-23 21:12 <phoenix24> I'm looking for hooking points for, resizing Zumastor Volume.
2008-03-23 21:12 <phoenix24> ah! I meant, starting points.
2008-03-23 21:13 <flips> ddsnap intialize
2008-03-23 21:13 <phoenix24> create_snapshot()
2008-03-23 21:13 <flips> create_snapshot has little to do with resizing
2008-03-23 21:14 <phoenix24> yes, actually.. create_snapshot() was for a different question.
2008-03-23 21:15 <flips> http://en.wikipedia.org/wiki/Call_graph
2008-03-23 21:16 <phoenix24> thank you!
2008-03-23 21:56 -!- phoenix24(auadkqz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-23 22:56 -!- phoenix24_(ixoq@h1119083.serverkompetenz.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Mon Mar 24 00:00:01 PDT 2008
2008-03-24 01:05 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-24 01:43 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-24 03:50 -!- phoenix24(nxr@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 05:17 -!- phoenix24(lbzrvqv@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 09:07 -!- MaZe(~MaZe@89.100.181.155) has joined #zumastor
2008-03-24 09:43 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-24 09:59 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-24 10:02 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-24 10:23 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-24 10:23 -!- shapor_(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-24 11:32 <willn> 
ACTION blinks

2008-03-24 11:39 -!- pgquiles_(~pgquiles@239.red-83-35-112.dynamicip.rima-tde.net) has joined #zumastor
2008-03-24 11:56 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-24 12:05 -!- cbsmith(~user@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-24 12:13 <willn-colo> zumastor-commits: [zumastor commit] r1492 - trunk/doc/callgraphs
2008-03-24 12:18 <willn-colo> zumastor-commits: [zumastor commit] r1493 - trunk/doc/callgraphs
2008-03-24 12:18 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1492 (source)
2008-03-24 12:18 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1492 (source)
2008-03-24 12:23 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1493 (source)
2008-03-24 12:23 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1493 (source)
2008-03-24 12:46 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-24 12:48 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has left #zumastor
2008-03-24 12:48 <DavidSev> ooh very nice.
2008-03-24 12:49 <DavidSev> doing CIA out of a job :)
2008-03-24 12:52 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-24 12:53 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-24 13:20 <willn> So, we had some spurrious failures from the PPA
2008-03-24 13:22 <pgquiles_> willn: fail to upload failures are annoying and have no solution. I wrote to the mailing list suggesting a "retry upload" button just like the "retry build" one but nobody answered :-/
2008-03-24 13:25 <pgquiles_> willn: is the -zumastor kernel working on hardy yet?
2008-03-24 13:28 <willn> I've not done another PPA build, but I have done local builds ok
2008-03-24 13:29 <pgquiles_> sorry, I didn't explain myself. I meant if ddsnap works fine, not if it builds fine.
2008-03-24 13:40 <willn> ah, I've not done testing on a proper machine, only my hardy laptop, so there was some clutter that could be an issue. flips also wanted me to eliminate the loop device, but I don't have any free disk at the moment
2008-03-24 13:42 <willn> I'm tempted to write a scraper for the launchpad build system and output rss
2008-03-24 13:49 <willn-colo> zumastor-commits: [zumastor commit] r1494 - trunk/doc/callgraphs
2008-03-24 13:54 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1494 (source)
2008-03-24 13:54 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1494 (source)
2008-03-24 13:59 -!- phoenix24(vfi@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 13:59 <phoenix24> Hi! everyone!
2008-03-24 14:01 <DavidSev> hi
2008-03-24 14:02 <phoenix24> I'm just curious, are the elections for Debian Project Leader on ?
2008-03-24 14:03 <phoenix24> Hi DavidSev 
2008-03-24 14:04 <cbsmith> Hey, has anyone tried using Zumastor w/dm-crypt?
2008-03-24 14:07 <jiayingz> hi phoenix24
2008-03-24 14:07 <jiayingz> di DavidSev
2008-03-24 14:08 <jiayingz> phoenix24, Debian Project Leader?
2008-03-24 14:09 <cbsmith> Okay, second question of the day: is there some reason why I don't see an "extern int mmon_interval;" somewhere in the headers?
2008-03-24 14:10 <cbsmith> Sorry, that second question is within ddsnap of course.
2008-03-24 14:13 <jiayingz> cbsmith, i think the reason is because it should be -DDDSNAP_MEM_MONITOR in Makefile, not -DDSNAP_MEM_MONITOR
2008-03-24 14:13 <jiayingz> i remember it once memory monitor once worked, so the Makefile must get that typo sometime during the code change
2008-03-24 14:14 <phoenix24> jiayingz: Yeah, for the Debian Project Leader.
2008-03-24 14:14 <cbsmith> jiayingz: But it looks like it is defined in ddsnap.c but used in ddsnapd.c...
2008-03-24 14:15 <jiayingz> cbsmith, it is a compile option
2008-03-24 14:15 <jiayingz> you can also give it as an argument to 'ddsnap server'
2008-03-24 14:16 <cbsmith> jiayingz: A compile option enables its existence, but once it springs in to existence... how can you use it from ddsnapd.c without getting an error about an undefined variable?
2008-03-24 14:17 <jiayingz> phoenix24, have you decided which soc project to work on?
2008-03-24 14:17 <cbsmith> Oddly, I don't get the error, so I'm obviously missing something.
2008-03-24 14:17 <jiayingz> cbsmith, i think the code structure has changed several times
2008-03-24 14:18 <cbsmith> jiayingz: Sounds like a safe bet :-)
2008-03-24 14:18 <phoenix24> jiayingz I'm zeroing on "64bit snapshot limit", other very likely option is "Volume resizing"
2008-03-24 14:18 <jiayingz> the reason why we didn't get compile error is probably because it never got compiled because of the makefile error
2008-03-24 14:20 <jiayingz> phoenix24, in my opinon, "64bit snapshot limit" may require a lot of interaction with the team here. the current btree code is quite complex
2008-03-24 14:20 <phoenix24> Yes, very true.
2008-03-24 14:20 <jiayingz> "volume resizing" may be easier to work on independently
2008-03-24 14:20 <phoenix24> ok
2008-03-24 14:21 <phoenix24> jiayingz: do you suggest, I think again.. or opt for "Volume Resizing" ?
2008-03-24 14:22 <pgquiles_> flips:  what do you think of a SoC on splitting ddsnap in libddsnap and a ddsnap command which essentially wraps libddsnap?
2008-03-24 14:22 <phoenix24> pgquiles_: +1
2008-03-24 14:23 <pgquiles_> next step, implementing zumastor in C++ as libzumastor + zumastor command (again, essentially a wrapper) :-)
2008-03-24 14:24 <phoenix24> creating a libddsnap, should indirectly effect a lot of code cleaning.
2008-03-24 14:24 <pgquiles_> third step, I develop a GUI without scraping zumastor's output  :-)
2008-03-24 14:24 <phoenix24> GUI should not be difficult, give the current status of the API.
2008-03-24 14:25 <jiayingz> phoenix24, i think it still depends on which project you are more interested in. i just think 'volume resizing' may be easier as a summer code project.
2008-03-24 14:25 <pgquiles_> well, I have to leave now. I'll read the log. bye!
2008-03-24 14:25 <phoenix24> pgquiles_ bye!
2008-03-24 14:25 <DavidSev> Isn't interacting with the current people and community etc part of the point of gsoc?
2008-03-24 14:25 <jiayingz> pgquiles_, good night
2008-03-24 14:26 <phoenix24> jiayingz: yeah.
2008-03-24 14:28 <jiayingz> DavidSev, yes. but we need to take into account that it is remote interaction
2008-03-24 14:29 <willn-colo> zumastor-commits: [zumastor commit] r1495 - trunk/doc/callgraphs
2008-03-24 14:34 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1495 (source)
2008-03-24 14:34 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1495 (source)
2008-03-24 14:46 -!- phoenix24(ngg@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 14:56 -!- phoenix24_(icsxs@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 15:06 <flips> pgquiles_, need to think about that
2008-03-24 15:07 <flips> doesn't make a lot of sense to write a library before the component interfaces reach something resembling their final form
2008-03-24 15:07 <flips> however, some modularizing work would help prepare the way for a library
2008-03-24 15:08 <flips> not really  techncail for enough for a proper SoC project I think
2008-03-24 15:08 <flips> ack
2008-03-24 15:08 <flips> not really techncal enough for a SoC project
2008-03-24 15:09 <flips> pgquiles_, I agree that implementing zumastor in C++ is a logical step
2008-03-24 15:09 <flips> and somebody might want to take a look at my already-written python prototype
2008-03-24 15:10 <flips> ddsnap/scripts/hackygenpatch.sh ?
2008-03-24 15:11 <flips> hmm, doc/callgraphs is taking an awful long time to update
2008-03-24 15:12 <flips> your basic 11 meg checked into the doc branch
2008-03-24 15:12 <flips> into the code repository I mean
2008-03-24 15:12 <flips> ouch
2008-03-24 15:13 <flips> maybe we better move that to a peer repository
2008-03-24 15:13 <flips> don't really agree with having binaries, blobs etc in the code tree
2008-03-24 15:14 <flips> we should just have the script that generates the callgraphs in the code repo
2008-03-24 15:14 <flips> likewise, there should be no ddsnap.ps, but the source for that isntead
2008-03-24 15:23 <flips> http://atomized.org/2005/09/subversion-sucks/
2008-03-24 15:23 <flips> wow
2008-03-24 15:23 <flips> what I found when trying to remember how to make a branch in svn
2008-03-24 15:23 <flips> I wonder if I really want to do this
2008-03-24 15:28 <phoenix24> flips, the doc/Cleanup_roadmap.txt could thus be approproately update. attached to release or milestone ?
2008-03-24 15:29 <flips> either
2008-03-24 15:30 <phoenix24> +1 for milestone
2008-03-24 15:31 <flips> that doc doesn't really cover the main things that need cleaning up unfortunately
2008-03-24 15:31 <flips> such as protocol messiness, option parsing...
2008-03-24 15:32 <phoenix24> yesh, it talks about certain parts.. nothing detailed.
2008-03-24 15:32 <flips> ...cut and paste syndrome...
2008-03-24 15:32 <flips> why not open a cleanup roadmap thread on the mailing list?
2008-03-24 15:32 <phoenix24> would be great.
2008-03-24 15:32 <flips> we can take a digest of it and improve that doc
2008-03-24 15:33 <flips> feel free :-)
2008-03-24 15:33 <phoenix24> and, an experimental (unstable) svn branch too
2008-03-24 15:33 <flips> I will join in if you cc me
2008-03-24 15:33 <flips> or even if not
2008-03-24 15:33 <flips> will watch for it
2008-03-24 15:33 <flips> I was just about to make that very thing
2008-03-24 15:33 <phoenix24> wow!
2008-03-24 15:34 <flips> then I read this... http://atomized.org/2005/09/subversion-sucks/ "Merging is virtually impossible!"
2008-03-24 15:34 <flips> now I am thinking twice
2008-03-24 15:34 <phoenix24> :P
2008-03-24 15:35 <phoenix24> 
ACTION wonders if he's only one intensely _wining_ about the code cleaning.

2008-03-24 15:35 <flips> why don't I just set up a git tree for the experimental branch
2008-03-24 15:35 <flips> code cleanup is really really important
2008-03-24 15:35 <flips> it's part of the zen of hacking
2008-03-24 15:35 <phoenix24> Zen as in Japanese :)
2008-03-24 15:36 <flips> as in buhddist I think
2008-03-24 15:37 <phoenix24> does Zumastor have a devel branch on Git too ?
2008-03-24 15:37 <flips> http://en.wikipedia.org/wiki/Zen
2008-03-24 15:37 <flips> not yet
2008-03-24 15:37 <flips> I think I will create a git tree
2008-03-24 15:37 <flips> give me 5 mins to think about it
2008-03-24 15:37 <flips> http://phunq.net/ddtree <- would be here
2008-03-24 15:38 <phoenix24> ok!
2008-03-24 15:38 <flips> need to practice my find-fun to delete all the .snv dirs
2008-03-24 15:38 <flips> find-fu that is
2008-03-24 15:42 <flips> shapor, ping
2008-03-24 15:42 <shapor> flips pong
2008-03-24 15:43 <flips> I think we should move the callgraph.png to www
2008-03-24 15:43 <shapor> gah, no git tree
2008-03-24 15:43 -!- phoenix24(fjhswkd@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 15:43 <flips> it's for me
2008-03-24 15:43 <flips> then I send patches from it
2008-03-24 15:43 <shapor> yeah, you can just use gitsvn for that
2008-03-24 15:43 <flips> I just don't see working on a branch in svn as a good idea
2008-03-24 15:44 <flips> ah, somebody had this problem before
2008-03-24 15:44 <shapor> http://www.kernel.org/pub/software/scm/git/docs/git-svn.html
2008-03-24 15:44 <flips> it may be a good idea, but people seem to be scared of merging in svn
2008-03-24 15:44 <shapor> merge with git
2008-03-24 15:44 <shapor> commit to svn
2008-03-24 15:44 <flips> right
2008-03-24 15:44 <flips> best tool for the job
2008-03-24 15:46 <flips> but... 12 meg of callgraphs/*.svg
2008-03-24 15:46 <flips> that needs to go out of our code tree I think
2008-03-24 15:46 <flips> ok?
2008-03-24 15:46 <flips> 
ACTION gets ready to pull the trigger

2008-03-24 15:47 <phoenix24> callgraphs/*.png can be moved to:  www/
2008-03-24 15:47 <phoenix24> callgraphs/*.* can be moved to:  www/
2008-03-24 15:47 <flips> is is the svgs that are 12 meg
2008-03-24 15:47 <flips> ouch
2008-03-24 15:48 <flips> worse than any file format ever invented
2008-03-24 15:50 <flips> svn copy ../trunk/doc/callgraphs/* callgraphs
2008-03-24 15:50 <flips> svn: Client error in parsing arguments
2008-03-24 15:50 <flips> :P
2008-03-24 15:59 <willn-colo> zumastor-commits: [zumastor commit] r1496 - www/callgraphs
2008-03-24 16:01 <flips> seems to have worked
2008-03-24 16:02 <flips> should move the .ps too, but that was stressful
2008-03-24 16:04 <willn-colo> zumastor-commits: [zumastor commit] r1497 - trunk/doc/callgraphs
2008-03-24 16:05 <flips> irc commit messages are cool
2008-03-24 16:10 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1497 (source)
2008-03-24 16:10 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1497 (source)
2008-03-24 16:10 <flips> accepted?
2008-03-24 16:10 <phoenix24> yes, I think too
2008-03-24 16:11 <phoenix24> just upp'd my svn
2008-03-24 16:14 <phoenix24> Why is there an openvz-zumastor-merge script lying around  in the source code ?
2008-03-24 16:14 <flips> good question
2008-03-24 16:14 <flips> dld might know
2008-03-24 16:14 <phoenix24> does Zumastor already has support for openvz
2008-03-24 16:14 <flips> I think so
2008-03-24 16:14 <dld> Nope, no support yet.  Still trying to make 2.6.24 coexist with ddsnap and openvz.
2008-03-24 16:15 <dld> That support probably won't ever be in zumastor (the script), since ddsnap replication would be used as part of openvz's migration, and in it's scripts.
2008-03-24 16:16 <flips> sucks that cp does not know how to leave out hidden files
2008-03-24 16:17 <flips> dld, so zumastor is to be used by openvs and not the other way round?
2008-03-24 16:17 <dld> Two uses I can think of:
2008-03-24 16:17 <dld> 1) we do the continuous build inside OpenVZ for isolation at full performance.
2008-03-24 16:18 <dld> 2) OpenVZ uses ddsnap to migrate faster than rsync.
2008-03-24 16:18 <dld> Neither would be part of zumastor-the-fileserver-script.
2008-03-24 16:20 <phoenix24> ok, then it should be remove from the Zumastor trunk/ ?
2008-03-24 16:21 <flips> dld?
2008-03-24 16:22 <dld> If removed I'll have to recreate it at some point, and don't want to have to keep track of it outside the repository.
2008-03-24 16:22 <dld> It's still part of zumastor-the-project, just not zumastor-the-script.
2008-03-24 16:23 <cbsmith> dld: Sounds right?
2008-03-24 16:23 <dld> cbsmith: ?
2008-03-24 16:23 <cbsmith> dld: I meant the rationale sounds about right.
2008-03-24 16:24 <cbsmith> dld: I'm very interested in seeing a combined OpenVZ + Zumastor build. Any pointers?
2008-03-24 16:24 <willn> voodoo and pixie dust are a step in the right direction ;)
2008-03-24 16:24 <willn> iirc, the 2.6.24 patches are not fun to integrate
2008-03-24 16:25 <cbsmith> willn: Yes, I was leaving that heavy work to others. ;-)
2008-03-24 16:25 <dld> 2.6.22 OpenVZ branch is frozen and abandoned, and 2.6.24 is only just about (we hope) to work for Zumastor, but OpenVZ's 2.6.24 has all the features that 2.6.18 had (migration, the main reason to integrate in the first place).
2008-03-24 16:26 <dld> So it's still a work in progress, and stalled on our 2.6.24 port.
2008-03-24 16:27 <willn> From my end, I need to get some testing done on 2.6.24 to see if the performance problem I think exists exists
2008-03-24 16:28 <dld> And we're still not getting successful 2.6.24 boots from the build from trunk.  But there are some working 2.6.24 builds somewhere?
2008-03-24 16:29 <willn> Yep. I can grant you a working .deb for the hardy machine of your choice
2008-03-24 16:29 <flips> I always forget how to find out the current commit in svn
2008-03-24 16:30 <willn> flips: current commit?
2008-03-24 16:30 <flips> rev number in a given repo
2008-03-24 16:30 <flips> latest rev
2008-03-24 16:30 <phoenix24> "svn info "
2008-03-24 16:30 <dld> flips: Re git-tree.  I did some (very) preliminary work trying to import every subversion revision into a git tree.  Would make sense if git is to become the replacement to keep the historical revisions - at least the actual releases.
2008-03-24 16:31 <flips> thanks
2008-03-24 16:31 <flips> dld, I'm just bringing in the plaintext from trunk
2008-03-24 16:31 <flips> not sure how I will track trunk ;-)
2008-03-24 16:31 <flips> will be interesting
2008-03-24 16:31 <dld> svnversion sometimes work.
2008-03-24 16:32 <willn> fwiw, I think that mixing version control systems usually ends in headache
2008-03-24 16:32 <dld> Yes, I think it should be a clean cutover.  And separate repos for the kernel tree and userspace.
2008-03-24 16:32 <flips> git is scary fast
2008-03-24 16:32 <flips> except at cloning
2008-03-24 16:33 <dld> It's fast there if you use tmpfs.  :)
2008-03-24 16:33 <willn> Doesn't matter how fast it is if all your code is somewhere else =p
2008-03-24 16:33 <flips> on upstream or downstream?
2008-03-24 16:34 <phoenix24> slooow on downstream :(
2008-03-24 16:34 <dld> cloning from git-daemon is much faster than over http:, so despite the evils, git-daemon has a purpose.
2008-03-24 16:35 <dld> SVNREV=`awk '/^[0-9]+$/ { print $1; }' SVNREV || svnversion | tr [A-Z] [a-z] ||
2008-03-24 16:35 <dld> svn info zumastor | grep ^Revision:  | cut -d\  -f2`
2008-03-24 16:36 <dld> Seriously.  :(  That's what it took to just get the repository version to embed in all the build scripts under all the environments supported.
2008-03-24 16:37 <flips> http://phunq.net/ddtree?p=zumastor/.git;a=summary
2008-03-24 16:37 <flips> :-)
2008-03-24 16:37 <flips> ok, I take back what I said about Linus having summa that stupid
2008-03-24 16:38 <flips> this rocks
2008-03-24 16:38 <flips> I like O(1)
2008-03-24 16:40 <flips> whoops I need to do it over
2008-03-24 16:40 <flips> and not check in the .o files this time
2008-03-24 16:40 <flips> well, that should take another 2 minutes ;-)
2008-03-24 16:41 <flips> only a total of 5 secs of computer time, about
2008-03-24 16:41 <flips> if that
2008-03-24 16:43 -!- phoenix24_(cfvjqf@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 16:44 <willn> willn-colo: rss-add ppa http://compbrain.net/zumastor-team_ppa.xml
2008-03-24 16:44 <willn-colo> rss item added
2008-03-24 16:46 <phoenix24_> flips: the git url would be "git clone http://phunq.net/ddtree/zumastor zumastor-git" ?
2008-03-24 16:49 <flips> it would
2008-03-24 16:49 <flips> and I will make an exception to my "no cloning" request for you ;-)
2008-03-24 16:49 <flips> and everybody else on zumastor@googlegroups
2008-03-24 16:50 <phoenix24_> git clone fails for me : http://paste.ubuntu.com/6052/
2008-03-24 16:57 <flips> hmm
2008-03-24 16:58 <flips> I need to say some git magical thing
2008-03-24 16:58 <flips> oh right
2008-03-24 16:58 <flips> I need to touch a file
2008-03-24 16:58 <flips> well... later
2008-03-24 16:59 -!- phoenix24(jsmjmfz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-24 17:02 <willn> why arent you letting people clone?
2008-03-24 17:05 <willn> I want to fix the PPA autobuilder to only submit when packaged sources change, but that list is somewhat grey if I recall
2008-03-24 17:23 <willn> willn-colo: rss-del ppa
2008-03-24 17:23 <willn-colo> rss item deleted
2008-03-24 17:23 <willn> willn-colo: rss-add ppa http://pipes.yahoo.com/pipes/pipe.run?_id=eK2OgAH63BGZqpnc8SvLAg&_render=rss
2008-03-24 17:23 <willn-colo> rss item added
2008-03-24 17:23 <willn> There, now ppa notices are sorted.
2008-03-24 17:31 <willn> willn-colo: rss-watch ppa 300
2008-03-24 17:31 <willn-colo> watcher thread started
2008-03-24 19:10 <flips> willn, that is unspeakably cool
2008-03-24 19:11 <willn> willn-colo: hm?
2008-03-24 19:11 <willn-colo> no infoitems known for hm
2008-03-24 19:11 <willn> er
2008-03-24 19:11 <flips> yes
2008-03-24 19:11 <willn> flips: hm?
2008-03-24 19:11 <flips> cloning = bad for the moment because the repo is running on my workstation
2008-03-24 19:11 <flips> but, ok in small numbers
2008-03-24 19:11 <flips> like 50 or so
2008-03-24 19:11 <flips> if it is 500...
2008-03-24 19:12 <willn> ah
2008-03-24 19:12 <willn> was that the unspeakably cool thing?
2008-03-24 19:12 <flips> I did pick up a box to be a dedicated server, but dan comandeered it
2008-03-24 19:12 <flips> dana did I mean
2008-03-24 19:17 <willn> yahoo pipes is pretty nifty.
2008-03-24 19:32 <willn-colo> ppa: Pipes Output
2008-03-24 19:33 <willn> er
2008-03-24 19:33 <willn> what the hell?
2008-03-24 19:40 <flips> willn, the cool thing is willn-colo
2008-03-24 19:40 <flips> I guess I must have a bot too if I want to be leet
2008-03-24 20:02 <willn> nah. Their generally a pita
2008-03-24 20:02 <willn> Eggdrop is a horrible tcl creation
2008-03-24 20:58 <flips> wow, tcl
2008-03-24 20:58 <flips> "tool control language" ?
2008-03-24 21:12 <flips> tool command language
2008-03-24 21:12 <flips> http://en.wikipedia.org/wiki/Tcl
2008-03-24 21:30 <flips> hmm, journal replay patch still applies to ddsnapd.c after 6 months of development
2008-03-24 21:31 <flips> I think that means the code did not change much
2008-03-24 21:31 <flips> and it needs to
2008-03-24 21:31 <shapor> willn: i like ii
2008-03-24 21:31 <flips> ii?
2008-03-24 21:32 <shapor> irc bot... er client
2008-03-24 21:32 <flips> ah
2008-03-24 21:32 <shapor> http://zumastor.org/irclogs/#about
2008-03-24 21:32 <flips> have we got a resolution on which is the best server kit?
2008-03-24 21:32 <flips> http://serverkit.org/ <- shapor's suggestion
2008-03-24 21:33 <flips> what was jeff's suggestion?
2008-03-24 21:34 <shapor> you mean phil's
2008-03-24 21:34 <shapor> i dont remember
2008-03-24 21:34 <shapor> we could ask him
2008-03-24 21:35 <flips> daniel@moonbase:/src/serverkit-2.1.0/lib$ wc *c *h
2008-03-24 21:35 <flips> ...
2008-03-24 21:35 <flips>   3329  13651 105889 total
2008-03-24 21:35 <flips> tolerable
2008-03-24 21:35 <flips> server_db??
2008-03-24 21:35 <flips> database shell
2008-03-24 21:35 <flips> ok
2008-03-24 21:36 <shapor> yeah
2008-03-24 21:36 <shapor> so managing database connections
2008-03-24 21:36 <shapor> you can have a pool of mysql db connections
2008-03-24 21:36 <shapor> the basic functionality is quite small
2008-03-24 21:36 <flips> we'd better add mysql to the mix quick ;)
2008-03-24 21:37 <shapor> there are some other things it provides which are convenient, like heaps
2008-03-24 21:37 <tim> i'm architecting a media production facility
2008-03-24 21:37 <shapor> tim: :)
2008-03-24 21:37 <flips> better heap than malloc I presume?
2008-03-24 21:38 <tim> and I'm thinking "raid 5 or raid 1 with snapshotting"
2008-03-24 21:38 <shapor> yeah the idea is not to malloc :)
2008-03-24 21:38 <tim> hey shapor
2008-03-24 21:38 <flips> somebody has to malloc
2008-03-24 21:39 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-24 21:40 <tim> think there's more performance/$ out of 20TB of storage using zumastor vs raid 5?
2008-03-24 21:40 <shapor> flips: http://serverkit.org/serverkit-2.0.0-api.txt
2008-03-24 21:40 <flips> I'm reading the code, it's small enough
2008-03-24 21:40 <tim> Or 60TB with say, 30 clients?
2008-03-24 21:40 <flips> just checked out the dll loader
2008-03-24 21:40 <flips> I code that stuff in just a few lines normally
2008-03-24 21:41 <flips> but there is probably something useful there
2008-03-24 21:41 <shapor> tim: "zumastor vs raid 5" doesn't make sense
2008-03-24 21:42 <tim> raid 5 is mirrored, right?
2008-03-24 21:42 <flips> it is, true
2008-03-24 21:42 <flips> it would be ddsnap vs raidx
2008-03-24 21:42 <flips> typically you would use one, the other, or both ;-)
2008-03-24 21:42 <tim>  so if I could eliminate my mirroring, and throw less expensive spindles at the back up
2008-03-24 21:43 <flips> in theory, yes
2008-03-24 21:43 <tim> maybe throw a violin in there as a metadata server
2008-03-24 21:43 <shapor> metadata server for what?
2008-03-24 21:43 <flips> in practice you better wait for us to optimize writes more before substituting replication for mirroring
2008-03-24 21:44 <flips> oh
2008-03-24 21:44 <flips> that would help :-)
2008-03-24 21:44 <tim> all the clips I'm backing up
2008-03-24 21:44 <flips> replicating is cool
2008-03-24 21:44 <shapor> so your application supports a separate storage for metadata?
2008-03-24 21:44 <flips> mirring in itself is not a backup solution, whereas replicating very nearly is, by itself
2008-03-24 21:45 <tim> the way I see it- you have a drive failure, you could replace the drive and revert to the last snapshot quicker than a raid rebuild
2008-03-24 21:45 <tim> lets assume I have a metadata server
2008-03-24 21:45 <tim> its going to be a lot faster to find clips that way
2008-03-24 21:46 <shapor> tim the problem is reverting to a snapshot doesn't prevent you from rebuilding your array
2008-03-24 21:46 <shapor> typically the raid sits underneath zumastor
2008-03-24 21:46 <shapor> unless you mean wipe and copy from the backup mahcine you're replicating to
2008-03-24 21:46 <tim> I get the reverting doesn't prevent you- its a matter of could I 
2008-03-24 21:46 <shapor> which sounds... unnatural if not just plain wrong
2008-03-24 21:47 <flips> tim is most probably right, for some configurations
2008-03-24 21:47 <flips> had not thought about that
2008-03-24 21:47 <flips> rebuilding raid5 can be excrutiating
2008-03-24 21:47 <tim> it just hit me as I was budgeting a buildout
2008-03-24 21:47 <shapor> s/from rebuilding/from needing to rebuild/
2008-03-24 21:47 <flips> tim, I'd say go with the very tried and true though
2008-03-24 21:47 <shapor> raid 10 = more better :)
2008-03-24 21:48 <tim> raid 10 it is
2008-03-24 21:48 <shapor> you only get 50% of your raw disk
2008-03-24 21:48 <flips> just have the replication sitting in the back as a sweetener, an upgrade path
2008-03-24 21:48 <flips> people like upgrade paths
2008-03-24 21:48 <tim> you get what I'm proposing though
2008-03-24 21:48 <flips> and people like "coming soon", knowing it is coming to them
2008-03-24 21:48 <flips> yes
2008-03-24 21:48 <flips> it makes sense
2008-03-24 21:48 <flips> just let us make it work a little better for local filesystems first
2008-03-24 21:49 <flips> we just recently got it acceptable for NFS
2008-03-24 21:49 <tim> if my snapshots are fine grained, it could eliminate the need for mirroring, and cut the number of drives and thus power consumption in half
2008-03-24 21:50 <flips> I need to see that with my own eyes :)
2008-03-24 21:50 <shapor> tim, no
2008-03-24 21:50 <tim> not in half, but performance drives in half
2008-03-24 21:50 <tim> then add back in the backing store
2008-03-24 21:50 <flips> if your load is read-mostly, that might be true
2008-03-24 21:50 <shapor> tim: what do you mean by mirroring?
2008-03-24 21:51 <tim> it would be mostly reads
2008-03-24 21:51 <tim> look at the xserve- standard edit facility storage
2008-03-24 21:51 <tim> 14 drives on two controllers
2008-03-24 21:51 <tim> raid 10
2008-03-24 21:51 <tim> one side goes down, you always have the insurance of the other
2008-03-24 21:52 <tim> problem is, bandwidth suz
2008-03-24 21:52 <tim> sux
2008-03-24 21:52 <tim> 320Gb/s
2008-03-24 21:52 <tim> and with multiple clients, performance blows
2008-03-24 21:53 <tim> so throw more drives at it
2008-03-24 21:53 <tim> I'd like 1GB/s linear performance
2008-03-24 21:53 <tim> and use slower drives for snapshotting
2008-03-24 21:53 <flips> what is the balance of read vs write?
2008-03-24 21:53 <tim> 90/10
2008-03-24 21:54 <tim> you're editing
2008-03-24 21:54 <tim> not writing until you're done editing
2008-03-24 21:54 <flips> replicating is looking attractive then, if all you want to do is get the data back if the worst happens
2008-03-24 21:54 <tim> exactly
2008-03-24 21:54 <flips> and if you can tolerate being down for a while if a drive goes out
2008-03-24 21:54 <shapor> tim i thought xserve was just a 3 drive 1u box?
2008-03-24 21:54 <tim> you're down with raid 10 if a drive goes down too
2008-03-24 21:55 <tim> xraid, sorry
2008-03-24 21:55 <tim> discountinued 
2008-03-24 21:55 <tim> Avid was a big storage vender until recently
2008-03-24 21:55 <shapor> you're not down if you lose a drive in raid10
2008-03-24 21:56 <tim> not for a rebuild?
2008-03-24 21:56 <shapor> rebuilds are fast with raid10
2008-03-24 21:57 <tim> if one side goes down, the other gets pegged for I/O on rebuild, and you take a performance hit- at least that's what I understand
2008-03-24 21:57 <shapor> its just a stripe of mirrors
2008-03-24 21:57 <tim> ok, so let's say you have 60 TB of storage
2008-03-24 21:57 <tim> a drive or controller takes a digger
2008-03-24 21:58 <tim> how long before you're back up vs back up with a snapshot revert
2008-03-24 21:58 <tim> might be a good test for SoC
2008-03-24 21:59 <tim> hello?
2008-03-24 21:59 <tim> ack, chased off all the geeks
2008-03-24 21:59 <tim> lol
2008-03-24 22:00 <shapor> a raid 10 rebuild should only take as copying an entire disk
2008-03-24 22:00 <shapor> one disk to one disk
2008-03-24 22:00 <shapor> because there are essentially 2 copies of every drive in the stripe
2008-03-24 22:00 <shapor> so you can read from its pair 
2008-03-24 22:01 <flips> shapor, his point is, given twice the drives he would rather have twice the speed than twice the safety
2008-03-24 22:01 <shapor> dont have to read from all the other drives and compute parity like you do with raid 5
2008-03-24 22:01 <flips> I think
2008-03-24 22:01 <tim> yes
2008-03-24 22:01 <tim> flips yes
2008-03-24 22:01 <shapor> so just raid 0?
2008-03-24 22:02 <shapor> theres no point of doing raid5 if you dont plan on rebuilding
2008-03-24 22:02 <tim> so throw your $$$ at performance drives, then $ for snapshoting the changed blocks on those performance drives
2008-03-24 22:03 <tim> ok so stripe just for performance, and your data integerity comes from the snapshot revert
2008-03-24 22:03 <shapor> tim: not just "snapshotting the changed blocks"
2008-03-24 22:03 <shapor> replicating and keeping more snapshot you mean.. i think
2008-03-24 22:03 <tim> if you're fine grained on your snapshot, like every 5 minutes, then the slow disk never get too far behind the performance disk
2008-03-24 22:03 <shapor> snapshot data is worthless without the origin data
2008-03-24 22:04 <shapor> yeah
2008-03-24 22:04 <shapor> you mean 5 minute replication interval
2008-03-24 22:04 <shapor> provided it can keep up
2008-03-24 22:04 <tim> yeah
2008-03-24 22:04 <shapor> during a period of heavy io
2008-03-24 22:04 <shapor> it might not be able to
2008-03-24 22:04 <shapor> i think editing is a load where it might not do much good
2008-03-24 22:04 <shapor> you do nothing for a while
2008-03-24 22:04 <shapor> then write a huge multi terabyte file
2008-03-24 22:04 <tim> meaning that the writes aren't so slow the next replication lands before the last write is complete
2008-03-24 22:04 <shapor> long linear io
2008-03-24 22:05 <shapor> your replicant will get behind if it has slow disk
2008-03-24 22:05 <shapor> why not just use ramback
2008-03-24 22:05 <tim> a render is a series of random reads with a long linear write
2008-03-24 22:05 <tim> seems that's maybe what I'm playing with in my head
2008-03-24 22:05 <shapor> then you dont need to worry about crashing
2008-03-24 22:06 <shapor> i thought of this application of ramback when everyone was arguing about persistency
2008-03-24 22:06 <tim> it feels like a ramback/zumastor approach
2008-03-24 22:06 <shapor> yeah, zumastor on the backing store only
2008-03-24 22:06 <shapor> your "fast cache" is a fast array
2008-03-24 22:06 <shapor> backing store can be lazy sata drives
2008-03-24 22:06 <shapor> with zumastor too
2008-03-24 22:07 <shapor> to keep lots of snapshots
2008-03-24 22:07 <tim> seems raid becomes redundant, pardon the pun
2008-03-24 22:07 <tim> seems we can simplify the raid controllers as a result
2008-03-24 22:07 <tim> especially if you're not calculating parity
2008-03-24 22:07 <shapor> i think that is a separate issue
2008-03-24 22:08 <shapor> i mean, you can just use a raid level without parity
2008-03-24 22:08 <shapor> or redundancy if you want to go that far
2008-03-24 22:08 <shapor> like raid 0 
2008-03-24 22:08 <shapor> although no
2008-03-24 22:08 <shapor> thats no good
2008-03-24 22:08 <shapor> because you'll lose data
2008-03-24 22:08 <shapor> if you lose the fast storage
2008-03-24 22:08 <shapor> because ramback writes out lazily
2008-03-24 22:09 <tim> hmm
2008-03-24 22:09 <shapor> you can get rid of rebuilding, if you dont mind being down to copy the data back
2008-03-24 22:10 <tim> I think I'm losing this arguement ;-)
2008-03-24 22:11 <shapor> i'm not arguing :)
2008-03-24 22:11 <shapor> just thinking "out loud"
2008-03-24 22:11 <tim> its more like I had a feeling there was an advantage
2008-03-24 22:11 <tim> thinking out loud, yes, but I'm losing the thinking inside my head
2008-03-24 22:11 <tim> meaning I'm talking myself out of my feeling
2008-03-24 22:12 <tim> wow
2008-03-24 22:12 <tim> how profound
2008-03-24 22:12 <tim> :-)
2008-03-24 22:13 <shapor> heh
2008-03-24 22:14 <shapor> but its a good idea
2008-03-24 22:14 <shapor> you just can't completely throw away redundancy on the fast storage
2008-03-24 22:14 <tim> it was an *almost* good idea
2008-03-24 22:14 <tim> right, just as you can't throw it away on violin
2008-03-24 22:14 <tim> actually, you can turn off the raid on violin
2008-03-24 22:15 <tim> but if you lose a module, you lose data
2008-03-24 22:36 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-24 23:15 <flips> can somebody please try: git clone http://phunq.net/git/zumastor
2008-03-24 23:17 <shapor> its going...
2008-03-24 23:17 <shapor> your upload sucks :P
2008-03-24 23:20 <flips> something gittish is broken
2008-03-24 23:20 <flips> it's not my upload
2008-03-24 23:20 <flips> waits forever even cloning locally
2008-03-24 23:21 <flips> git-http-fetch using 100% cpu
2008-03-24 23:21 <flips> bleah
2008-03-24 23:21 <flips> a bug for another day
2008-03-24 23:22 <flips> actually, it is a bug for #git, today
2008-03-24 23:58 <flips> actually, it is not
2008-03-24 23:58 <flips> it was just a messed symlink
2008-03-24 23:58 <flips> nothing to see there...
2008-03-24 23:58 <flips> time git clone http://phunq.net/git/zumastor
2008-03-24 23:58 <flips> real    0m25.670s
2008-03-24 23:58 <flips> user    0m0.284s
2008-03-24 23:58 <flips> sys     0m0.132s
irc.oftc.net #zumastor log beginning Tue Mar 25 00:00:01 PDT 2008
2008-03-25 00:09 <shapor> real    9m55.467s
2008-03-25 00:09 <shapor> user    0m4.164s
2008-03-25 00:09 <shapor> sys     0m2.064s
2008-03-25 00:09 <shapor> remotely..
2008-03-25 00:09 <shapor> strange.. it used a lot more cpu time
2008-03-25 00:19 -!- pgquiles(~pgquiles@236.Red-88-17-199.staticIP.rima-tde.net) has joined #zumastor
2008-03-25 00:20 <flips> what is the du?
2008-03-25 00:21 <shapor> 8.5m
2008-03-25 00:25 <flips> that is about right
2008-03-25 00:26 <flips> a little bigger than the original
2008-03-25 00:26 <flips> apparently git unpacks during clone
2008-03-25 00:27 <flips> time git clone http://phunq.net/git/zumastor
2008-03-25 00:27 <flips> real    0m1.423s
2008-03-25 00:27 <flips> user    0m0.260s
2008-03-25 00:27 <flips> sys     0m0.064s
2008-03-25 00:27 <flips> local clone
2008-03-25 00:31 <shapor> did it actually work, or get a 404?
2008-03-25 01:01 -!- pgquiles(~pgquiles@121.Red-81-39-99.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-25 01:07 <flips> it actually worked
2008-03-25 01:07 <flips> if you look in /home/zumastor/tmp you will see a repo
2008-03-25 01:08 <flips> now trying to explain the concept of forward compatibility to junio...
2008-03-25 01:08 <flips> he's not getting it
2008-03-25 01:08 <flips> also not getting the idea of repository versioning
2008-03-25 01:08 <flips> in the db schema sense
2008-03-25 01:08 <flips> sigh
2008-03-25 01:11 <flips> sad but true: git in debian stable is obsolete and will fail weirdly in various ways
2008-03-25 01:11 <flips> even sadder: installing a more recent git in debian stable is an ugly little project
2008-03-25 01:24 <pgquiles> flips: backport the 'experimental' package to stable, it's easy
2008-03-25 01:25 <flips> git clone git://git.kernel.org/pub/scm/git/git.git && cd git && git checkout v1.5.4.2 && make && sudo apt-get remove git && sudo make install
2008-03-25 01:26 <flips> I did
2008-03-25 01:26 <flips> easy is a... slight exaggeration
2008-03-25 01:26 <flips> did you do the pinning?
2008-03-25 01:26 <pgquiles> what pinning?
2008-03-25 01:26 <flips> what you have to do to install a backport
2008-03-25 01:27 <flips> there'
2008-03-25 01:27 <flips> there's a backport faq
2008-03-25 01:27 <flips> http://backports.org
2008-03-25 01:27 <pgquiles> oh, I don't do that :-)
2008-03-25 01:27 <pgquiles> it's usually safer and easier to:
2008-03-25 01:27 <pgquiles> 1) download the .tar.gz, the .diff.gz and the .dsc
2008-03-25 01:27 <pgquiles> 2) dpkg-source -x file.dsc
2008-03-25 01:28 <pgquiles> 3) go into the dir
2008-03-25 01:28 <pgquiles> 4) dpkg-buildpackage -rfakeroot
2008-03-25 01:28 <flips> heh
2008-03-25 01:28 <pgquiles> with recent versions of dpkg-buildpackage you don't need '-rfakeroot'
2008-03-25 01:28 <flips> not what you'd expect everybody to know though
2008-03-25 01:28 <pgquiles> O:-)
2008-03-25 01:28 <flips> I didn't know it
2008-03-25 01:29 <pgquiles> I usually do that when I backport from ubuntu unstable to ubuntu stable or from debian to ubuntu
2008-03-25 01:30 <flips> I am noticing as I compile git that every additional devel package needed is a nontrivial headache
2008-03-25 01:30 <flips> good reason to try to keep dependencies down to a dull roar
2008-03-25 01:31 <pgquiles> actually, I lied because I wanted to make it simpler but if dependency problems arise, you should replace step 4 with this:
2008-03-25 01:32 <pgquiles> 4a) debuild -S -sa
2008-03-25 01:32 <pgquiles> 4b) sudo pbuilder build ../file.dsc
2008-03-25 01:32 <pgquiles> but you need to have installed pbuilder and have run pbuilder create before 4b
2008-03-25 01:32 <flips> it's never easy...
2008-03-25 01:33 <pgquiles> pbuilder is an UML environment which downloads all the dependencies, builds the packages and stores them usually in /var/cache/pbuilder/result
2008-03-25 01:33 <flips> anyway, getting from git and building as above seems morally right in the case of git
2008-03-25 01:34 <flips> ok, that is enough git
2008-03-25 01:34 <flips> got to use this tool now
2008-03-25 01:34 <flips> make it earn its keep
2008-03-25 01:54 <flips> http://phunq.net/ddtree?p=zumastor/.git;a=summary <- "rewrite journal replay to be much more readable and ..."
2008-03-25 02:10 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor
2008-03-25 02:14 <pgquiles> 24'' monitors are so large they become uncomfortable next to a laptop :-/
2008-03-25 02:14 <shapor> uncomfortable how?
2008-03-25 02:14 <shapor> i just find the pixels way to large
2008-03-25 02:16 <pgquiles> uncomfortable as in I need a dockstation for the laptop to have both screens at the same height and even then, I have to turn my head too much to look at the left most of the 24'' screen
2008-03-25 02:19 <shapor> yeah, i have 2x 24" widescreen monitors in a dual head setup
2008-03-25 02:19 <shapor> its a bit strainging on the neck
2008-03-25 02:19 <shapor> i find myself concentrating windows near the center of the 2
2008-03-25 02:20 <shapor> unimportant things in the peripheral
2008-03-25 02:20 <pgquiles> I've just rotated the 24'' and put it in vertical
2008-03-25 02:21 <pgquiles> I noticed it last week when I brought it home from the office
2008-03-25 02:21 <pgquiles> s/it/it's possible
2008-03-25 02:23 <shapor> i have a 42" display about 6 feet away from me at home
2008-03-25 02:23 <pgquiles> attached to a computer? whow :-)
2008-03-25 02:23 <shapor> has approximately the same resolution as my laptop screen, which is annoying though
2008-03-25 02:24 <shapor> its more ergonomic than hunching over the laptop on the couch
2008-03-25 02:47 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-25 04:39 -!- pgquiles(~pgquiles@121.Red-81-39-99.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-25 06:26 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-25 06:58 -!- tim_(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-25 08:31 <willn> 
ACTION yawns

2008-03-25 08:45 -!- tim_(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-25 08:45 -!- tim__(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-25 09:50 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-25 10:16 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-25 10:47 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-25 11:31 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-25 11:40 -!- phoenix24(yigz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 11:54 -!- phoenix24(wxv@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 12:26 <flips> good morning
2008-03-25 12:26 <tim__> howdeee!
2008-03-25 12:26 <shapor> aloha
2008-03-25 12:27 <tim__> (mr enthusiasm this morning)
2008-03-25 12:55 -!- phoenix24_(arhme@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 13:05 <flips> scramjet time
2008-03-25 13:33 <dld> flips: yes.  Stable is stable as of around two years ago, and git has changed in that time.  backports.org is your friend if you need one thing new, but want the rest to stay unchanged.
2008-03-25 13:33 <dld> deb http://www.backports.org/debian etch-backports main contrib non-free
2008-03-25 13:34 <dld> apt-get -t etch-backports install git-core
2008-03-25 13:34 <dld> See http://backports.org/ for more detail.
2008-03-25 14:23 -!- phoenix24(xhohywp@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 14:50 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-25 15:22 <DavidSev> Hi.  I'm currently writing my GSoC app.  What level of detail should I go into?
2008-03-25 15:31 -!- phoenix24(gxp@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 17:09 -!- phoenix24(zpl@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 18:09 -!- phoenix24_(uxgdv@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 19:39 -!- phoenix24(fvh@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 19:49 <flips> davidsev, enough to show that your proposal is technically plausible and useful in some way
2008-03-25 20:41 <willn> Holy cow
2008-03-25 20:41 <willn> Kernel just went kabluey
2008-03-25 20:42 <willn> 'tty pty[$sequential] uevent: unsupported action-string; this will be ignored in a future kernel version'
2008-03-25 20:43 <willn> Constantly
2008-03-25 20:45 <flips> uevent...
2008-03-25 20:45 <flips> I've see that one
2008-03-25 20:45 <flips> forget where
2008-03-25 20:46 <flips> willn, Alt-SysRq-t
2008-03-25 20:46 <willn> http://lkml.org/lkml/2007/8/14/106
2008-03-25 20:47 <willn> yar, the system clock went bye bye too
2008-03-25 20:47 <flips> how do you know?
2008-03-25 20:47 <flips> timestamps?
2008-03-25 20:47 <willn> tar: lib/libnss_nisplus-2.7.so: time stamp 2007-12-07 21:02:46 is 34380996 s in the future
2008-03-25 20:48 <flips> let's get that number in hex
2008-03-25 20:48 <willn> ick
2008-03-25 20:48 <willn> init is confused aswell
2008-03-25 20:48 <flips> nothing interesting
2008-03-25 20:48 <willn>  * INIT: Id "3" respawning too fast: disabled for 5 minutes                     
2008-03-25 20:49 <willn> for all ids
2008-03-25 20:49 <flips> got the sysrq-t?
2008-03-25 20:49 <flips> which machine?
2008-03-25 20:49 <willn> im on a serial console for zuma6
2008-03-25 20:49 <flips> kind of machine?
2008-03-25 20:50 <flips> kernel on it?
2008-03-25 20:50 <willn> 2950, i think
2008-03-25 20:50 <willn> 2.6.24 that I just built
2008-03-25 20:50 <willn> But I think it may be libc woes
2008-03-25 20:51 <willn> since it was a dapper machine that I wedged a hardy kernel onto
2008-03-25 20:51 <flips> just say no to distro kernels
2008-03-25 20:51 <willn> eh?
2008-03-25 20:52 <flips> distro kernels tend to suck hard
2008-03-25 20:52 <flips> mainline is solid as a rock
2008-03-25 20:52 <flips> well
2008-03-25 20:52 <flips> except for storage apps ;)
2008-03-25 20:52 <flips> without patches
2008-03-25 20:52 <willn> Theres nothing special about this build, except its got all our crap built in
2008-03-25 20:52 <willn> It was just built the ubuntu way on a hardy machine
2008-03-25 20:52 <flips> and all the hardy crap
2008-03-25 20:53 <willn> Well, its this way, or no way
2008-03-25 20:54 -!- nataliep_(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-25 20:54 <willn> [  807.783909] call-dhclient-s[6597] general protection eip:b7ff8456 esp:bffffbac error:0                                                                       
2008-03-25 20:54 <phoenix24> I think, I'm writing an article on Zumastor :)
2008-03-25 20:54 <flips> for posting?
2008-03-25 20:55 <phoenix24> heh! or that my notes have.. grown too big
2008-03-25 20:55 <flips> would be pleased to read them
2008-03-25 20:55 <phoenix24> No, No.. not posting yet.
2008-03-25 20:56 <phoenix24> What are the other open source projects.. cousins to Zumastor ?
2008-03-25 21:13 <phoenix24> flips: is the git repository online, for some reason my checkout seems to fail everytime ?
2008-03-25 21:14 <willn> mkfs.ext3 takes forever on large volumes
2008-03-25 21:14 -!- phoenix24_(xvl@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 21:14 <flips> phoenix24, which repository?
2008-03-25 21:15 <phoenix24_> Git
2008-03-25 21:15 <flips> I mean, which url?
2008-03-25 21:16 <phoenix24_> at, git://phunq.net/dtree/zumastor
2008-03-25 21:16 <flips> could you show your git --version please
2008-03-25 21:16 <phoenix24_> git version 1.5.4.3
2008-03-25 21:59 <flips> hmm
2008-03-25 21:59 <flips> ok, wrong url
2008-03-25 21:59 <flips> git clone http://phunq.net/dtree/zumastor
2008-03-25 21:59 <flips> git: protocol is broken because git-daemon is broken, have not had time to fix it yet
2008-03-25 22:20 <willn> Well
2008-03-25 22:20 <willn> I got it working
2008-03-25 22:20 <willn> then it broke agian
2008-03-25 22:20 <willn> ill diagnose tommorow morning
2008-03-25 22:27 <flips> will stay tuned
2008-03-25 22:39 -!- phoenix24(hebiz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-25 22:47 <shapor> phoenix24 the google code subversion is the authoratative source for zumastor anyway
2008-03-25 22:48 <shapor> authoritative*
2008-03-25 22:48 <phoenix24> that would be good, but git's definitely a better alternative
2008-03-25 22:49 <shapor> you can use git-svn if you want to use git with svn repo
2008-03-25 22:50 <phoenix24> No, svn's good for now. /me is comfortable
2008-03-25 22:52 <flips> phoenix24, where you able to clone from the http url?
2008-03-25 22:52 <flips> let me check my server log :-)
2008-03-25 22:52 <flips> 
ACTION hates checking server logs

2008-03-25 22:53 <phoenix24> flips: nope, http fails too :)
2008-03-25 22:53 <flips> 220.225.53.35 - - [25/Mar/2008:22:53:14 -0700] "GET /dtree/zumastor//info/refs HTTP/1.0" 404 294 "-" "curl/7.18.0 (i486-pc-linux-gnu) libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.1"
2008-03-25 22:53 <phoenix24> **:(
2008-03-25 22:53 <phoenix24> that's me
2008-03-25 22:53 <flips> I'll try it locally
2008-03-25 22:54 <phoenix24> fails with an error stating "Perhaps git-update-server-info needs to be run on the server"
2008-03-25 22:54 <flips> phoenix24, fails here too
2008-03-25 22:54 <flips> sorry
2008-03-25 22:54 <flips> will find and fix the problem
2008-03-25 22:55 <flips> oh
2008-03-25 22:55 <flips> wrong url
2008-03-25 22:55 <flips> feh
2008-03-25 22:56 <flips> git clone http://phunq.net/git/zumastor
2008-03-25 22:56 <flips> I'm so sorry
2008-03-25 23:07 <flips> did we clone successfully now?
2008-03-25 23:08 <flips> $ time git clone http://phunq.net/git/zumastor
2008-03-25 23:08 <flips> Initialized empty Git repository in /home/daniel/tmp/zumastor/.git/
2008-03-25 23:08 <flips> Getting alternates list for http://phunq.net/git/zumastor
2008-03-25 23:08 <flips> Getting pack list for http://phunq.net/git/zumastor
2008-03-25 23:08 <flips> Getting index for pack 3a2b9c0cbee91dc78f1b162c897c9549959bfd2f
2008-03-25 23:08 <flips> Getting pack 3a2b9c0cbee91dc78f1b162c897c9549959bfd2f
2008-03-25 23:08 <flips>  which contains 763a99e4fc7d5dd172767e4818752435a4ce9cdd
2008-03-25 23:08 <flips> walk 763a99e4fc7d5dd172767e4818752435a4ce9cdd
2008-03-25 23:08 <flips> real    0m1.086s
2008-03-25 23:08 <flips> user    0m0.268s
2008-03-25 23:08 <flips> sys     0m0.072s
2008-03-25 23:09 <flips> clones in 1 sec
2008-03-25 23:10 <flips> zumastor@yzf:~/tmp$ time git clone http://phunq.net/git/zumastor
2008-03-25 23:10 <flips> got 69f78f2bb05ff21c01ddae305a29e8e7d434711a
2008-03-25 23:10 <flips> walk 69f78f2bb05ff21c01ddae305a29e8e7d434711a
2008-03-25 23:10 <flips> Getting alternates list for http://phunq.net/git/zumastor/
2008-03-25 23:10 <flips> got 366146d7e8564f00de6ff054e6b061fb811d6c17
2008-03-25 23:10 <flips> Getting pack list for http://phunq.net/git/zumastor/
2008-03-25 23:10 <flips> Getting index for pack 3a2b9c0cbee91dc78f1b162c897c9549959bfd2f
2008-03-25 23:10 <flips> got a450d024d2208f2a0ff4730485b3cb7c2533ff39
2008-03-25 23:10 <flips> Getting pack 3a2b9c0cbee91dc78f1b162c897c9549959bfd2f
2008-03-25 23:10 <flips>  which contains 763a99e4fc7d5dd172767e4818752435a4ce9cdd
2008-03-25 23:10 <flips> walk 763a99e4fc7d5dd172767e4818752435a4ce9cdd
2008-03-25 23:10 <flips> got 68a39328e2f2c8677e24a99956b4ddb5446472fc
2008-03-25 23:10 <flips> Checking files out...
2008-03-25 23:10 <flips>  100% (381/381) done
2008-03-25 23:10 <flips> real    0m26.896s
2008-03-25 23:10 <flips> user    0m0.300s
2008-03-25 23:10 <flips> sys     0m0.108s
2008-03-25 23:10 <flips> zumastor@yzf:~/tmp$ 
2008-03-25 23:10 <flips> clones in 26 seconds over the net
2008-03-25 23:11 <flips> not the most efficient method, but that is what git: is for
2008-03-25 23:12 <flips> 6 megs... that is 216K/sec
2008-03-25 23:12 <flips> probably the limit of my upload bandwidth
2008-03-25 23:15 <phoenix24> it's probable
2008-03-25 23:39 -!- phoenix24_(ddqr@h1119083.serverkompetenz.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Wed Mar 26 00:00:01 PDT 2008
2008-03-26 00:40 <flips> somebody wrote a very handy function called change_bits
2008-03-26 00:40 <flips> rlnelson?
2008-03-26 00:40 -!- phoenix24(twm@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-26 00:40 <flips> I'm thinking, thanks for that
2008-03-26 00:43 <flips> oh, I think it was jiayginz
2008-03-26 00:43 <flips> as part of the resize
2008-03-26 00:43 <flips> thanks jiayingz
2008-03-26 01:14 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-26 01:21 -!- flips(~phillips@phunq.net) has joined #zumastor
2008-03-26 01:39 -!- phoenix24_(mggppib@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-26 01:43 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-26 01:51 -!- pgquiles__(~pgquiles@121.Red-81-39-99.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 02:20 -!- pgquiles__(~pgquiles@51.Red-83-53-122.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 02:21 -!- pgquiles__(~pgquiles@51.Red-83-53-122.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 02:40 -!- phoenix24(ajwja@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-26 04:19 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-26 04:19 -!- phoenix24_(obk@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-26 04:19 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 04:50 -!- pgquiles__(~pgquiles@51.Red-83-53-122.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 06:26 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 09:04 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-26 09:21 <willn> 
ACTION gives the kernel another test

2008-03-26 09:35 <willn-colo> zumastor-commits: [zumastor commit] r1498 - trunk/ddsnap/scripts
2008-03-26 09:35 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: zumastor 0.8.0-r1498 (source)
2008-03-26 09:40 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: ddsnap 0.8.0-r1498 (source)
2008-03-26 09:41 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-26 09:59 -!- MaZe(~MaZe@193.120.148.177) has left #zumastor
2008-03-26 10:00 <willn-colo> ppa: QUEUED: amd64 build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:00 <willn-colo> ppa: QUEUED: i386 build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:00 <willn-colo> ppa: QUEUED: lpia build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:00 <willn-colo> ppa: QUEUED: amd64 build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:00 <willn-colo> ppa: QUEUED: i386 build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:00 <willn-colo> ppa: QUEUED: lpia build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:01 <willn> holy crap
2008-03-26 10:02 -!- pgquiles__(~pgquiles@51.Red-83-53-122.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 10:05 <willn-colo> ppa: FINISHED: lpia build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:05 <willn-colo> ppa: FINISHED: i386 build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:06 <willn-colo> ppa: FINISHED: amd64 build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:06 <willn-colo> ppa: FINISHED: lpia build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:06 <willn-colo> ppa: FINISHED: i386 build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE
2008-03-26 10:06 <willn-colo> ppa: FINISHED: amd64 build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE .. STARTED: lpia build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE .. STARTED: lpia build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE .. STARTED: amd64 build of ddsnap 0.8.0-r1498 in ubuntu hardy RELEASE .. STARTED: amd64 build of zumastor 0.8.0-r1498 in ubuntu hardy RELEASE .. STARTED: i386 build (+1)
2008-03-26 10:06 <willn> well, thats annoying.
2008-03-26 10:12 <DavidSev> Can you make it squish them all into one line?
2008-03-26 10:15 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: linux 2.6.24-12.22~zumappa1497 (source)
2008-03-26 10:17 <willn> It's just listening to the rss feed. In the end it looks like it tried
2008-03-26 10:17 <willn> I turned off autobuilds for ddsnap and zumastor, as we're not really commiting code that goes into those
2008-03-26 10:24 <willn> 
ACTION rewrites the launchpad watcher while waiting for the kernel build

2008-03-26 10:41 <willn-colo> ppa: QUEUED: amd64 build of linux 2.6.24-12.22~zumappa1497 in ubuntu hardy RELEASE
2008-03-26 10:41 <willn-colo> ppa: QUEUED: i386 build of linux 2.6.24-12.22~zumappa1497 in ubuntu hardy RELEASE
2008-03-26 10:41 <willn-colo> ppa: QUEUED: lpia build of linux 2.6.24-12.22~zumappa1497 in ubuntu hardy RELEASE
2008-03-26 10:47 -!- pgquiles_(~pgquiles@51.Red-83-53-122.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-26 11:45 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 11:46 -!- MaZe(~MaZe@193.120.148.177) has left #zumastor
2008-03-26 11:59 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 12:06 -!- MaZe(~MaZe@193.120.148.177) has left #zumastor
2008-03-26 12:42 -!- MaZ1(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 12:42 -!- MaZ1(~MaZe@193.120.148.177) has left #zumastor
2008-03-26 12:43 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-26 12:46 <willn> willn-colo: rss-delete ppa
2008-03-26 12:47 <willn> willn-colo: rss-add ppa http://compbrain.net:8080/rss/zumastor-team/
2008-03-26 12:47 <willn-colo> rss item added
2008-03-26 12:47 <willn> willn-colo: rss-watch ppa 60
2008-03-26 12:47 <willn-colo> watcher thread started
2008-03-26 13:10 <willn> grr. Google code doesn't provide rss feeds either.
2008-03-26 13:27 <willn> flips: Did you get a chance to patch issue 82
2008-03-26 13:49 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-26 14:21 <willn-colo> zumastor-commits: [zumastor commit] r1499 - trunk
2008-03-26 14:57 <willn-colo> ppa: FINISHED: amd64 build of linux 2.6.24-12.22~zumappa1497 in ubuntu hardy RELEASE
2008-03-26 15:08 <willn> hooray
2008-03-26 15:10 -!- cbsmith(~user@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-26 15:21 <flips> hmm, issue 82
2008-03-26 15:22 <cbsmith> flips: -H for hourly
2008-03-26 15:23 <flips> -h for hourly, --help for help
2008-03-26 15:23 <flips> or -?
2008-03-26 15:23 <flips> which is becoming predominant practice exactly because -h is too useful for other shortcuts
2008-03-26 15:24 <flips> who wants to hack the script?
2008-03-26 15:24 <flips> this is open to nongooglers by the way
2008-03-26 15:24 <flips> pgquiles_, ping?
2008-03-26 15:31 <pgquiles_> flips: pong
2008-03-26 15:31 <cbsmith> flips: -h for help, --help for help, -H and --hourly for hourly
2008-03-26 15:32 <cbsmith> Oh, and -? for help too
2008-03-26 15:32 <flips> cbsmith, no, -?|--help for help, -h|--hourly for hourly
2008-03-26 15:32 <flips> 
ACTION points pgquiles_ at query chat ==>

2008-03-26 15:32 <cbsmith> flips: Maybe I'm an idiot, but I *frequently* do "foo -h" when testing out a new command.
2008-03-26 15:32 <flips> cbsmith, retraining for you then
2008-03-26 15:38 <flips> cbsmith, note that zumastor define volume -h will produce usage text, even with -h -> hourly
2008-03-26 15:38 <cbsmith> flips: yeah, 'cause *that* won't be confusing. ;-)
2008-03-26 15:41 <flips> why does firebox always start doing weird things with more than 100 tabs open?
2008-03-26 15:41 <cbsmith> flips: I'm sure you are aware there are roughly a million reasons for that. Try firefox3beta4. It seems to be better at such things.
2008-03-26 15:42 <flips> thanks
2008-03-26 16:10 <flips> http://osindia.blogspot.com/2008/03/microsoft-files-complaint-on-ooxml-vote.html <- wow
2008-03-26 16:32 <willn-colo> ppa: FINISHED: i386 build of linux 2.6.24-12.22~zumappa1497 in ubuntu hardy RELEASE
2008-03-26 16:55 <willn> flips: I just did another test with 2.6.24
2008-03-26 16:55 <flips> and?
2008-03-26 16:55 <willn> same fail
2008-03-26 16:55 <willn> ill get you sysrq
2008-03-26 16:55 <flips> ok
2008-03-26 16:58 <flips> 
ACTION orders one of those green 750G sata twos

2008-03-26 16:58 <flips> going to take solid sate a while to catch up with stuff like that
2008-03-26 16:58 <willn> flips: http://compbrain.net/tmp/sofia-sysrq.txt?
2008-03-26 16:58 <flips> $150
2008-03-26 16:58 <willn> rougue question mark
2008-03-26 16:59 <flips> side by side bug hunt tomorrow?
2008-03-26 16:59 <flips> I'll ponder tonight
2008-03-26 16:59 <willn> I wont be in SMO tommorow
2008-03-26 16:59 <willn> Still in mountain view until monday
2008-03-26 16:59 <flips> right
2008-03-26 16:59 <flips> ok, well I will do what I can
2008-03-26 17:00 <flips> and then monday if the bug is still wiggling?
2008-03-26 17:00 <willn> Sure
2008-03-26 17:01 <flips> http://www.silentpcreview.com/article786-page1.html <- check it out
2008-03-26 17:01 <flips> picked mine up from newegg, price is lower than in the review
2008-03-26 17:02 <willn> its a much smaller trace from before
2008-03-26 17:02 <willn> doesnt look like any io is happening at any speed right now
2008-03-26 17:03 <willn> and mkfs isnt responding to signals
2008-03-26 17:03 <flips> I'll look through it
2008-03-26 17:04 <flips> oh wow, init is making a fuse call
2008-03-26 17:04 <flips> scary
2008-03-26 17:05 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-26 17:05 <flips> hey tim_vimm
2008-03-26 17:05 <tim_vimm> yo
2008-03-26 17:05 <tim_vimm> http://www.imagineersystems.com/
2008-03-26 17:06 <flips> wearing yo vimm 2day ic
2008-03-26 17:06 <tim_vimm> i'm having a vimmtastic day
2008-03-26 17:06 <tim_vimm> or...better yet, I'm feeling rather vimmy today
2008-03-26 17:06 <flips> niftey
2008-03-26 17:07 <flips> mplayer is playing it :-)
2008-03-26 17:07 <flips> 
ACTION remembers when that didn't work

2008-03-26 17:07 <tim_vimm> actually, that's the result of a user restore to another system
2008-03-26 17:07 <flips> did I see you playing with that at the apple store?
2008-03-26 17:08 <tim_vimm> no, someone just posted that on CML
2008-03-26 17:08 <tim_vimm> (cinematographers mailing list)
2008-03-26 17:08 <tim_vimm> my virtual haunt
2008-03-26 17:08 <flips> done that skydive think, complete with city in back
2008-03-26 17:09 <flips> for real, not edited :-)
2008-03-26 17:11 <willn> flips: Strace stops at fsync
2008-03-26 17:12 <tim_vimm> flips, remember the monsters and aliens show Scottie showed us at dreamworks?
2008-03-26 17:12 <flips> a little
2008-03-26 17:12 <tim_vimm> a buddy of mine is doing the 3D conversion of that film- he's the guy who posted the imagineeringsystems link
2008-03-26 17:13 <flips> 3d conversion... the future has arrived
2008-03-26 17:13 <flips> 
ACTION would like to have his apartment 3d converted

2008-03-26 17:13 <tim_vimm> actually, they rendered both eyes, Jeff does the final tweeks
2008-03-26 17:14 <flips> look at all those kblockd threads
2008-03-26 17:14 <flips> I suspect they are completely bogus
2008-03-26 17:14 -!- tim(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-26 17:15 <flips> willn, first thing I'd suggest trying: some other bio scheduler than cfq
2008-03-26 17:15 <tim_vimm> whoa, there's another me here
2008-03-26 17:15 <flips> freaky
2008-03-26 17:15 <tim_vimm> that's my old laptop (the one that will run linux very shortly)
2008-03-26 17:17 <flips> wow, udevd calling fuse
2008-03-26 17:17 <flips> somebody hold the barf bag for me please?
2008-03-26 17:19 <flips> I think those fuse: items in the traceback may be bogus
2008-03-26 17:21 <flips> willn, no smoking gun jumps out of that trace at me
2008-03-26 17:21 <willn> its still wedged
2008-03-26 17:21 <willn> 0 progress
2008-03-26 17:22 <flips> bio stats not changing?
2008-03-26 17:22 <willn> Where am I looking?
2008-03-26 17:23 <willn> the ddsnap interface?
2008-03-26 17:23 <flips> the monitoring utility that scraps sysfs whose name I always forget
2008-03-26 17:23 <flips> just a sec
2008-03-26 17:23 <willn> 'ddsnap inflight bio vecs 1000, pending: 0, query: 0, release: 0, locked: 0'
2008-03-26 17:24 <flips> not good
2008-03-26 17:25 <willn> 0 io on the system
2008-03-26 17:25 <willn> according to iostat
2008-03-26 17:27 <flips> right
2008-03-26 17:27 <willn> so, im stuck.
2008-03-26 17:27 <flips> I see something
2008-03-26 17:28 <flips> ddsnap_server is doing a slab alloc
2008-03-26 17:28 <willn> does it look like a patch? ;)
2008-03-26 17:28 <flips> sock aio write
2008-03-26 17:28 <flips> got it
2008-03-26 17:28 <flips> ok, I'm on it
2008-03-26 17:28 <flips> patch will take a few days, ok?
2008-03-26 17:29 <flips> thanks for the report, it's great
2008-03-26 17:29 <flips> :)
2008-03-26 17:29 <flips> known issue which you just voted up
2008-03-26 17:29 <willn> its blocking 2.6.24 then, though I guess we modded that down today?
2008-03-26 17:30 <willn> Suppose its time for me to get 2.6.22 builds working
2008-03-26 17:34 <flips> good plan
2008-03-26 17:34 <flips> we've just been lucky that 2.6.earlier doesn't trigger this pattern
2008-03-26 17:34 <flips> then somebody improved the vm ;-)
2008-03-26 17:35 <flips> scary amount of cruft in there now
2008-03-26 17:35 <flips> won't hack it all out this time though, just make socket write work properly
2008-03-26 17:36 <flips> see, this is why the world needs ddtree
2008-03-26 17:36 <flips> http://phunq.net/ddtree
2008-03-26 17:43 <flips> willn, can you paste or post the output of cat /proc/meminfo?
2008-03-26 17:48 <willn> I just killed the machine, but looked at it recently, what were you looking for?
2008-03-26 20:06 <willn-colo> zumastor-commits: [PPA zumastor-team] Accepted: linux-source-2.6.22 2.6.22-14.52~zumappa1 (source)
2008-03-26 20:09 <willn-colo> ppa: STARTED: lpia build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:10 <willn-colo> ppa: QUEUED: lpia build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:11 <willn-colo> ppa: STARTED: amd64 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:11 <willn-colo> ppa: STARTED: i386 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:11 <willn-colo> ppa: QUEUED: i386 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:11 <willn-colo> ppa: QUEUED: amd64 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 20:50 -!- MaZe(~MaZe@193.120.148.177) has left #zumastor
2008-03-26 20:58 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-26 21:06 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-26 21:26 <willn-colo> ppa: FINISHED: lpia build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-26 22:28 -!- phoenix24(tdbxu@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-26 22:28 <phoenix24> Hi everyone!
2008-03-26 22:29 <shapor> hi phoenix24
2008-03-26 22:31 <phoenix24> shapor: Do any applications of Neural Networks.. of Artificial Intelligence theory exist in the domain of Data-Storage, Data-Backup or Data-Protection ?
2008-03-26 22:31 <shapor> I don't know
2008-03-26 22:33 <shapor> googling seems to return some research papers
2008-03-26 22:33 <phoenix24> I have the AI exam today, so was wondering.. if it had applications here.
2008-03-26 22:33 <phoenix24> is it ?.. I'd check the papers later
2008-03-26 22:34 <phoenix24> shapor: what key-words you searched for ?
2008-03-26 22:34 <shapor> a patent even
2008-03-26 22:34 <shapor> "Data storage system with trained predictive cache management engine"
2008-03-26 22:35 <shapor> "neural network" storage system
2008-03-26 22:35 <phoenix24> wow!
2008-03-26 22:37 <shapor> just getting storage working and reliable is enough of a challenge for most
2008-03-26 22:38 <phoenix24> _reliability_ engineering itself is a domain there.
2008-03-26 22:38 <shapor> work seems more focused details and corner cases and simple fast algoritms
2008-03-26 22:38 <shapor> flips probably knows more
2008-03-26 22:42 <phoenix24> I should probably head back.. studying for the exam.
2008-03-26 23:43 <flips> phoenix24 (when you come back) oddly enough, the inventor of  device mapper is a neural net programmer
2008-03-26 23:47 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
irc.oftc.net #zumastor log beginning Thu Mar 27 00:00:01 PDT 2008
2008-03-27 00:07 <willn-colo> ppa: FINISHED: amd64 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-27 01:59 <willn-colo> ppa: FINISHED: i386 build of linux-source-2.6.22 2.6.22-14.52~zumappa1 in ubuntu gutsy RELEASE
2008-03-27 03:39 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-27 03:41 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-27 03:44 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-27 08:09 -!- erwan_taf(~erwan@81.80.43.67) has left #zumastor
2008-03-27 08:34 <willn> woot has $77 500G drives currently
2008-03-27 08:42 <willn> not any more.
2008-03-27 08:44 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 09:24 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-27 10:10 <dld> willn: how many sata cards/ports are already in the test boxes?
2008-03-27 10:13 <dld> 
ACTION thinks to go "upmarket" and get NCQ-capable cards and disks this time

2008-03-27 10:16 <willn> dld: One machine has one, one has 0
2008-03-27 10:16 <willn> the machine with one could possibly use another, but likely not. 
2008-03-27 10:16 <dld> And they have one onboard sata port (the 4100s)?
2008-03-27 10:16 <willn> (Each would need a max of two to support all disks)
2008-03-27 10:16 <willn> Yea
2008-03-27 10:16 <dld> The 4300 has multiple onboard ports?
2008-03-27 10:17 <willn> Though, if you make disk changes to the onboard controller it prompts for input at boot (stupid)
2008-03-27 10:17 <willn> They each have two onboard, but suffer from the idiot disk change notification
2008-03-27 10:17 <dld> I only saw one SIIG card in the 4300 in lspci.  Are there really two then?
2008-03-27 10:18 <dld> Or I could open the case..
2008-03-27 10:18 <willn> The board has two sata ports onboard
2008-03-27 10:18 <willn> the 4300 at your desk has one sata card + onboard
2008-03-27 10:19 <dld> Yep, all four onboard occupied by somethings.
2008-03-27 10:20 <dld> http://www.newegg.com/Product/Product.aspx?Item=N82E16816102062
2008-03-27 10:21 <dld> some slower GP drives: http://www.newegg.com/Product/Product.aspx?Item=N82E16822136149
2008-03-27 10:23 <dld> NCQ drives: http://www.amazon.com/Seagate-ST3500641AS-RK-Barracuda-Internal-Drive/dp/B000EY5F5C/ref=pd_bbs_9?ie=UTF8&s=electronics&qid=1206638534&sr=8-9
2008-03-27 10:23 <willn> You missed out on the $77 from woot this morning
2008-03-27 10:23 <dld> Yep.
2008-03-27 10:23 <willn> I was going to just buy them, but reimbursement and all
2008-03-27 10:24 <dld> Yeah, too hard.  I'll do them and just pay more so we have them next week.
2008-03-27 11:31 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 11:37 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-27 12:03 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 12:10 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-27 12:16 <willn> Anyone want to try out the 2.6.22 kernel we built overnight?
2008-03-27 12:26 <willn> well, it boots.
2008-03-27 12:26 <willn> 
ACTION throws another disk in to test ddsnap

2008-03-27 12:33 <willn> It works. yay :)
2008-03-27 14:13 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 14:22 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 14:33 <dld> Which 2.6.22 was that, PPA or CBTB?
2008-03-27 15:02 <willn> PPA
2008-03-27 15:03 <willn> amd64 and i386 built the ubuntu way
2008-03-27 15:16 <willn> If people want to test them out more, thatd be cool. I've qualified they boot and run zumastor. 
2008-03-27 16:15 <willn-colo> zumastor-commits: [zumastor commit] r1500 - trunk/doc
2008-03-27 16:55 <willn-colo> zumastor-commits: [zumastor commit] r1501 - trunk/doc
2008-03-27 17:00 <willn-colo> zumastor-commits: [zumastor commit] r1502 - branches/0.7/doc
2008-03-27 17:04 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-27 17:04 <dkegel> 0.7 released
2008-03-27 17:05 <shapor> w00
2008-03-27 17:05 <willn-colo> zumastor-commits: [zumastor commit] r1503 - www
2008-03-27 17:15 <willn-colo> zumastor-commits: [zumastor commit] r1504 - trunk/doc
2008-03-27 17:20 <willn> I sent out mail about the ppa kernels
2008-03-27 17:21 <willn> flips: ping re: 2.6.24 sysrq
2008-03-27 17:25 <willn-colo> zumastor-commits: [zumastor commit] r1505 - branches/0.7/doc
2008-03-27 17:58 -!- phoenix24(ouzmzkl@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-27 18:50 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 18:57 -!- phoenix24_(afmc@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-27 19:00 <phoenix24_> Hi everyone!
2008-03-27 19:24 -!- phoenix24(bxg@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-27 19:31 <jiayingz> hi phoenix24_
2008-03-27 19:32 <jiayingz> phoenix24_, have you submitted the application for soc?
2008-03-27 19:33 <phoenix24> I'm writing it now.
2008-03-27 19:33 <phoenix24> jiayingz: I'm sorry, I had exams till yesterday.
2008-03-27 19:34 <phoenix24> jiayingz: also, I see Zumastor also registered as a Launchpad Project.
2008-03-27 19:34 <phoenix24> I've created a test project@launchpad for myself to see if we could use it.
2008-03-27 19:34 <jiayingz> phoenix24, I think will registered that
2008-03-27 19:35 <jiayingz> phoenix24, I did not see Zumastor on Launchpad
2008-03-27 19:36 <phoenix24> jiayingz: https://launchpad.net/~zumastor-team
2008-03-27 19:37 <phoenix24> it's the zumastor-team registered.
2008-03-27 19:37 <jiayingz> ah, i c. I searched for zumastor, not zumastor-team
2008-03-27 19:38 <phoenix24> since I don't have write access to it, i'm upstreaming initial import of Zumastor-0.7 to my test-project@Launchpad.
2008-03-27 19:38 <phoenix24> I hope to see if it works fine.
2008-03-27 19:39 <jiayingz> great. we just announced 0.7 release today, good timing :)
2008-03-27 19:39 <jiayingz> let us know if you get any problems
2008-03-27 19:39 <phoenix24> sure!
2008-03-27 19:39 <phoenix24> is BZR a decent alternative to Git ?
2008-03-27 19:40 <jiayingz> what does BZR stands for?
2008-03-27 19:41 <phoenix24> bazaar, distributed repository used a Launchpad
2008-03-27 19:41 <jiayingz> interesting. never heard of it before
2008-03-27 19:41 <phoenix24> http://bazaar-vcs.org/
2008-03-27 19:42 <jiayingz> thx
2008-03-27 19:42 <jiayingz> sponsored by Canonical Ltd
2008-03-27 19:43 <jiayingz> ubuntu uses it
2008-03-27 19:43 <jiayingz> written in python
2008-03-27 19:43 <phoenix24> yes
2008-03-27 19:47 <phoenix24> jiayingz: under what licenses is Zumastor released ?
2008-03-27 19:57 <phoenix24> Zumastor-Test-Project @ Launchpad  : https://launchpad.net/myzuma/
2008-03-27 19:58 <phoenix24> Code hosted @ Launchpad  : https://code.launchpad.net/~csharma/myzuma/Zumastor.upstream
2008-03-27 20:01 <jiayingz> phoenix24, GPL2
2008-03-27 20:01 <phoenix24> thanks!
2008-03-27 20:14 <phoenix24> Zumastor @ Launchpad  :  https://launchpad.net/zumastor/
2008-03-27 20:15 <phoenix24> Zumastor @ Launchpad  :  https://code.launchpad.net/zumastor/
2008-03-27 21:14 -!- phoenix24(phksi@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-27 21:28 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 22:14 -!- phoenix24_(uknsrch@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-27 22:15 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-27 23:41 -!- phoenix24(florq@h1119083.serverkompetenz.net) has joined #zumastor
irc.oftc.net #zumastor log beginning Fri Mar 28 00:00:01 PDT 2008
2008-03-28 00:49 -!- phoenix24(wrak@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-28 02:39 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 03:06 <flips> turns out my new LG tv runs linux
2008-03-28 03:07 <flips> The last page of the manual is a big GPL notice for Linux and Busybox
2008-03-28 03:07 <flips> not sure why busybox needs special mention... there's libc etc etc too
2008-03-28 04:21 <pgquiles> flips: is it only a TV or a video recorder too? (I've seen many set top boxes with Linux but never a TV set)
2008-03-28 05:10 -!- phoenix24(uruce@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-28 05:11 <phoenix24> Hi! 
2008-03-28 05:12 <phoenix24> I've created a tiny patch, for purpose of cleanup.. if I could discuss it.
2008-03-28 05:32 <pgquiles> phoenix24: most people live in the US, they are sleeping now :-)
2008-03-28 05:32 <phoenix24> hmm :(
2008-03-28 05:33 <phoenix24> pgquiles: You'r still awake ?
2008-03-28 05:34 <pgquiles> phoenix24: I live in Spain
2008-03-28 05:35 <phoenix24> wow!
2008-03-28 05:35 <phoenix24> did you see my patch ?
2008-03-28 05:36 <phoenix24> I just sent thru, a while ago
2008-03-28 05:36 <pgquiles> no, where is it?
2008-03-28 05:36 <pgquiles> to the mailing list?
2008-03-28 05:37 <phoenix24> yes, to the mailing list.
2008-03-28 05:41 <pgquiles> what was the subject?
2008-03-28 05:42 <phoenix24> pgquiles, subject is "Patch - for "ddsnapd.h" 
2008-03-28 05:45 <DavidSev> The mailing list seems to have stopped mailing me...
2008-03-28 05:46 <pgquiles> phoenix24: I can't see that message
2008-03-28 06:10 -!- phoenix24_(jgubnij@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-28 06:13 <phoenix24_> the mailing list is going wierd
2008-03-28 06:13 <DavidSev> Try posting from the web page thing.
2008-03-28 06:14 <phoenix24_> done, it shows now.
2008-03-28 06:16 <DavidSev> Yeah, just landed in my inbox.
2008-03-28 06:22 <DavidSev> make: *** No rule to make target `ddsnapd.h', needed by `ddsnapd.o'.  Stop.
2008-03-28 06:22 <DavidSev> Your patch doesn't seem to contain anything for ddsnapd.h.
2008-03-28 06:22 <DavidSev> You probably need to svn add it before doing the diff.
2008-03-28 06:36 -!- xiuf(~f@58.38.200.216) has joined #zumastor
2008-03-28 07:01 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 08:05 -!- MaZe(~MaZe@193.120.148.177) has left #zumastor
2008-03-28 08:48 -!- xiuf(~f@58.38.200.216) has left #zumastor
2008-03-28 08:51 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 09:32 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 09:43 <willn> flips: ping
2008-03-28 09:54 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-28 10:18 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 10:45 <flips> willn, pong
2008-03-28 10:49 -!- pgquiles__(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 10:57 <willn> flips: Any luck with the sysrq output
2008-03-28 11:14 <flips> willn, yes, I found the problem
2008-03-28 11:14 <flips> wilnn, one thing: if you can get me the output of cat /proc/meminfo, that would be helpful
2008-03-28 11:30 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 12:27 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-28 12:38 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 12:38 <willn-colo> zumastor-commits: [zumastor commit] r1506 - trunk/zumastor/bin
2008-03-28 12:59 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 13:38 <cbsmith> Okay, just submitted the AIO patch to the list.
2008-03-28 13:42 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:20 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:22 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:23 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:31 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:36 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:37 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:40 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:42 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:46 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:52 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 14:56 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:00 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:05 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 15:06 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:14 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:15 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:17 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:20 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:21 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-28 15:21 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:32 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:34 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:35 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:40 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:42 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:45 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:52 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 15:56 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:00 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:01 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:02 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:10 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:11 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:17 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:22 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:23 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:25 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:29 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 16:32 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-28 21:42 -!- xiuf(~f@58.38.200.216) has joined #zumastor
irc.oftc.net #zumastor log beginning Sat Mar 29 00:00:01 PDT 2008
2008-03-29 00:57 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor
2008-03-29 01:00 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor
2008-03-29 02:42 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 02:47 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor
2008-03-29 02:58 -!- pgquiles__(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 03:40 -!- flips(~phillips@phunq.net) has joined #zumastor
2008-03-29 04:36 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 06:17 -!- DavidSev(david@lolcat.mercenariesguild.net) has joined #zumastor
2008-03-29 06:40 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 08:15 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 08:25 -!- xiuf(~f@58.38.200.216) has joined #zumastor
2008-03-29 08:32 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-29 09:50 -!- pgquiles_(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 12:40 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 12:54 -!- charlesnw(~charles@206.170.103.124) has joined #zumastor
2008-03-29 13:13 -!- willn(~wan@pinball.ccs.neu.edu) has joined #zumastor
2008-03-29 14:24 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 14:31 -!- charlesnw(~charles@adsl-75-50-162-175.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-29 14:31 -!- charlesnw(~charles@adsl-75-50-162-175.dsl.lsan03.sbcglobal.net) has left #zumastor
2008-03-29 16:16 -!- pgquiles(~pgquiles@150.128.40.40) has joined #zumastor
2008-03-29 17:10 <willn> autotest is a PITA
2008-03-29 21:25 <flips> willn, ping
2008-03-29 21:25 <flips> or shapor
2008-03-29 21:25 <flips> what kind of usb wireless works reasonbly?
2008-03-29 21:40 <willn> pardon?
2008-03-29 22:09 -!- xiuf(~f@58.38.200.216) has joined #zumastor
2008-03-29 23:34 <flips> willn, I need some kind of wireless lan that isn't a pain to configure, i.e., no madwifi
2008-03-29 23:36 <flips> 802.11b interface that is not pci
2008-03-29 23:36 <flips> or that is pci and have an in-tree driver
2008-03-29 23:52 <willn> b?
2008-03-29 23:59 <flips> or g
irc.oftc.net #zumastor log beginning Sun Mar 30 00:00:01 PDT 2008
2008-03-30 00:08 <cbsmith> flips: if you can still find a prism54 based adapter, that is always an option.
2008-03-30 00:17 <cbsmith> ick, actually, it looks like these days all the usb ones use SoftMAC firmware, which is the prism54 equivalent to madwifi. :-(
2008-03-30 00:40 -!- balajirrao(~balaji@219.64.72.125) has joined #zumastor
2008-03-30 00:48 <balajirrao> Hello all, I am a student wishing to contribute to zumastor ' through Google SoC. I am particularly interested in 'Remove 64 snapshot limit'
2008-03-30 00:56 -!- xiuf(~f@58.38.200.216) has joined #zumastor
2008-03-30 03:33 <flips> hi balajirrao
2008-03-30 03:34 <flips> let's have a chat later today, it is early in the morning here
2008-03-30 03:41 -!- pratikpattani(HydraIRC@203.187.237.222) has joined #zumastor
2008-03-30 03:43 <pratikpattani> hi..any mentor there for GSOC..??
2008-03-30 03:45 <pratikpattani> guess timezone is the prob..no prob..will come up later..had some queries..
2008-03-30 03:51 <flips> there is zumastor@googlegroups.com too, all queries welcome
2008-03-30 04:23 <balajirrao> hi flips! oh! that's okk. But i prefer we have a chat bode I apppy.
2008-03-30 06:43 -!- balajirrao(~balaji@219.64.72.125) has left #zumastor
2008-03-30 09:23 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-30 09:59 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-30 10:53 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-30 11:14 <willn> 
ACTION blinks

2008-03-30 14:18 -!- pgquiles(~pgquiles@239.Red-83-35-112.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-30 14:39 -!- cbsmith(~xman@adsl-76-243-131-38.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-30 14:43 <willn-colo> zumastor-commits: [zumastor commit] r1507 - www
2008-03-30 16:21 <flips> mercurial is indeed considerably smoother than git
2008-03-30 16:21 <flips> same commands to create a new repo: cd sometree; hg init; hg add .; hgcommit
2008-03-30 16:22 <flips> same speed it seems
2008-03-30 16:22 <flips> good demo of python capability
2008-03-30 16:22 <flips> and brilliant coding by Matt
2008-03-30 16:22 <flips> everybody should study mercurial I think
2008-03-30 16:30 -!- cbsmith(~xman@adsl-76-243-131-38.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-30 16:36 <flips> http://phunq.net/hgzuma <- smokin
2008-03-30 16:38 <willn> gpxe is very cool, but their docs are severly lacking
2008-03-30 16:38 <willn> and some parts of their product are quite broken
2008-03-30 16:55 -!- cbsmith(~xman@adsl-76-243-131-38.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-30 17:00 -!- xiuf(~f@58.38.200.216) has joined #zumastor
2008-03-30 17:08 <willn> epic fail, again
2008-03-30 17:18 -!- xiuf(~f@58.38.200.216) has left #zumastor
2008-03-30 17:41 <flips> willn, I will work with you on it on monday
2008-03-30 17:42 <flips> what is the g in gpxe?
2008-03-30 17:47 <flips> I hope it does not mean "grub"
2008-03-30 18:00 <willn> its the new spawn of etherboot
2008-03-30 18:03 <willn> I'm having a conversation with myself in their irc channel
2008-03-30 18:38 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has left #zumastor
2008-03-30 18:45 -!- charlesnw(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-30 18:51 <willn> hpa popped in to assist
2008-03-30 18:51 <willn> 
ACTION feels much better

2008-03-30 19:21 -!- charlesn1(~charles@cpe-75-84-67-98.socal.res.rr.com) has joined #zumastor
2008-03-30 19:37 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-30 20:21 <willn> ok, that crisis is now fixed
2008-03-30 20:22 <willn> now on to integer errors
2008-03-30 20:51 -!- charlesn1(~charles@cpe-75-84-67-98.socal.res.rr.com) has left #zumastor
2008-03-30 21:11 <willn> Its probably time to go home. 
2008-03-30 21:30 <willn> ok, for real this time. See everyone tommorow
2008-03-30 21:50 <flips> hpa? leet
2008-03-30 21:52 <willn> issues with chunked http transfers in gpxe and gpxelinux
irc.oftc.net #zumastor log beginning Mon Mar 31 00:00:01 PDT 2008
2008-03-31 03:24 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-31 09:42 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-31 10:18 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-31 10:40 <willn> ETOOMANYCHANNELS
2008-03-31 10:40 <willn> 4 networks, ~6 channels/network
2008-03-31 11:26 <flips> only 24?
2008-03-31 11:26 <flips> ;-)
2008-03-31 11:29 <willn> i'm on at least 12 channels on freenode, so that average is a bit wrong
2008-03-31 11:30 <willn> In other news, the installer now works really well
2008-03-31 11:30 <willn> I need one more null modem cable, plus accounting for whatever amd64 hardware I can get, and we're fully operational
2008-03-31 11:30 <flips> the miracle that is modern communications
2008-03-31 11:31 <flips> that is great to hear
2008-03-31 11:31 <willn> autotest and gpxe, cool projects that need better documentation
2008-03-31 11:32 <flips> I was looking at autotest, the scheduling aspect is interesting
2008-03-31 11:33 <willn> They've got many things figured out that I got wrong
2008-03-31 11:34 <flips> maybe we should try it
2008-03-31 11:34 <willn> we (i) are
2008-03-31 11:34 <flips> :-)
2008-03-31 11:34 -!- pgquiles(~pgquiles@226.Red-88-16-39.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-31 11:35 <willn> I axed most of what I wrote earlier this year, and autotest replaces a bunch
2008-03-31 11:39 -!- pgquiles(~pgquiles@226.Red-88-16-39.dynamicIP.rima-tde.net) has joined #zumastor
2008-03-31 11:46 <willn> Oh debian, why do you hate me. Ubuntu has been installing fine with apt-cacher, but debian blows up with 'corrupt' packages
2008-03-31 12:21 -!- MaZe(~MaZe@193.120.148.177) has joined #zumastor
2008-03-31 12:39 -!- pgquiles(~pgquiles@226.Red-88-22-55.staticIP.rima-tde.net) has joined #zumastor
2008-03-31 14:04 <pgquiles> willn: what packages is apt complaining about?
2008-03-31 14:05 <willn> a bunch. I blame apt-cacher
2008-03-31 14:05 <pgquiles> willn: there is a well-known bug in Soyuz (the PPA codebase) which generates code with some wrong headers
2008-03-31 14:05 <willn> These arent ppa issues, their debian base package issues
2008-03-31 14:05 <pgquiles> oh, ok
2008-03-31 14:05 <pgquiles> not PPA-generated
2008-03-31 14:08 <willn> yep
2008-03-31 14:33 -!- natalie(~nataliep@207.47.98.129.static.nextweb.net) has joined #zumastor
2008-03-31 15:30 <flips> hi natalie
2008-03-31 15:41 <natalie> flips: hi
2008-03-31 15:41 <flips> what's up in bugland?
2008-03-31 15:41 <natalie> heh it is going...
2008-03-31 15:41 <willn> flips: ping 2.6.24
2008-03-31 15:42 <natalie> i think im close to making summaries again
2008-03-31 15:42 <flips> would be much appreciated I think
2008-03-31 15:42 <flips> natalie, check this out: http://phunq.net/hgzuma
2008-03-31 15:42 <flips> compare to: http://phunq.net/ddtree
2008-03-31 15:43 <flips> first is a mercurial, second is a git repo
2008-03-31 15:43 <flips> web interfaces are quite different
2008-03-31 15:43 <MaZe> /who natalie
2008-03-31 15:43 <MaZe> /lol
2008-03-31 15:43 <willn> oops.
2008-03-31 15:43 <natalie> haha
2008-03-31 15:43 <flips> maze, this is natalie the linux kernel bug manager
2008-03-31 15:43 <natalie> better than typing my password as i did once
2008-03-31 15:43 <flips> heh
2008-03-31 15:44 <natalie> what is the first one dan
2008-03-31 15:44 <natalie> is it new version control?
2008-03-31 15:44 <flips> so, in my not too informed opinion after having used both at a superficial level, mercurial is a better base to build on than git
2008-03-31 15:44 <flips> only reason for using git for me is to communicate with other kernel devs
2008-03-31 15:44 <natalie> is this what mercurial looks like?
2008-03-31 15:45 <flips> that's just a demo web interface
2008-03-31 15:45 <flips> mercurial commands look much like git commands, except simpler
2008-03-31 15:45 <natalie> ah, oh
2008-03-31 15:45 <natalie> ok
2008-03-31 15:45 <flips> better designed in my opinion
2008-03-31 15:45 <flips> just as fast as git
2008-03-31 15:45 <MaZe> apparently, though a lot of projects are switching to git...
2008-03-31 15:45 <natalie> does it let you do patch versioning
2008-03-31 15:46 <flips> took me 15 minutes from the apt-get install mercurial to having a zumastor repository created and on the web
2008-03-31 15:46 <flips> no, it does not do patch versioning
2008-03-31 15:46 <flips> only darcs does
2008-03-31 15:46 <flips> which is going to v2.0 pretty soon
2008-03-31 15:46 <natalie> so it is simpler and faster you said
2008-03-31 15:46 <flips> I have not tried it since v0.6 or so
2008-03-31 15:46 <flips> mercurial is simpler and just as fast as git
2008-03-31 15:47 <flips> and _much_ less code
2008-03-31 15:47 <flips> mercurial is written in python
2008-03-31 15:47 <natalie> i tink its like bsd vs linux :)
2008-03-31 15:47 <flips> something like that
2008-03-31 15:47 <natalie> (very controvercial thing to say ;)
2008-03-31 15:47 <flips> both are excellent by the way
2008-03-31 15:47 <flips> just, mercurial wins in total goodness according to me
2008-03-31 15:47 <natalie> just many people are working on linux
2008-03-31 15:47 <natalie> yes
2008-03-31 15:47 <flips> mind you I have not done a lot of merging with either yet
2008-03-31 15:48 <flips> I will opine again after I have
2008-03-31 15:48 <flips> for now I will only maintain the git zumastor repo
2008-03-31 15:48 <flips> since my workflow is already organized around that, and I want to try the git merging
2008-03-31 15:48 <willn> flips: Just remember, linus says if you use anything but git your ugly and stupid
2008-03-31 15:48 <flips> :-)
2008-03-31 15:48 <MaZe> I believe the problem is linux has a lot of developers pending on a few people doing management at the core - thus ultimately, what's important is the throughput of the core (ie. linus, andrew morton, etc), and thus merge-speed is of paramount importance...
2008-03-31 15:48 <willn> www.youtube.com/watch?v=4XpnKHJAok8
2008-03-31 15:49 <flips> exaggeration, he only says that about subversion, and did not say ugly
2008-03-31 15:49 <willn> he did
2008-03-31 15:49 <flips> oh
2008-03-31 15:49 <natalie> i think this means mercusial more elegant at the least (knowing dan :)
2008-03-31 15:49 <natalie> mercurial that is
2008-03-31 15:49 <flips> well I am ugly, so perhaps I should stick with svn
2008-03-31 15:49 <MaZe> heh
2008-03-31 15:49 <flips> mercurial is far and away more elegant
2008-03-31 15:50 <flips> no "packs" for one thing <- what everybody hates about git
2008-03-31 15:50 <MaZe> elegant in internal design or elegant in user interface?
2008-03-31 15:50 <flips> in both
2008-03-31 15:50 <MaZe> git can be run without packs though can't it?
2008-03-31 15:51 <flips> mercurial does not have the bogus "index" interface concept
2008-03-31 15:51 <flips> maze, not if you want to be able to pull/clone from it efficiently
2008-03-31 15:51 <willn-colo> zumastor-commits: [zumastor commit] r1508 - trunk/cbtb/uml
2008-03-31 15:51 <flips> mercurial supports efficient cloning and pulling without repacking
2008-03-31 15:52 <flips> and, see "both are excellent" above
2008-03-31 15:52 <flips> mercurial is just more excellent
2008-03-31 15:52 <flips> oh, and mercurial was written by one person in less time than git was written by a good sized team
2008-03-31 15:52 <flips> and is better documented
2008-03-31 15:53 <flips> and doesn't have backward compatibility problems like git does
2008-03-31 15:54 <flips> or strange behaviour when doing something like git add .gitignore, which quietly fails in git
2008-03-31 15:54 <flips> by design
2008-03-31 15:55 <natalie> backward compatibility? with other systems you mean?
2008-03-31 15:55 <flips> backward compatibility with itself
2008-03-31 15:55 <natalie> haha is it really
2008-03-31 15:55 <flips> really
2008-03-31 15:55 <flips> new git cannot create a repo that can be read by an old git
2008-03-31 15:56 <flips> or rather, cloned/pulled by an old git
2008-03-31 15:56 <flips> most other stuff I tried works
2008-03-31 15:56 <flips> unfortunately, old git is what ships in debian stable
2008-03-31 15:56 <flips> and its not very old
2008-03-31 15:56 <flips> 1.4.4.4 vs 1.5.something
2008-03-31 15:57 <natalie> probably left as not very many people were using it in the past
2008-03-31 15:57 <flips> shapor, ping?
2008-03-31 15:57 <shapor> flips pong
2008-03-31 15:57 <flips> how about we do something nice for maze today?
2008-03-31 15:58 <MaZe> ?
2008-03-31 15:58 <flips> anything standing in the way?
2008-03-31 15:58 <flips> maze, like give you a zumastor volume to play with
2008-03-31 15:58 <MaZe> hmm, I'm in Dublin ;-)
2008-03-31 15:58 <MaZe> lol
2008-03-31 15:58 <flips> when back?
2008-03-31 15:58 <MaZe> I'll be back in mtv in two weeks
2008-03-31 15:58 <flips> but we can give it do you even where you are
2008-03-31 15:58 <flips> you can poke at it from there
2008-03-31 15:58 <MaZe> oh, cool, that'd be nice
2008-03-31 15:59 <MaZe> I've actually been clearing out my homedir (and laptop) today since it got fullish
2008-03-31 15:59 <flips> heh
2008-03-31 15:59 <flips> we can fix :-)
2008-03-31 15:59 <MaZe> [otoh I'm already way down on utilization, deleter a few binaries here and there and presto - lots of free space]
2008-03-31 15:59 <flips> yep
2008-03-31 15:59 <flips> usually that way
2008-03-31 16:00 <flips> eventually is does get crowded though
2008-03-31 16:00 <flips> and you have to start thinking about throwing out good stuff
2008-03-31 16:00 <MaZe> I never seem to run out of cruft to throw out
2008-03-31 16:00 <MaZe> I think I have 5 full os'es installed on my laptop for example
2008-03-31 16:01 <flips> to tell the truth, I have never come close either
2008-03-31 16:01 <MaZe> trying to reduce down to 3 but it's tough going...
2008-03-31 16:01 <flips> well
2008-03-31 16:01 <flips> I do fill up laptops
2008-03-31 16:01 <flips> few years of lkml is a good start
2008-03-31 16:01 <MaZe> (mostly because I can't figure out what was my files, what was part of the base os install, anybody know the equivalent of rpm -Va for deb?)
2008-03-31 16:02 <flips> a few dozen kernel tree variants sucks space like you would not believe, especially if they are all built with debug symbols
2008-03-31 16:02 <flips> it's been so long since I used rpm I forget what Va does
2008-03-31 16:02 <flips> don't have a rpm machine to try it on either
2008-03-31 16:02 <flips> probalby should fix that
2008-03-31 16:07 <MaZe> it verifies disk to database - basically a big sha1sum -c
2008-03-31 16:07 <MaZe> and yes, I know how much kernels take up
2008-03-31 16:07 <MaZe> had to lvextend a few days ago in order to build a kernel...
2008-03-31 16:08 <MaZe> (I think I needed 4G of disk space for a build)
2008-03-31 16:11 <willn-colo> zumastor-commits: [zumastor commit] r1509 - trunk/cbtb/host-scripts
2008-03-31 16:29 <flips> maze, debian does not treat the actual files on your system as authoritative because it relies on package depedencies instead of file dependencies
2008-03-31 16:29 <flips> so I do not know of any feature to go out and verify files against the package database
2008-03-31 16:29 <MaZe> oh, I don't care about authoritative, I want to know which files are the same as the packages I installed so that I can delete them ;-)
2008-03-31 16:29 <MaZe> be left with my changes
2008-03-31 16:29 <flips> right
2008-03-31 16:30 <flips> oddly, I've never heard of anyone doing that on debian
2008-03-31 16:30 <flips> but surely somebody has
2008-03-31 16:30 <flips> irc.freenode.net #debian would be useful
2008-03-31 16:30 <MaZe> so rpm -Vav basically lists any files that have wrong checksum, owner, group, permissions, selinux permissions, timestamp, etc... and what differs
2008-03-31 16:30 <MaZe> good point ;-)
2008-03-31 16:31 <flips> if it doesn't exist, it sounds like it would be a nice SoC project
2008-03-31 16:32 <flips> another nice SoC project would be "improve dpkg so it doesn't waste time and CPU doing linear scans of text files all the time"
2008-03-31 16:32 <flips> next summer maybe
2008-03-31 16:33 <MaZe> (in this case the important part being the extra -v also lists files which are OK in all those respects) - true, but hopefully it's already been done...
2008-03-31 16:35 <flips> maze, an off the wall idea... suppose you we able to manage your installs with a nice, scalable vcs like mercurial
2008-03-31 16:35 <flips> ...how would that work, in detail?
2008-03-31 16:36 <flips> would it not rock to have a log of everything you ever did to you system, and be able to run it backwards etc?
2008-03-31 16:38 -!- cbsmith(~xman@64.148.65.13) has joined #zumastor
2008-03-31 16:38 <dld> MaZE: aptitude install cruft.  /usr/sbin/cruft will print out a report on every file that isn't part of a Debian package.  World-class not recommended from Goobuntu, which is full of cruft.
2008-03-31 16:38 <flips> oh right
2008-03-31 16:38 <flips> forgot about cruft
2008-03-31 16:38 <flips> but that generates a _lot_ of false positives
2008-03-31 16:38 <flips> last time I tried
2008-03-31 16:39 <MaZe> hmm, now if only dpkg actually worked on this system...
2008-03-31 16:39 <flips> ooh
2008-03-31 16:39 <MaZe> they suggested to use debsums
2008-03-31 16:39 <flips> when dpkg stops working on a deb system it is time to reinstall (nearly)
2008-03-31 16:39 <MaZe> which of course isn't installed ;-)
2008-03-31 16:40 <MaZe> well, problem is, I'm trying to uninstall the system ;-)
2008-03-31 16:40 <flips> get somebody to email it to you from a similar machine ;)
2008-03-31 16:41 <flips> debsums installs with some very wierd failures
2008-03-31 16:41 <flips> oh, it is because I moved cwd
2008-03-31 16:41 <dld> The files in each package are listed in ...
2008-03-31 16:42 <dld> /var/lib/dpkg/info/*.list
2008-03-31 16:42 <flips> maze, want to try my debsums binary?
2008-03-31 16:43 <dld> rm all the files listed in those, and you should have a much more compact system with real changes in many of the remaining files.
2008-03-31 16:43 <dld> (I'd make a copy before doing that purge..)
2008-03-31 16:44 <MaZe> nah, I seem to be missing libraries as well
2008-03-31 16:44 <MaZe> looks like I must have done a preliminary deletion already
2008-03-31 16:45 <MaZe> I'll take a look at the *.list files
2008-03-31 16:45 <flips> already email the binary
2008-03-31 16:45 <MaZe> thanks
2008-03-31 16:47 <flips> strange...
2008-03-31 16:47 <flips> # ldd `which debsums`
2008-03-31 16:47 <flips> /usr/bin/ldd: line 161: /lib64/ld-linux-x86-64.so.2: cannot execute binary file
2008-03-31 16:47 <flips>         not a dynamic executable
2008-03-31 16:47 <flips> I wonder what it's up to?
2008-03-31 16:47 <flips> this it not a 64 bit system
2008-03-31 16:54 <flips> man, git is really pleasant to use compared to svn. 
2008-03-31 17:03 <cbsmith> flips: If ((!cvs) && (!rcs)) { return better_than_svn; }
2008-03-31 17:03 <cbsmith> flips: And I'm not even 100% sure about the CVS bit. ;-)
2008-03-31 17:05 <flips> cbsmith, reasonable algorithm, though perhaps it would be better implemented in python?
2008-03-31 17:05 <flips> svn is better than cvs according to me
2008-03-31 17:05 <flips> cvs was/is brutal
2008-03-31 17:05 <flips> redhat still uses it on their public code site
2008-03-31 17:06 <flips> unable to get concensus on what to move to
2008-03-31 17:06 <cbsmith> flips: lol
2008-03-31 17:06 <cbsmith> flips: Consensus is clear: git... no wait, mercurial.... no wait, bazaar!... no wait, darcs! ;-)
2008-03-31 17:07 <flips> "any of the above" [check here]
2008-03-31 17:08 <flips> mercurial would be a truly enlightened choice, git pragmantic, svn pedestrian, all of the above wise, but do nothing: stupid
2008-03-31 17:08 <flips> and ugly
2008-03-31 17:08 <flips> :)
2008-03-31 17:08 <cbsmith> 
ACTION feels darcs is the truly enlightened choice

2008-03-31 17:08 <flips> see if it scales first
2008-03-31 17:08 <flips> it did not last time I tried
2008-03-31 17:10 <cbsmith> meh, they've made huge performance improvements with it, and it continues to seem easy to address scalability issues as they encounter them. Technically, it should be able to scale better than git and be competitive with the others. The rest is just implementation details, and its open source so... ;-)
2008-03-31 17:24 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor
2008-03-31 17:26 <flips> cbsmit, then why not check a kernel tree into it and see how it works?
2008-03-31 17:33 <cbsmith> flips: That might be a fun thing to do as an experiment.
2008-03-31 17:51 <dank> xman, how's life in libaio land?
2008-03-31 17:52 <cbsmith> dank: I've not been able to reproduce the multi_write bug in a UP world, so I'm pretty confident about the nature of the bug. I'm putting together a small test case to send to the aio list.
2008-03-31 17:53 <cbsmith> I'm thinking we can kind of avoid the whole issue though by just using eventfd and using a single thread.
2008-03-31 17:53 <cbsmith> But first I want to do the equivalent of the test case using eventfd to prove to myself that the problem doesn't happen that way.
2008-03-31 17:54 <cbsmith> The next logical step is to throw away all the paio work I did and put together a patch adding aio for the journal write in ddsnapd.
2008-03-31 17:54 <cbsmith> ..using eventfd.
2008-03-31 18:09 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-31 18:10 <dank> Sounds good to me.
2008-03-31 18:19 -!- phoenix24(kur@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-31 18:26 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-31 19:07 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-31 19:11 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-31 19:12 -!- phoenix24(dihdz@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-31 19:29 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-31 20:12 -!- phoenix24(ancrxso@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-31 20:16 <willn> http://code.google.com/p/zumastor/issues/detail?id=96
2008-03-31 20:28 <willn> flips: ^
2008-03-31 20:28 <willn> jiayingz: ^ (fyi too)
2008-03-31 20:32 <jiayingz> hi willn
2008-03-31 20:32 <jiayingz> willn, thanks!
2008-03-31 20:43 -!- cbsmith(~xman@adsl-99-165-23-73.dsl.lsan03.sbcglobal.net) has joined #zumastor
2008-03-31 21:31 -!- phoenix24(pdd@h1119083.serverkompetenz.net) has joined #zumastor
2008-03-31 22:19 -!- tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor
2008-03-31 23:36 -!- phoenix24(czux@h1119083.serverkompetenz.net) has joined #zumastor