irc.oftc.net #zumastor log beginning Thu Nov 1 00:00:01 PDT 2007 2007-11-01 10:48 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-01 10:49 hi flipz 2007-11-01 12:34 -!- cbsmith(~user@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-01 18:50 -!- otoole(~programad@ns01.megasimples.com.br) has joined #zumastor irc.oftc.net #zumastor log beginning Fri Nov 2 00:00:02 PDT 2007 2007-11-02 02:55 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-02 09:00 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-02 11:08 -!- liona29(~liona29@ANantes-257-1-53-33.w90-25.abo.wanadoo.fr) has joined #zumastor irc.oftc.net #zumastor log beginning Sat Nov 3 00:00:01 PDT 2007 2007-11-03 18:40 -!- peter_(~peter@p57B5D388.dip.t-dialin.net) has joined #zumastor 2007-11-03 18:41 wo ist oma 2007-11-03 18:41 und wilhelm irc.oftc.net #zumastor log beginning Sun Nov 4 00:00:02 PDT 2007 2007-11-04 20:13 quiet weekend 2007-11-04 21:14 thank goodness ;-) irc.oftc.net #zumastor log beginning Mon Nov 5 00:00:01 PST 2007 2007-11-05 09:35 -!- jdries3(~jdries3@72.14.228.89) has joined #zumastor 2007-11-05 09:38 -!- jdries3(~jdries3@72.14.228.89) has left #zumastor 2007-11-05 10:56 goog morning shapor 2007-11-05 11:14 hi flipz 2007-11-05 11:14 goog morning to you too :) 2007-11-05 11:14 and a very goog morning it is too 2007-11-05 11:15 hmm, dana is pestering me to go play oblivion 2007-11-05 11:15 let me see 2007-11-05 11:15 how about im? 2007-11-05 11:16 shapor, can you buddy realdaniel@gmail.com? 2007-11-05 11:20 you are already 2007-11-05 11:24 hmm, I wonder why you don't show in my gaim buddy list 2007-11-05 11:24 oh wait... 2007-11-05 11:25 there you are under "aim bots" 2007-11-05 11:25 I wonder what an aim bot is supposed to be 2007-11-05 12:11 lol 2007-11-05 13:04 ACTION is an aimbot 2007-11-05 14:41 -!- sevz(~Mirc@38.99.101.131) has joined #zumastor 2007-11-05 14:41 -!- sevz(~Mirc@38.99.101.131) has left #zumastor 2007-11-05 16:58 -!- rafiu(~rafiu@84.12.112.162) has joined #zumastor 2007-11-05 16:58 ping ... 2007-11-05 17:32 -!- rafiu(~rafiu@84.12.112.162) has left #zumastor 2007-11-05 17:35 pong. 2007-11-05 17:35 error: user not found irc.oftc.net #zumastor log beginning Tue Nov 6 00:00:01 PST 2007 2007-11-06 00:48 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-06 02:00 is it possible to use the snapshots feature of zumastor to emulate a versioning filesystem? (with a good performance, I mean) 2007-11-06 05:37 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor 2007-11-06 05:47 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor 2007-11-06 06:44 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-06 09:33 -!- MaZe(~MaZe@216.239.45.19) has joined #zumastor 2007-11-06 11:38 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-06 11:49 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-06 15:35 -!- lola22(~lola22@ANantes-257-1-120-128.w90-25.abo.wanadoo.fr) has joined #zumastor 2007-11-06 22:51 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor irc.oftc.net #zumastor log beginning Wed Nov 7 00:00:01 PST 2007 2007-11-07 06:20 -!- zenTourist(~dwanek@165-234-90-3.state.nd.us) has joined #zumastor 2007-11-07 06:33 -!- zenTourist(~dwanek@165-234-90-3.state.nd.us) has left #zumastor 2007-11-07 09:26 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-07 09:45 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-07 10:16 hi flipz 2007-11-07 11:05 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-07 11:09 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-07 11:44 hi jiayingz 2007-11-07 11:49 I just found the checksum error I got yesterday was caused by a mistake in my test script 2007-11-07 11:50 I gave the snapshot device as the origin device when I started 'ddsnap server' 2007-11-07 11:51 it is surprising that I got this kind of error instead of more serious ones 2007-11-07 11:51 anyway, looks like blockdev works and we can get rid of 'dmsetup suspend' now 2007-11-07 12:08 nice 2007-11-07 13:32 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-07 15:16 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-07 15:24 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-07 15:36 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-07 16:12 now installing r850 on a vmware dapper on my laptop... 2007-11-07 16:40 svn diff --diff-cmd diff -x -rup 2007-11-07 16:40 flips: ^ irc.oftc.net #zumastor log beginning Thu Nov 8 00:00:02 PST 2007 2007-11-08 00:00 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-08 10:32 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-08 11:36 -!- juuva_(juuva@peili.org) has joined #zumastor 2007-11-08 11:55 hi flips 2007-11-08 12:15 -!- cbsmith(~cbsmith@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-08 12:15 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-08 12:16 Good morning North Hollywood! 2007-11-08 12:19 lol 2007-11-08 12:19 As per usual I logged in to #zumastore earlier 2007-11-08 13:52 d'oh! I have a bookmark on my firefox toolbar for irc://irc.oftc.net/#zumastor so I don't have to remember anything 2007-11-08 13:53 thats high tech 2007-11-08 13:53 perhaps we should make the text irc.oftc.net #zumastor under "IRC channel" on our project page link to that 2007-11-08 13:54 Yeah, I do from our internal wiki page already... 2007-11-08 14:00 our external project page sucks 2007-11-08 14:43 I mean to fix it soon myself 2007-11-08 14:47 our last news item is about Scale.. how lame 2007-11-08 14:47 If I went to this page I'd think it was a dead project 2007-11-08 14:52 indeed. I'm thinking of posting a graphical walkthrough of installing our packages on dapper and testing them... 2007-11-08 17:37 -!- cbsmith(~cbsmith@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-08 17:41 ping 2007-11-08 17:41 pong 2007-11-08 17:44 yes? 2007-11-08 17:45 Well, you did touch off a rather stark realization on my part. :-) 2007-11-08 17:45 Yeah, I'm just now reading your email. 2007-11-08 17:46 Fix it however you like, just send me patches. :-) 2007-11-08 17:46 KK. I'm going to check in to the branches for you. 2007-11-08 17:46 I bet at this point you are highly dependent on the test suite setup/teardown only running once per invocation though, right? 2007-11-08 17:46 Yep. 2007-11-08 17:46 KK. Then I'll make sure to fix it such that that is how it works (how it should be anyway). 2007-11-08 17:47 (Although it _should_ survive once per test, but it would be better if not.) 2007-11-08 17:47 For definitions of "survive" that don't include dying of old age, right? ;-) 2007-11-08 17:47 Er, yeah. :-/ 2007-11-08 17:51 Okay, so the error recover if the suite setup/teardown fails is gonna be ugly (read: core dump), but otherwise I think the end result should be good. 2007-11-08 17:56 OK, sounds good. Gotta go, l8r irc.oftc.net #zumastor log beginning Fri Nov 9 00:00:01 PST 2007 2007-11-09 00:08 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-09 02:26 http://www.socallinuxexpo.org/scale5x/conference+info/speakers/Daniel+Phillips/ 2007-11-09 02:30 note the "listen now" link .. heh 2007-11-09 07:04 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-09 08:07 -!- cbsmith(~cbsmith@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-09 08:18 -!- xman(~xman@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-09 08:55 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-09 09:48 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-09 09:50 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor 2007-11-09 14:14 http://shapor.com/pics/cali/2007-01-06_zuma_canyon/.html/DSC06704.JPG.html 2007-11-09 15:13 shapor: I like the new site a lot. 2007-11-09 15:33 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-09 16:10 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-09 16:58 flips: echo "alias svndiff='svn diff --diff-cmd diff -x -rup'" >> .bashrc 2007-11-09 16:58 echo "alias svndsel='svn diff --diff-cmd diff -x -rup | xsel'" >> .bashrc 2007-11-09 16:58 :) irc.oftc.net #zumastor log beginning Sat Nov 10 00:00:01 PST 2007 2007-11-10 01:46 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-10 06:27 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-10 06:27 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor 2007-11-10 07:29 -!- cbsmith(~xman@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-10 08:56 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-10 17:06 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-10 17:41 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor irc.oftc.net #zumastor log beginning Sun Nov 11 00:00:01 PST 2007 2007-11-11 00:54 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-11 02:33 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-11 07:25 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-11 09:52 -!- pgquiles_(~pgquiles@81.202.65.108) has joined #zumastor 2007-11-11 10:42 -!- pgquiles__(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-11 10:56 -!- pgquiles_(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-11 11:37 -!- pgquiles__(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-11 11:41 -!- pgquiles_(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-11 12:41 ok, been working out details of some optimizations to reduce number of metadata blocks written for each copyout 2007-11-11 12:41 reducing number of bitmap writes is easy 2007-11-11 12:41 reducing number of btree leaf writes is a little harder 2007-11-11 13:08 -!- pgquiles__(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-11 13:53 -!- pgquiles_(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-11 18:02 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-11 18:32 flipz: hi 2007-11-11 18:33 can't the same idea be applied to leaf writes? 2007-11-11 18:33 just leave the leaf buffer dirty and log that in the journal? 2007-11-11 18:43 that could also be used as a way of coalescing metadata updated, they could be done in order, with minimal seeking when flushing the journal 2007-11-11 18:43 s/updated/updates/ 2007-11-11 21:28 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-11 22:10 shapor, a similar idea can be applied to leaf writes, the problem is determining which updates have and nave not been applied at replay time 2007-11-11 22:10 good obvservation 2007-11-11 23:44 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor 2007-11-11 23:54 flipz: the only case we really need to handle is inserting an exception to the leaf 2007-11-11 23:55 and we can just defer writes in that case irc.oftc.net #zumastor log beginning Mon Nov 12 00:00:01 PST 2007 2007-11-12 00:03 even then it is hard to know on replay whether or not a particular exception has already been inserted 2007-11-12 00:04 not impossible, just hard 2007-11-12 00:04 more complexity 2007-11-12 00:04 typical problem for logical journalling 2007-11-12 00:04 "hard" 2007-11-12 00:04 not just more complex :-) 2007-11-12 00:04 heh 2007-11-12 00:04 i've never tried 2007-11-12 00:05 why not just check all the exceptions in the journal? 2007-11-12 00:05 do you even need to know? 2007-11-12 00:05 you don't want to insert an exception twice, or insert one that was supposed to be deleted 2007-11-12 00:06 yeah but deletions are rare, you can just flush the journal when you delete 2007-11-12 00:06 anyway, the thing to do is come up with a nice solution to the "how do you tell" problem 2007-11-12 00:07 several ought to present themselves, sounds like you're working on one now 2007-11-12 00:07 here's one from me: record a checksum in the commit block for each journalled block, if the checksum matches then that block has already been replayed 2007-11-12 00:08 obvious question of false matches, requiring bigger checksums 2007-11-12 00:08 or, check for the false match in advance 2007-11-12 00:08 and do something to prevent it 2007-11-12 00:09 extra checksums seem to be the fashion these days, maybe we need to add some 2007-11-12 00:10 why not just check if the physical chunk is already in the leaf? 2007-11-12 00:10 and flush the journal before you do a delete 2007-11-12 00:10 sounds like a typical solution 2007-11-12 00:10 sounds easier 2007-11-12 00:10 deletes are expensive anyway 2007-11-12 00:11 deletes are also done it big batches making them less of a problem 2007-11-12 00:11 yeah, i was thinking before you start one of the big batches 2007-11-12 00:13 splitting a leaf is a complexity 2007-11-12 00:13 just physically journal it 2007-11-12 00:28 flipz, do you have a patch for journalling the bitmap updates? 2007-11-12 00:29 s/ll/l/ 2007-11-12 00:32 shapor, I have a fairly detailed description I'm posting about now 2007-11-12 00:44 ok, there it goes, out to zumastor.googlegroups 2007-11-12 00:45 about time we got the main development discussions moved there 2007-11-12 01:24 yeah i was thinking exactly that 2007-11-12 01:24 done it 2007-11-12 01:24 will consider getting a nap now 2007-11-12 01:27 yeah me too 2007-11-12 02:34 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-12 09:06 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor 2007-11-12 09:06 -!- shapor_(~shapor@yzf.shapor.com) has joined #zumastor 2007-11-12 09:13 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-12 09:13 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-12 09:16 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-12 11:30 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-12 13:14 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-12 13:15 -!- Ganneff(~joerg@ganneff.noc.oftc.net) has joined #zumastor 2007-11-12 13:15 anyone alive? 2007-11-12 13:16 and comments against shapor registering this channel? 2007-11-12 13:17 -!- Ganneff changed mode/#zumastor -> -t 2007-11-12 13:18 hrm. 2007-11-12 13:18 shapor: actually 2007-11-12 13:18 this chan *is* registered. 2007-11-12 13:18 and flips is your master. 2007-11-12 13:19 ah 2007-11-12 13:19 now, change the topic once to follow your website change. 2007-11-12 13:19 ok i'll ping him 2007-11-12 13:20 -!- shapor changed topic to "http://www.zumastor.org" 2007-11-12 13:20 -!- Ganneff changed mode/#zumastor -> +t 2007-11-12 13:20 for the rest talk to flips. 2007-11-12 13:20 -!- Ganneff(~joerg@ganneff.noc.oftc.net) has left #zumastor 2007-11-12 13:28 -!- ChanServ changed mode/#zumastor -> +o shapor 2007-11-12 13:30 :) 2007-11-12 14:14 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-12 14:42 -!- pgquiles__(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-12 14:50 -!- cbsmith(~cbsmith@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-12 17:48 -!- flipz(~phillips@207.47.98.129.static.nextweb.net) has joined #zumastor irc.oftc.net #zumastor log beginning Tue Nov 13 00:00:01 PST 2007 2007-11-13 06:25 -!- camgirl29(~camgirl29@d033.dhcp212-198-248.noos.fr) has joined #zumastor 2007-11-13 10:37 any shapor? 2007-11-13 10:48 ACTION is alive 2007-11-13 10:55 hey 2007-11-13 11:04 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-13 11:21 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-13 12:19 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-13 16:36 -!- shapor_(~shapor@yzf.shapor.com) has joined #zumastor 2007-11-13 16:45 -!- flipz(~phillips@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-13 16:46 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor irc.oftc.net #zumastor log beginning Wed Nov 14 00:00:01 PST 2007 2007-11-14 02:55 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-14 03:06 -!- flips(~phillips@phunq.net) has joined #zumastor 2007-11-14 09:54 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-14 11:00 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-14 11:05 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-14 11:14 hi maze 2007-11-14 11:14 hey 2007-11-14 11:14 shapor_, ping? 2007-11-14 11:14 I saw you wrote me an email... haven't read it yet though ;-) 2007-11-14 11:14 ah, that is what irc is for... reminding about email 2007-11-14 11:15 lol 2007-11-14 11:55 hi all 2007-11-14 11:55 hi flips 2007-11-14 11:59 hi 2007-11-14 12:00 how is 0.4 prerelease test going? 2007-11-14 12:38 still going 2007-11-14 12:38 same speed as usual? 2007-11-14 12:38 much faster 2007-11-14 12:39 got to be seen as good 2007-11-14 12:39 the big test should be finished tonight 2007-11-14 12:39 write caching was disabled on the raid controller 2007-11-14 12:39 where is the bottleneck now? 2007-11-14 12:40 oh, and how did you find out about the raid controller write caching while the test was running? 2007-11-14 12:40 i used the dell utility 2007-11-14 12:41 that's really interesting how a small amount of write cache can have a what? 5x effect on performance 2007-11-14 12:41 whats interesting is that the 2950 with a 5 sata drive raid 5 2007-11-14 12:41 appears to be much faster than a 2850 with a 5 15krpm scsi drive raid5 2007-11-14 12:42 each configuration has a hot spare 2007-11-14 12:42 so a 2950 with 15K rpm sas drivers would be faster yet 2007-11-14 12:42 there is an urgent firmware update for the 2850s as well though 2007-11-14 12:42 sas drives I mean 2007-11-14 12:42 ah 2007-11-14 15:12 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-14 15:23 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-14 15:33 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-14 15:53 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has left #zumastor 2007-11-14 16:36 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-14 17:30 -!- camgirl29(~camgirl29@ANantes-257-1-109-241.w90-25.abo.wanadoo.fr) has joined #zumastor 2007-11-14 17:46 -!- paula35(~paula35@ANantes-257-1-109-241.w90-25.abo.wanadoo.fr) has joined #zumastor 2007-11-14 18:12 -!- obi(~obioni@213.219.189.9.adsl.dyn.edpnet.net) has joined #zumastor 2007-11-14 18:13 Evening all... 2007-11-14 18:14 Can anyone tell me about the status of zumastor? I'm investigating different ways of replicating data between two or more servers 2007-11-14 18:15 I'd be interested to hear if zumastor can be trusted already with "real" data? 2007-11-14 21:17 -!- obi(~obioni@213.219.189.9.adsl.dyn.edpnet.net) has left #zumastor irc.oftc.net #zumastor log beginning Thu Nov 15 00:00:01 PST 2007 2007-11-15 00:37 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-15 01:49 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-15 05:37 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor 2007-11-15 08:55 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has left #zumastor 2007-11-15 08:56 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-15 08:56 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has left #zumastor 2007-11-15 10:12 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-15 10:48 -!- cbsmith(~xman@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-15 12:17 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-15 12:23 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-15 13:58 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-15 13:59 -!- phillips_(~phillips@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-15 16:09 -!- mas(~masouds@216-239-45-4.google.com) has joined #zumastor irc.oftc.net #zumastor log beginning Fri Nov 16 00:00:01 PST 2007 2007-11-16 08:53 -!- pgquiles(~pgquiles@10.Red-217-127-124.staticIP.rima-tde.net) has joined #zumastor 2007-11-16 09:55 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-16 10:14 shapor, there? 2007-11-16 10:14 jiayingz? 2007-11-16 10:17 ACTION is here 2007-11-16 10:17 hi shapor, do we know which 2.6.22 openfiler is using? 2007-11-16 10:18 see some really good comments in this slashdot ramdisk thread: http://hardware.slashdot.org/article.pl?sid=07/11/16/151251 2007-11-16 10:18 particularly this one: http://hardware.slashdot.org/comments.pl?sid=362911&cid=21380057 2007-11-16 10:29 openfiler announces a new kernel but doesn't say which one it is: http://www.openfiler.com/news/?id=40 2007-11-16 10:31 also no tarball, there is a "conary" thingy instead 2007-11-16 10:40 ok, here we go, "Build kernel 2.6.22.10 for 2.3 release": https://project.openfiler.com/tracker/ticket/595 2007-11-16 10:45 yeah i looked around for a few minutes and couldnt find out 2007-11-16 10:46 nothing in their svn reposity seemed to point at a specific version 2007-11-16 10:46 seems to be an rpath-only project 2007-11-16 10:47 they package up centos and make an iso with their gui in it 2007-11-16 10:47 and yet they want only testers/developers to download, seems to me most of those want tarballs 2007-11-16 10:48 well, maybe the iso strategy has benefits, I dunno 2007-11-16 10:48 i dont know if they'll want zumastor 2007-11-16 10:48 Voluna Software is a new commercial entity set-up by the instigators of the Openfiler project to provide commercial level support and consulting services for Openfiler installations. Voluna Software also actively develops enterprise level enhancements to Openfiler. 2007-11-16 10:48 might conflict with their business model :) 2007-11-16 10:49 "High availability and replication functionality is available as a commercial subscription plugin from Voluna Software." 2007-11-16 10:49 http://www.openfiler.com/docs/faq/openfiler-general-faq.html#d0e59 2007-11-16 10:50 well 2.6.22.10 is as good a kernel as any for us 2007-11-16 10:50 will continue with the rebase 2007-11-16 10:50 they it sounds like you've done the work anyway 2007-11-16 10:50 s/they/yeah/ 2007-11-16 10:51 just a little tidying to do now 2007-11-16 10:51 i wonder what their commercial plugin does 2007-11-16 10:51 spits out money? 2007-11-16 10:55 "for the user" 2007-11-16 10:55 I'm putting ".patch" at the end of all the patch file names by the way ;) 2007-11-16 10:55 this time round 2007-11-16 10:56 why the change? 2007-11-16 10:56 everybody else does it 2007-11-16 10:56 lets you have a README or SERIES without name pollution 2007-11-16 11:06 well you were the one who argued for removing the .patch before 2007-11-16 11:06 make up your damn mind :-P 2007-11-16 11:06 the openfiler guys explicitly asked for 2.6.22.10, and they are willing to try it out. 2007-11-16 11:07 fips: thanks for the .patch naming change, that's nice. 2007-11-16 11:07 :) 2007-11-16 11:07 now let me make this happen in time for lunch 2007-11-16 11:08 still some busy work to do 2007-11-16 11:11 what is this dm-ddsnap.c.mine? 2007-11-16 11:11 hmm 2007-11-16 11:12 has something to do with the patch autogenerate? 2007-11-16 11:12 In which case I think improving that is in order 2007-11-16 11:12 no 2007-11-16 11:12 thats not in the repository 2007-11-16 11:13 what about dm-ddsnap.c.r613? 2007-11-16 11:13 sounds like somethign generated by a failed svn merge 2007-11-16 11:13 delete those and svn revert dm-ddsnap.c 2007-11-16 11:13 I'll clean that up 2007-11-16 11:13 thats svn droppings 2007-11-16 11:16 r613 sounds old.. last time you did any work? ;) 2007-11-16 11:17 last time a merge failed more like it 2007-11-16 11:20 shapor, the way I'd like to do it this time is have a ddsnap.base.patch that has everything currently in ddsnap-2.6.21.1 except dm-ddsnap.c and dm-ddsnap.h 2007-11-16 11:25 so basically only one non-auto generated patch 2007-11-16 11:26 dont we lose knowledge by doing that though 2007-11-16 11:26 it is just a bit more work to make all the patches 2007-11-16 11:26 the other bit of that plan is to include the two missing files as an independent patch 2007-11-16 11:27 leaving well enough alone is a viable alternative 2007-11-16 11:27 :) 2007-11-16 11:28 hmm, we have an in-tree export of dm_put for an out-of-tree patch, that breaks the merge 2007-11-16 11:29 so the in-tree export is going bye bye 2007-11-16 11:30 we have confusion about whether we us -p1 or -p0 to apply our patches 2007-11-16 11:30 shapor, how is this resolved when we build a kernel? 2007-11-16 11:43 shapor, ping? 2007-11-16 11:53 getrusage-diff failed all over the place 2007-11-16 11:53 prognosis is: drop this patch for now 2007-11-16 11:55 we always use p1 2007-11-16 11:55 right 2007-11-16 11:55 as I found 2007-11-16 11:55 good 2007-11-16 11:55 ok, so I did the ddsnap.base.patch 2007-11-16 11:55 it's the right thing to do, if we want to carry dm-ddsnap.c/h in plain text 2007-11-16 11:55 thats the easy way 2007-11-16 11:56 so have fewer patches now anyway 2007-11-16 11:56 there is still ddraid stuff in that patch iirc 2007-11-16 11:56 we could strip it out but it doesn't hurt 2007-11-16 11:56 get-rusage is dropped for now, likewise kdb and the sata patch 2007-11-16 11:56 ok 2007-11-16 11:56 we can put the sata patch back if we go forward with the kernel 2007-11-16 11:56 ddraid stuff in which patch? 2007-11-16 11:57 the big ddsnap one 2007-11-16 11:57 ddsnap-2.6.21.1 2007-11-16 11:57 right, that will be our rollup patch 2007-11-16 11:57 maybe 2007-11-16 11:57 well no 2007-11-16 11:57 everything used to be in it 2007-11-16 11:57 it isn't quite 2007-11-16 11:57 right 2007-11-16 11:57 i cut out dm-ddsnap.[ch] 2007-11-16 11:57 ah 2007-11-16 11:57 but left in dm-ddraid 2007-11-16 11:58 because, well i was lazy 2007-11-16 11:58 ok, so that is dropped now 2007-11-16 11:58 ok 2007-11-16 11:58 .base is only about 80 lines long 2007-11-16 11:58 ok, I'll try for a build now 2007-11-16 11:59 but you kept all the kconfig stuff 2007-11-16 11:59 and makefile changes 2007-11-16 11:59 for the md dir 2007-11-16 11:59 cat /src/zumastor/ddsnap/patches/2.6.22.10/*.patch | patch -p1 <- nice n simple 2007-11-16 11:59 bio throttle alone was more than 80 lines 2007-11-16 11:59 correct 2007-11-16 11:59 hows that possible 2007-11-16 11:59 you got rid of it? 2007-11-16 11:59 bio throttle has its own patch 2007-11-16 11:59 oh i see 2007-11-16 12:00 nowi follow 2007-11-16 12:00 gives ziayingz some continuity 2007-11-16 12:00 well 2007-11-16 12:00 more to the point 2007-11-16 12:00 s/z/j/ :) 2007-11-16 12:00 lets us see clearly how we've tweaked things 2007-11-16 12:00 you and names... 2007-11-16 12:01 being precise gets in the way of creativity ;-) 2007-11-16 12:01 hah 2007-11-16 12:03 ok the computer can do some work now 2007-11-16 12:04 let's see what's in that rusage patch 2007-11-16 12:10 bleah 2007-11-16 12:14 shapor, about a dozen warnings compiling dm-ddsnap.c, we should alwlays fix such warnings as soon as we see them 2007-11-16 12:14 I filed a bug 2007-11-16 12:14 you responded 2007-11-16 12:15 a while ago 2007-11-16 12:15 although i think there wereonly 3 warnings then 2007-11-16 12:15 more api changes? 2007-11-16 12:15 too much process in such a case 2007-11-16 12:15 interferes with fixing warnings 2007-11-16 12:15 let's see 2007-11-16 12:15 well i wasn't sure of the fix 2007-11-16 12:15 before it was 2007-11-16 12:16 drivers/md/dm-ddsnap.c: In function `rwpipe': 2007-11-16 12:16 drivers/md/dm-ddsnap.c:70: warning: initialization discards qualifiers from pointer target type 2007-11-16 12:16 drivers/md/dm-ddsnap.c: At top level: 2007-11-16 12:16 drivers/md/dm-ddsnap.c:252: warning: `kmem_cache_t' is deprecated (declared at include/linux/slab.h:17) 2007-11-16 12:16 drivers/md/dm-ddsnap.c:253: warning: `kmem_cache_t' is deprecated (declared at include/linux/slab.h:17) 2007-11-16 12:17 "You can replace kmem_cache_t with struct 2007-11-16 12:17 kmem_cache, which will work both before and after the witch hunt against 2007-11-16 12:17 typedefed structs got underway." 2007-11-16 12:18 CC [M] drivers/md/dm-ddsnap.o 2007-11-16 12:18 drivers/md/dm-ddsnap.c: In function 'rwpipe': 2007-11-16 12:18 drivers/md/dm-ddsnap.c:72: warning: initialization discards qualifiers from pointer target type 2007-11-16 12:18 drivers/md/dm-ddsnap.c: At top level: 2007-11-16 12:18 drivers/md/dm-ddsnap.c:228: warning: 'kmem_cache_t' is deprecated 2007-11-16 12:18 drivers/md/dm-ddsnap.c:229: warning: 'kmem_cache_t' is deprecated 2007-11-16 12:18 drivers/md/dm-ddsnap.c: In function 'incoming': 2007-11-16 12:18 drivers/md/dm-ddsnap.c:575: warning: unused variable 'vm_sock' 2007-11-16 12:18 drivers/md/dm-ddsnap.c: In function 'ddsnap_map': 2007-11-16 12:18 drivers/md/dm-ddsnap.c:1166: warning: left shift count >= width of type 2007-11-16 12:18 drivers/md/dm-ddsnap.c: In function 'ddsnap_seq_show': 2007-11-16 12:18 drivers/md/dm-ddsnap.c:1261: warning: implicit declaration of function 'dm_inflight_total' 2007-11-16 12:18 fixing now 2007-11-16 12:18 eek irc.oftc.net #zumastor log beginning Sat Nov 17 00:00:01 PST 2007 2007-11-17 13:53 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor irc.oftc.net #zumastor log beginning Sun Nov 18 00:00:01 PST 2007 2007-11-18 02:21 -!- erwan__taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-18 05:40 -!- erwan_taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-18 15:05 taking a browse through the "New Kernel Bugs" lkml wankfest... 2007-11-18 15:25 i'm sorry 2007-11-18 20:25 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor irc.oftc.net #zumastor log beginning Mon Nov 19 00:00:01 PST 2007 2007-11-19 00:03 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-19 07:49 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-19 08:49 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-19 10:15 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-19 10:24 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-19 10:30 2.6.22.10 builds 2007-11-19 10:30 just a little tidying and patches will arrive 2007-11-19 10:45 cool :) 2007-11-19 10:57 there we go, heading to the office now 2007-11-19 14:53 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-19 20:39 hmm, I don't see any traffic from our mailing list in my inbox 2007-11-19 20:39 yet I think there is some 2007-11-19 20:44 ah, it was a filter malfunction 2007-11-19 21:43 -!- liona29(~liona29@ANantes-257-1-65-105.w90-25.abo.wanadoo.fr) has joined #zumastor irc.oftc.net #zumastor log beginning Tue Nov 20 00:00:01 PST 2007 2007-11-20 00:29 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-20 02:17 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-20 03:31 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-20 07:22 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-20 11:00 -!- erwan_taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-20 11:19 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-20 11:21 shapor, ping? 2007-11-20 11:25 flips, pong 2007-11-20 11:25 hi shapor, I'm thinking about the patch generation mechanism for kernel/dm-ddsnap.c/h 2007-11-20 11:25 what's the script that does this again? 2007-11-20 11:28 flips: homedir success :) 2007-11-20 11:28 :) 2007-11-20 11:28 good, so I can bother you about this now ;) 2007-11-20 11:28 literally 30 seconds ago 2007-11-20 11:28 I'm telepathic 2007-11-20 11:30 ddsnap/scripts 2007-11-20 11:30 genallpatches.sh 2007-11-20 11:30 its quite lame, feel free to make it nicer 2007-11-20 11:30 :) 2007-11-20 11:31 genpatch.sh is factored out for some silly reason.. historic i think 2007-11-20 11:31 the build script was calling it directly before 2007-11-20 11:33 the ddsnap.base.patch idea is a little tricky to maintain 2007-11-20 11:34 might have to lose that idea 2007-11-20 11:35 it's too hard to pull a nice new patch out of a hacked kernel tree, you have to do something like temporarily removing drivers/md/dm-ddsnap.c/h 2007-11-20 11:37 one can however take advantage of the "feature" of diff that by default it ignores files present in the destination tree but missing in the source (clean) tree 2007-11-20 11:37 and that will generate a diff that omits dm-ddsnap.c+h 2007-11-20 11:37 however, it will also omit any other new files that happened to have been added 2007-11-20 11:38 not that we have any of those, but... 2007-11-20 11:38 well 2007-11-20 11:38 any new files we have by policy should be in the repo 2007-11-20 11:38 not in the patch 2007-11-20 11:38 ok, there's a way to do it 2007-11-20 11:38 in theory some new file might be needed only for one particular kernel version 2007-11-20 11:39 doesn't happen now of course 2007-11-20 11:39 and likely never will 2007-11-20 11:40 well, the way quilt "subtracts" one patch from another is to apply the patch to be subtracted to the base tree (what I call clean) 2007-11-20 11:40 then it diffs 2007-11-20 11:43 yeah.. thats expensive 2007-11-20 11:43 shapor, genpatch.sh is roughly the same way I would do it 2007-11-20 11:45 the whole thing including the build script was about an hour of work 2007-11-20 11:45 except that it's not necessary to have an actual kpath 2007-11-20 11:45 back when we were having commit errors 2007-11-20 11:45 due to double-tracking the dm-ddsnap files 2007-11-20 11:45 in the patch 2007-11-20 11:45 it's not that this is hard, it's just that not having it kind of keeps those files from being maintained properly 2007-11-20 11:45 right 2007-11-20 11:45 yeah 2007-11-20 11:47 ok, kpath doesn't actually do anything except let you provide a custom spelling of the path name 2007-11-20 11:47 we can lose that 2007-11-20 11:47 we can lose the .AUTO too 2007-11-20 11:47 and the .sh 2007-11-20 11:47 incremental improvements ;) 2007-11-20 11:48 can also lose the destination patch directory, the output can go into . 2007-11-20 11:48 well 2007-11-20 11:48 hmm 2007-11-20 11:49 .AUTO was to be able to delete the auto generated patches easily with make clean iirc 2007-11-20 11:49 I suppose you're right, each update has to update _all_ patch directories 2007-11-20 11:49 make clean probably should not touch autogenerated source 2007-11-20 11:50 make mrproper is the usual syntax for that kind of thing 2007-11-20 11:50 or dist-clean 2007-11-20 11:50 whatever that means 2007-11-20 11:50 getting down to shades of gray 2007-11-20 11:51 ok, time to lose the earlier kernel patches, no longer used 2007-11-20 11:51 why 2007-11-20 11:52 someone might want to apply them to a distribution kernel 2007-11-20 11:52 because we can't reasonably maintain dm-ddsnap.c+h against kernel versions we don't support 2007-11-20 11:52 if somebody wants that, they need to find an old tarball 2007-11-20 11:52 or look back in time in the repo 2007-11-20 11:52 why not leave them in the repo as depricated 2007-11-20 11:53 *shrug* 2007-11-20 11:53 ACTION still runs 2.6.8 at home 2007-11-20 11:53 we have a version control system so we don't need to leave scraps lying around 2007-11-20 11:53 svn diff 2007-11-20 11:53 svn: Can't find a temporary directory: Internal error 2007-11-20 11:53 eek 2007-11-20 11:53 hah 2007-11-20 11:53 i hate how svn pollutes iwth .svn 2007-11-20 11:53 we have something resembling a version control system anyway 2007-11-20 11:54 shiwhwih 2007-11-20 11:54 it's actually ENOSPC, thanks a lot svn 2007-11-20 11:55 flips: drake recommands using git as an svn client 2007-11-20 12:03 ACTION recommends darcs ;-) 2007-11-20 12:04 see you in a few minutes 2007-11-20 12:10 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-20 14:05 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-20 15:45 -!- juuva(juuva@peili.org) has joined #zumastor 2007-11-20 17:35 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-20 17:38 Trying out Jiaying's refurbished tests/uml/demo.sh now... 2007-11-20 17:38 guess I'll let it run while I drive home. 2007-11-20 17:58 -!- phillips_(~phillips@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-20 21:47 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor 2007-11-20 22:13 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor irc.oftc.net #zumastor log beginning Wed Nov 21 00:00:01 PST 2007 2007-11-21 00:14 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-21 01:03 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-21 05:46 -!- user_(~liandong@123.5.72.229) has joined #zumastor 2007-11-21 08:45 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-21 08:54 It exited with status 1 right after building the kernel, no clue why. 2007-11-21 09:07 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-21 10:11 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-21 10:22 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-21 10:34 -!- erwan_taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-21 11:32 shapor, there? 2007-11-21 11:48 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-21 11:59 flips pong 2007-11-21 12:07 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-21 13:07 -!- cbsmith(~xman@adsl-71-133-80-65.dsl.irvnca.pacbell.net) has joined #zumastor 2007-11-21 14:25 -!- cbsmith(~xman@64.148.65.11) has joined #zumastor 2007-11-21 15:16 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-21 23:08 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-21 23:26 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor irc.oftc.net #zumastor log beginning Thu Nov 22 00:00:01 PST 2007 2007-11-22 00:26 hello 2007-11-22 00:26 flips: ping 2007-11-22 00:26 hi shapor 2007-11-22 00:27 hey 2007-11-22 00:28 so 0.4 time? 2007-11-22 00:28 got email from dkegel? 2007-11-22 00:28 yeah 2007-11-22 00:28 rading 2007-11-22 00:28 so Jeff had a problem using build_packges.sh 2007-11-22 00:29 no horribly surprising 2007-11-22 00:29 I've been using buildcurrent.sh recently 2007-11-22 00:29 why doesn't he know about this channel by the way? 2007-11-22 00:29 theres so much duplication between the code 2007-11-22 00:29 between the two* 2007-11-22 00:29 it has really exploded sincemy simple 10 line build script 2007-11-22 00:29 and I can't see any additional functionality 2007-11-22 00:30 having trouble finding the email 2007-11-22 00:30 gross 2007-11-22 00:30 it also seems to have multiplied 2007-11-22 00:30 we need a round of shrinking then 2007-11-22 00:30 in to two scripts :) 2007-11-22 00:31 found the mail 2007-11-22 00:32 ok, 0.4 release, yes 2007-11-22 00:32 which means.... email on googlegroups? 2007-11-22 00:32 + news on zumastor.org? 2007-11-22 00:32 plust news on lwn + random geek site tomorrow? 2007-11-22 00:33 well i'm testing out the build script 2007-11-22 00:33 but yeah 2007-11-22 00:34 so r958 test passed 2007-11-22 00:34 ok, I will write the announcement this time then 2007-11-22 00:34 and circulate 2007-11-22 00:34 automated test that is 2007-11-22 00:34 I think it's my turn 2007-11-22 00:34 go for it :) 2007-11-22 00:35 do we need a catchy stupid name? 2007-11-22 00:35 something like "bedrock" 2007-11-22 00:35 haha no 2007-11-22 00:36 what do you mean, "no"? 2007-11-22 00:36 ;) 2007-11-22 00:36 that seems silly 2007-11-22 00:40 release notes? 2007-11-22 00:41 must have workable ones around for one of the big pre-releases 2007-11-22 00:41 what are the subject lines for the pre-releases? 2007-11-22 00:42 Linux 1.3.51 aka "Greased Weasel" 2007-11-22 00:42 http://www.uwsg.iu.edu/hypermail/linux/alpha/9512.3/0004.html 2007-11-22 00:47 Changelog is way out of date, only describes the 0.7 release 2007-11-22 00:55 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-22 01:40 so freshmeat has 6 project states 2007-11-22 01:40 planning, pre-alpha, alpha, beta, production/stable, and mature 2007-11-22 01:40 where the heck are we? 2007-11-22 01:47 Your new project (ID 66950) has been stored for verification by the freshmeat crew. 2007-11-22 01:49 so what's the current status ? 2007-11-22 01:51 well we're release 0.4, which is our stability release 2007-11-22 01:51 but we haven't had a lot of people using it yet 2007-11-22 01:51 so its hard to call that production ready :) 2007-11-22 01:53 we are alpha 2007-11-22 01:54 rather horribly stable alpha to be sure 2007-11-22 01:54 actually more like post-beta quality 2007-11-22 01:54 but what the heck 2007-11-22 01:54 it turns out you dont have to choose on freshmeat 2007-11-22 01:56 want to make an email user at zumastor.org? 2007-11-22 01:56 I will forward the patches there 2007-11-22 01:57 what are the main features of 0.4 ? 2007-11-22 01:57 snapshotting and remote replication with a nice admin script 2007-11-22 01:58 oriented to NFS serving, but not exclusive to NFS servintg 2007-11-22 01:58 what are the main new features maybe you meant? 2007-11-22 01:59 stability 2007-11-22 01:59 not corrupting data 2007-11-22 01:59 to be honest I didn't followed zumastor for a while and I was wondering this morning 'what's the current status of zumastor" ;) 2007-11-22 01:59 not losing sync on the replication 2007-11-22 02:00 I thought zumastor was working at block level, do you integrate scripts to handle filesystems too ? 2007-11-22 02:00 meaning I create a volume, format it using ext3 or xfs etc.. 2007-11-22 02:00 yes 2007-11-22 02:00 and as you are speaking about nfs, you also have commands to start (maybe migrate) nfs servers on top of that ? 2007-11-22 02:01 yes, it's compatible with linux knfsd 2007-11-22 02:01 theres a walk through here: http://zumastor.org/demo-qemu-ubuntu.html 2007-11-22 02:01 integrated with knfsd is a better way of saying it 2007-11-22 02:01 The first half is just getting ubuntu running under qemu 2007-11-22 02:02 perhaps the non install related steps though be broken out in to a seperate doc 2007-11-22 02:02 that all the walkthroughs should point at 2007-11-22 02:02 we need to explain some basic ideas I think 2007-11-22 02:03 yeah 2007-11-22 02:03 i was thinking about what the user would expect 2007-11-22 02:03 such as the idea that you have _this_ disk of here and zumastor magically makes a continuous copy of it _over there_ 2007-11-22 02:03 yeah.. with pretty pictures 2007-11-22 02:04 sort of like quantum disk entanglement 2007-11-22 02:04 sort of not very much like that 2007-11-22 02:04 well, I was also planning on writing a summary of such a document to submit to the scale cfp by this weekend 2007-11-22 02:04 sounds good 2007-11-22 02:04 ugh much to do 2007-11-22 02:04 heh 2007-11-22 02:04 much done too 2007-11-22 02:04 yeah 2007-11-22 02:04 getting stable is really the thing 2007-11-22 02:09 zumastor.googlegroups has 37 members 2007-11-22 02:09 lots of lurkers 2007-11-22 03:01 zfs community is tiny compared to the ext3 community 2007-11-22 07:26 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-22 10:39 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-22 13:27 lots of reports of opensolaris + zfs panicking for various reasons, interesting 2007-11-22 13:43 not suprising, filesystems are hard to get right 2007-11-22 14:09 I'd like to ask somebody who has been working with opensolaris + zfs for a while what the situation is like today 2007-11-22 14:10 I've asked around, don't know anyone who runs it 2007-11-22 14:10 I think having to run opensolaris turns a lot of people away from it 2007-11-22 14:10 people want Linux 2007-11-22 14:14 nice to see the commit messages arriving at zumastor@googlegroups. I think. 2007-11-22 14:26 yeah I had to fix it this morning 2007-11-22 14:26 nosite-noreply wasn't allowed to post to the list 2007-11-22 15:34 ugh the man pages are horrible 2007-11-22 15:35 the technical content is fine, just the stuff like description are way off 2007-11-22 15:56 if anyone is around and wants to work on cleaning up the man pages go for it 2007-11-22 15:56 i have a turkey to attend to :) 2007-11-22 17:04 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor 2007-11-22 21:45 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-22 21:48 -!- cbsmith(~xman@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-22 22:09 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-22 22:14 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-22 22:54 Anyone looking at my gentoo commits? 2007-11-22 23:55 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor irc.oftc.net #zumastor log beginning Fri Nov 23 00:00:01 PST 2007 2007-11-23 00:09 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-23 02:38 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-23 03:58 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-23 04:12 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor 2007-11-23 06:58 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-23 11:02 -!- flips(~phillips@phunq.net) has joined #zumastor 2007-11-23 14:34 -!- pgquiles(~pgquiles@131.Red-83-41-234.dynamicIP.rima-tde.net) has joined #zumastor irc.oftc.net #zumastor log beginning Sat Nov 24 00:00:01 PST 2007 2007-11-24 02:47 -!- pgquiles(~pgquiles@131.Red-83-41-234.dynamicIP.rima-tde.net) has joined #zumastor 2007-11-24 02:47 -!- pgquiles(~pgquiles@131.Red-83-41-234.dynamicIP.rima-tde.net) has joined #zumastor 2007-11-24 07:25 -!- zumalog(~zumalog@yzf.shapor.com) has joined #zumastor 2007-11-24 09:52 -!- flips(~phillips@phunq.net) has joined #zumastor 2007-11-24 09:59 -!- cbsmith(~xman@adsl-76-240-82-190.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-24 21:31 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-24 21:55 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor irc.oftc.net #zumastor log beginning Sun Nov 25 00:00:01 PST 2007 2007-11-25 05:08 -!- pgquiles(~pgquiles@107.Red-83-33-70.dynamicIP.rima-tde.net) has joined #zumastor 2007-11-25 05:38 -!- pgquiles_(~pgquiles@107.Red-83-33-70.dynamicIP.rima-tde.net) has joined #zumastor 2007-11-25 10:24 -!- ancorey(~ancorey@33.Red-83-32-1.dynamicIP.rima-tde.net) has joined #zumastor 2007-11-25 10:24 hola 2007-11-25 10:24 quien existe 2007-11-25 12:04 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has joined #zumastor 2007-11-25 14:57 -!- MaZe(~MaZe@c-67-188-123-1.hsd1.ca.comcast.net) has left #zumastor irc.oftc.net #zumastor log beginning Mon Nov 26 00:00:01 PST 2007 2007-11-26 00:34 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-26 01:00 -!- erwan_taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-26 01:01 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-26 03:03 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-26 10:55 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-26 10:55 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-26 10:56 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-26 12:35 -!- cbsmith(~xman@64.148.65.12) has joined #zumastor 2007-11-26 12:43 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-26 12:49 ping 2007-11-26 12:57 pong 2007-11-26 12:57 wish chatzilla had timestamps 2007-11-26 12:57 ah, it does. 2007-11-26 13:48 hey, where should I upload my Gentoo/Zumastor vmware image to? 2007-11-26 14:04 Shapor's not here at the moment, but the answer is zumastor.org, I think... 2007-11-26 14:11 cbsmith: how big is it? 2007-11-26 14:11 shapor: I made it a 5 GB image, although only ~3GB of it has data. 2007-11-26 14:12 thats huge! 2007-11-26 14:12 Can you slim it down? 2007-11-26 14:12 Well you need space to compile glibc. ;-) 2007-11-26 14:12 i doubt anyone is going to download that 2007-11-26 14:12 shapor: bittorrent? ;-) 2007-11-26 14:12 You need space, but not 3GB of data. 2007-11-26 14:12 I will see if I can trim it down. 2007-11-26 14:13 some big is it compressed? 2007-11-26 14:13 s/some/how/ 2007-11-26 14:13 Oh, that's raw size. Not compressed. 2007-11-26 14:13 I would expect to get at least a 2:1 compression ratio. 2007-11-26 14:13 And can you leave a bunch of packages out, maybe? A 250MB compressed image would be nice. 2007-11-26 14:14 hosting it isn't a problem, i have about 1TB/month of unused bandwidth 2007-11-26 14:14 dkegel: I'll do some trimming. 2007-11-26 14:14 but i highly doubt anyone is going to download it 2007-11-26 14:15 Okay, let me get back to you after a shrinking effort. 2007-11-26 14:15 I'll also do some tests to see how well it compresses. 2007-11-26 14:16 I'm guess lrzip might be a good idea for this one. ;-) 2007-11-26 14:21 Has anyone got a VMware image right now? 2007-11-26 14:21 I.e. an Ubuntu one? 2007-11-26 14:23 yeah the site that dkegel linked to recently 2007-11-26 14:23 oh right, but we haven't built our own yet 2007-11-26 14:23 i dont think so 2007-11-26 14:23 http://www.thoughtpolice.co.uk/vmware/ 2007-11-26 14:23 that was it 2007-11-26 14:24 they seem to generally be around 150 or so mb 2007-11-26 14:25 well, Gentoo is going to be bigger than that. Just the sources for the linux kernel take up a good chunk of that. 2007-11-26 14:25 why do the kernel sources need to be in the image? 2007-11-26 14:26 'cause it's a source based system and you build a Zumastor kernel with it. 2007-11-26 14:27 I had two kernel sources though (plain Gentoo + Zumastor), so I'll wipe out the Gentoo one. 2007-11-26 14:27 i still dont understand why its needed 2007-11-26 14:27 the image should be he result of the compling no? 2007-11-26 14:27 otherwise whats the point of an image? :) 2007-11-26 14:30 Yeah, but it's weird to have the binary w/o the sources on a Gentoo system. I guess I could wipe them out and let people install them themselves if they want to. 2007-11-26 14:31 Hehe. I could probably shrink it a lot if I used ReiserFS for the root FS. ;-) 2007-11-26 14:31 Whatever it takes. Remember, gentoo users are insane :-) 2007-11-26 14:32 Plus there is the whole portage tree, which is pretty big. I think tar'd/compressed it's like 40MB. 2007-11-26 14:32 dkegel: Nice. Thanks. 2007-11-26 14:38 Phone guy is *finally* showing up. BBIAB. 2007-11-26 15:30 -!- cbsmith(~xman@adsl-75-56-200-52.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-26 15:34 cbsmith: You're lucky. He could have shown up at 7:00. 2007-11-26 15:35 fmayhar: I know. He was actually really great. Got me up and running very quickly. Life w/o DSL is almost as bad as life w/o air. :-( 2007-11-26 15:35 Well, not _that_ bad... 2007-11-26 15:36 fmayhar: Hey, I said *almost*. ;-) 2007-11-26 16:48 -!- cbsmith(~user@adsl-75-56-200-52.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-26 17:15 Hmmm... looks like lrzip is doing a pretty lousy job. 2007-11-26 17:17 I'm at 1.5G and it isn't done. :-( 2007-11-26 18:58 -!- cbsmith(~user@adsl-76-246-157-64.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-26 23:21 -!- shapor(~shapor@yzf.shapor.com) has joined #zumastor irc.oftc.net #zumastor log beginning Tue Nov 27 00:00:01 PST 2007 2007-11-27 02:14 -!- cbsmith(~user@adsl-75-56-194-53.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-27 04:49 -!- veltan(~veltan@pool23.cl.uni-heidelberg.de) has joined #zumastor 2007-11-27 04:49 Hi 2007-11-27 04:51 hi 2007-11-27 04:52 bin neu hier und wollte mal fragen wie das hier funkt. und ob ich mich auch zu ICQ usern verbinden kann 2007-11-27 04:52 ?? 2007-11-27 04:55 ?? 2007-11-27 10:02 good morning 2007-11-27 10:13 jeff thinks coming over to meet the team would be fun 2007-11-27 10:16 cool :) 2007-11-27 10:24 -!- erwan_taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-27 10:52 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor 2007-11-27 10:52 ping 2007-11-27 10:53 pong 2007-11-27 11:00 pong too 2007-11-27 11:00 say, how about releasing 0.4? 2007-11-27 11:01 seems to me the release announcement is the only thing standing between us and that 2007-11-27 11:01 ACTION goes to read 2007-11-27 11:02 better docs perhaps 2007-11-27 11:03 never will be perfect, but avoiding acute embarassment is a worthy goal 2007-11-27 11:03 sometimes 2007-11-27 11:09 shapor, I think I goofed a bit with the "obtain the device name from the server" idea 2007-11-27 11:09 which is used in ddsnap transmit 2007-11-27 11:10 it does seem like policy 2007-11-27 11:10 which would come from zumastor 2007-11-27 11:10 now, I'm looking at a command syntax like "ddsnap revert " and wondering how ddsnap will derive the device names 2007-11-27 11:11 the idea was, the snapshot server would known the device names from the dmsetup creates, by querying the clients 2007-11-27 11:11 hm 2007-11-27 11:11 the fly in that ointment is, some operations on snapshots are valid with no client 2007-11-27 11:12 for example, revert 2007-11-27 11:12 wait 2007-11-27 11:12 the syntax is wrong 2007-11-27 11:12 it's "ddsnap revert " 2007-11-27 11:13 same issue, the command would be valid even if there were no dm client for 2007-11-27 11:14 but, you need a client 2007-11-27 11:14 in order to get to the data, dont you? 2007-11-27 11:14 now, the only reason this is an issue is that I want to have ddsnap read via the virtual snapshot device to do the revert 2007-11-27 11:14 yes, unless the server does it 2007-11-27 11:14 right 2007-11-27 11:14 which i suppose it would have to 2007-11-27 11:14 so my quick hack to get a revert happening results in issues for the command syntax 2007-11-27 11:14 n.g. 2007-11-27 11:14 unless it locks all the blocks or something 2007-11-27 11:15 the locking is not as bad as I first thought 2007-11-27 11:15 only origin chunks of a snapshot get readlocked 2007-11-27 11:15 snapshot resident chunks don't get moved, so don't need locks 2007-11-27 11:16 why do you even to to readlock the origin chunks? 2007-11-27 11:16 oh right, so they dont get stomped on 2007-11-27 11:16 origin chunks need to be readlocked when an origin client is reading them 2007-11-27 11:16 right 2007-11-27 11:17 well if we're going to change the transmit syntax we should do that sooner rather than later 2007-11-27 11:18 ok, the disadvantages of continuing with my quick hack revert implementation is: 1) we gain a new ddsnapd message that might become obsolete later 2) we need to do another of the name stem hacks, reducing generality 2007-11-27 11:18 it could get the paths from the connected dm devices 2007-11-27 11:18 and error if there are not clients 2007-11-27 11:18 i mean, transmit requires them to exist anyway 2007-11-27 11:18 transmit pretty much requires a client, doesn't it? 2007-11-27 11:18 yep 2007-11-27 11:18 well in theory it could operate without clients 2007-11-27 11:18 just like the revert 2007-11-27 11:19 so what to do there isn't obvious either 2007-11-27 11:19 but supporting that not-obviously-useful case makes the interface suck 2007-11-27 11:19 it's not obviously useless 2007-11-27 11:19 it also has a simpler syntax 2007-11-27 11:20 hmm 2007-11-27 11:20 lunchtime topic? 2007-11-27 11:20 sure 2007-11-27 13:41 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-27 13:56 -!- juuva(juuva@peili.org) has joined #zumastor 2007-11-27 14:14 -!- erwan__taf(~erwan@konilope.linuxeries.org) has joined #zumastor 2007-11-27 17:34 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-27 17:58 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-27 19:30 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-27 21:22 -!- jiayingz(~jiayingz@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-27 23:02 shapor, still around? 2007-11-27 23:21 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-27 23:29 flips yeah 2007-11-27 23:29 hi shapor, the question is: is accepted@zumastor.org an address I can mail too, and if so... can everybody? and would that make sense? irc.oftc.net #zumastor log beginning Wed Nov 28 00:00:01 PST 2007 2007-11-28 00:34 flips: accepted-patches 2007-11-28 00:34 and yeah, you should be able to send mail to it 2007-11-28 00:42 can anybody? 2007-11-28 00:42 yeah 2007-11-28 00:42 problem :-) 2007-11-28 00:43 anyway, sleeping soon is an issue 2007-11-28 00:43 its early :) 2007-11-28 00:43 I was young once too 2007-11-28 00:56 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-28 01:03 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-28 04:15 is zumastor ready and safe for production yet? 2007-11-28 05:11 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-28 08:06 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-28 08:32 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-28 10:36 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-28 10:51 ACTION heads in 2007-11-28 13:44 -!- cbsmith(~xman@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-28 14:08 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-28 14:40 I keep narrowly missing pgquiles 2007-11-28 14:41 shapor: ? 2007-11-28 14:41 hi 2007-11-28 14:41 you were asking about zumastor stability 2007-11-28 14:41 yes 2007-11-28 14:41 but nobody replied 2007-11-28 14:41 i think most of us were sleeping 2007-11-28 14:41 :) 2007-11-28 14:41 oh 2007-11-28 14:41 timezone issue 2007-11-28 14:41 it's about time to sleep here in Europe now 2007-11-28 14:42 yeah.. but anyway, we have packages on our website for our stability release 2007-11-28 14:42 0.4 2007-11-28 14:42 we've been wroking on getting it stable for the last 6 months or so 2007-11-28 14:42 and haven't been able to break it with any of our tests for a while now 2007-11-28 14:43 we're jsut working on documentation before calling 0.4 an official release 2007-11-28 14:43 we've never lost data on it, the risk is actually quite low 2007-11-28 14:43 we only add snapshots on a separate block device 2007-11-28 14:44 worst case if there is a serious bug and the snapshot store gets corrupt, you still have the original filesystem on the origin 2007-11-28 14:44 sounds good :-) 2007-11-28 14:44 which you could just mount and use 2007-11-28 14:44 you would lose all your snapshots in that worst cause though 2007-11-28 14:45 one thing which is not clear to me: by snapshot, do you mean something like "versioned filesystem" or something like LVM-snapshots? 2007-11-28 14:46 ddsnap provides block device snapshots, very similar to lvm 2007-11-28 14:46 with some important differences 2007-11-28 14:46 1) they actually work 2007-11-28 14:46 2) data is not duplicated for each snapshot (it is shared) 2007-11-28 14:46 3) performance does not degrade with number of snapshots 2007-11-28 14:47 what about storage consumption? how much do snapshots need? (if these questions are answered somewhere else, just tell me where, no need to waste your time answering) 2007-11-28 14:48 well, that depends on filesystem churn 2007-11-28 14:48 the default is to use 16k chunksize 2007-11-28 14:48 and we use copy-on-write 2007-11-28 14:49 so 16k of data gets copied to the snapshot store at a time 2007-11-28 14:49 some of these issues are covered in the whitepaper 2007-11-28 14:49 which is linked to under documentation on zumastor.org 2007-11-28 14:50 yes, I read it a few months ago but I got the feeling zumastor was not quite ready yet 2007-11-28 14:50 we're at the point where we feel zumastor is ready, but we have few users 2007-11-28 14:51 it would be great if you want to give it a try 2007-11-28 14:51 sure 2007-11-28 14:51 tomorrow 2007-11-28 14:51 I'm going to setup 4 DL320s with 4TB hard disk, remote replication and fail-over 2007-11-28 14:52 sounds like a nice setup, what distribution do you run? 2007-11-28 14:52 4.5TB hard disk, my bad 2007-11-28 14:52 actually, I don't care 2007-11-28 14:53 i'm guessing thats a hardware raid of sata disks? 2007-11-28 14:53 I used to use Debian but I usually use Ubuntu Server now 2007-11-28 14:53 cool, we have packages for ubuntu dapper, but they are 32 bit 2007-11-28 14:53 yeah, 6x750 SATA with a SmartArray P400 and 512MB of battery-backed write cache 2007-11-28 14:53 no problem, I can build packaged myself 2007-11-28 14:53 we've tested on a similar setup using dell 2850s and 2950s 2007-11-28 14:54 packages 2007-11-28 14:54 can Zumastor be iSCSI target? 2007-11-28 14:54 cool, let us know if you run in to any issues 2007-11-28 14:54 its never been tested 2007-11-28 14:54 I'm going to test it 2007-11-28 14:54 however, it should work, it is just a block device 2007-11-28 14:54 :-) 2007-11-28 14:54 excellent 2007-11-28 14:54 using open-iscsi? 2007-11-28 14:55 yes, unless you suggest anything else 2007-11-28 14:55 no, last time i tried to do iscsi in linux i found things in a pretty sad state 2007-11-28 14:55 that was years ago though 2007-11-28 14:56 it should work fine now, LeftHand builds SAN/iQ on top of iSCSI IIRC 2007-11-28 14:57 actually, at first I was going to use OpenFiler but then we decided we needed virtualization and a better management console and decided to build it ourselves 2007-11-28 14:58 yeah, the openfiler guys are working on testing zumastor with their setup, they have done quite a bit of work bringing the bits and pieces together it seems 2007-11-28 14:59 I've read Daniel Philips e-mails to the openfiler mailing list, yes 2007-11-28 14:59 does Zumastor provide anything like ZFS's RAID-Z? 2007-11-28 15:00 not yet 2007-11-28 15:01 ok 2007-11-28 15:03 when you give it a try it would be nice to hear what additional features you'd like to see, we have lots of ideas 2007-11-28 15:05 sure 2007-11-28 15:05 are those ideas written anywhere I could read them? 2007-11-28 15:09 lots of them are in heads and they are slowly trickling out to our mailing list 2007-11-28 15:11 a single place to find them would be great (at least, until I can read your minds) 2007-11-28 15:12 is there any API to develop applications which need info from zumastor? (for example, a management console) 2007-11-28 15:12 s/console/webapp 2007-11-28 15:15 theres currently no api beyond the cli 2007-11-28 15:16 although thats a good feature request :) 2007-11-28 15:16 we make some stats available via /proc, but that is not finalized 2007-11-28 15:17 what kind of api would you be interested in, and would a stable cli suffice? 2007-11-28 15:17 I'll take a look at OpenFiler to see what Rafiu is doing, although he is using PHP and I will be using C++ 2007-11-28 15:18 CLIs are fine but they are tedious to parse 2007-11-28 15:18 is this an open source project you're working on? 2007-11-28 15:18 it's an internal project but we can open-source it if there is interest 2007-11-28 15:19 it's not our core business 2007-11-28 15:19 interesting 2007-11-28 15:20 so its an openfiler like project? 2007-11-28 15:20 the kind of API I'd like to have is something which allows me to create a volume, get the free space, info about snapshots, info about integrity, etc without having to exec() and parse the command's output 2007-11-28 15:21 openfiler-like? it's not intended to be that, as I said, it's for internal use 2007-11-28 15:21 but it might become something like that, yes 2007-11-28 15:21 cool 2007-11-28 15:21 only we will need Xen integration, too 2007-11-28 15:22 your ideas about the api are very similar to what i had in mind about a year ago 2007-11-28 15:22 and information about Active Directory (in a virtualized Windows 2003) status, etc 2007-11-28 15:22 we decided to not spend the time on that and let the cli settle a bit 2007-11-28 15:22 so we'd have a better chance of getting the api right 2007-11-28 15:23 of course, it's better to have a good CLI and no API than a bad CLI and a bad API 2007-11-28 15:23 CLI parsing is always possible, after all 2007-11-28 15:25 also on the ideas issue I agree we should gather that information better.. reading minds is especially hard remotely ;) 2007-11-28 15:26 A page with links to mailing list posts would be a good start 2007-11-28 15:28 yes 2007-11-28 15:29 a wiki page would be enough 2007-11-28 15:56 pgquiles: we haven't tested on volumes larger than ~1.5TB and we know we have an artificial limitation at 2TB due to using the 32 bit BLKGETSIZE ioctl() for determining volume size 2007-11-28 15:58 shapor: how difficult would be to fix that? 2007-11-28 15:58 since we would like to test on larger volumes and it sounds like you'll be running in to this, I'm opening an issue for it 2007-11-28 15:58 great 2007-11-28 15:58 quite easy, the fd_size function we have can just be replcated by getsize from util-linux-ng 2007-11-28 15:59 I'll use <2TB volumes until it's fixed 2007-11-28 16:01 I don't have a solid deadline to implement the systems, therefore I can test what works now, iSCSI, Xen, start developing the management webapp, etc 2007-11-28 16:23 -!- fmayhar(~fmayhar@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-28 16:35 time to go to bed here 2007-11-28 16:36 see you 2007-11-28 20:47 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor 2007-11-28 22:27 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #zumastor irc.oftc.net #zumastor log beginning Thu Nov 29 00:00:01 PST 2007 2007-11-29 00:22 created an "accepted" mbox as a local folder, moved the marzinksi patches to it, soon will do the apply 2007-11-29 00:23 I think a similar procedure can be used by anybody with svn commit access, given appropriate acks/lack of nacks 2007-11-29 00:24 yeah, its lot like we have so many patches we need such a mechanism to remember to commit them :) 2007-11-29 00:24 actually, the mbox is named "accept" since actual "accepted" still depends on whether diffstat and patch --dry-run work out ok 2007-11-29 00:25 so now I can do diffstat Mail/accept 2007-11-29 00:29 grep Subject: Mail/accept 2007-11-29 00:29 Subject: [PATCH 2/7] bio-throttle.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 2/7] bio-throttle.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 3/7] ddsnap.base.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 4/7] less.throttle.failed.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 5/7] remove.alloc.pages.stackdump.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 6/7] remove.nfsd.less.throttle.patch 2.6.23.9 port 2007-11-29 00:29 Subject: [PATCH 7/7] nfs_suspend.patch 2.6.23.9 port 2007-11-29 00:30 Subject: Re: [PATCH 1/7] drivers/md/dm-ddsnap.c 2.6.23.9 port 2007-11-29 00:34 whoops, a simple accept does not quite work, because different patches apply to different trees 2007-11-29 00:35 can't just apply the whole mbox in one patch command, which is my aim 2007-11-29 00:35 let's see how many separate mboxes would be required 2007-11-29 00:35 all patches should be against the trunk -p1 2007-11-29 00:35 that is true 2007-11-29 00:36 in fact though, ben made his diffs against real kernel trees 2007-11-29 00:36 can't avoid that 2007-11-29 00:36 ben did it right 2007-11-29 00:36 yeah, its the annoying part of maintaining patches 2007-11-29 00:36 we can't take patches against /partches/ 2007-11-29 00:36 yeah thats too confusing 2007-11-29 00:36 so I need too accept mboxes I think 2007-11-29 00:36 two even 2007-11-29 00:37 it appears like what you need is sleep 2007-11-29 00:37 --- a/dm-ddsnap.c 2007-11-27 11:24:41.000000000 -0600 2007-11-29 00:37 +++ b/dm-ddsnap.c 2007-11-27 11:24:05.000000000 -0600 2007-11-29 00:37 we should specify that the top level dir be zumastor (and not zumastor/zumastor) 2007-11-29 00:38 sleep would be ncie too 2007-11-29 00:39 should just be trunk 2007-11-29 00:39 see you tomorrow at 10 2007-11-29 00:39 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor 2007-11-29 00:39 not everybody gets "trunk" in the path 2007-11-29 00:39 well the root of the trunk 2007-11-29 00:39 top level should be "zumastor" just like the top level of a linux patch is linux-xx-x-x 2007-11-29 00:39 whatever they have checked it out as 2007-11-29 00:39 good morning pgquiles 2007-11-29 00:41 bio-throttle.patch got misnamed 2007-11-29 00:41 was bio.throttle.patch 2007-11-29 00:42 shapor: hi 2007-11-29 00:42 shapor: I'm installing ubuntu gutsy in one of the servers, will try zumastor soon 2007-11-29 00:43 pgquiles, hi, nice to meet you, now I must immediately go get some sleep ;-) 2007-11-29 00:43 flips: great 2007-11-29 00:43 sounds good, i'll probably be up for a couple hours more 2007-11-29 00:43 flips: it's about 10am here :-D 2007-11-29 00:43 when you sleep, I work 2007-11-29 00:43 when I work I sleep 2007-11-29 00:44 :-D 2007-11-29 00:44 when I work and it does not kill me I grow stronger 2007-11-29 00:44 on that nietzian note.. 2007-11-29 01:50 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-29 02:36 shapor: thanks for the 2TB limit fix 2007-11-29 06:22 pgquiles: doh! 2007-11-29 06:22 nice catch 2007-11-29 06:24 I've seen the 'sh is bash' bug in many places, it's quite common 2007-11-29 06:25 yeah, unfortunately we haven't tested everything in the environment where sh != bash 2007-11-29 06:25 that one is my fault too :( 2007-11-29 06:26 bad habit, typing /bin/sh and using bashisms :) 2007-11-29 06:27 pgquiles: by the way, where did you get the sources from, svn, or tarballs? 2007-11-29 06:28 svn 2007-11-29 06:30 ok, because the packages on the site are from our 0.4 stability branch, the trunk is a bit more bleeding edge 2007-11-29 06:30 but we haven't actually changed too much on it yet 2007-11-29 06:31 oh, I didn't notice there was a stability branch 2007-11-29 06:31 I receive lots of warnings while building the kernel, is that fine? 2007-11-29 06:32 when building dm-ddsnap ? 2007-11-29 06:33 it says "+ echo -n Building kernel package...", is taht dm-ddsnap? 2007-11-29 06:37 well, its building a whole new kernel 2007-11-29 06:37 so its not suprising there are errors 2007-11-29 06:39 ok 2007-11-29 06:39 its required because we have some minor patches to other parts of the kernel, ddsnap isn't just a drop in module 2007-11-29 06:39 have you tried zumastor in 64-bit linux? 2007-11-29 06:39 yes 2007-11-29 06:39 it does work and we did some load testing, not as extensive as with 32 bit 2007-11-29 06:40 but we didnt encounter any issues 2007-11-29 06:40 sweet 2007-11-29 06:40 it suprised us too ;) 2007-11-29 06:40 :-D 2007-11-29 06:41 but we did go through and clean up the things that gave warning on 64 bit first 2007-11-29 06:41 drivers/md/dm-ddsnap.c: In function ‘ddsnap_map’: 2007-11-29 06:41 drivers/md/dm-ddsnap.c:1167: warning: left shift count >= width of type 2007-11-29 06:41 yeah thats really a bogus warning, there was recently discussion of that on the list 2007-11-29 06:42 happens in 32 bit also 2007-11-29 06:42 but i believe daniel phillips has a workaround for it 2007-11-29 06:42 nothing to worry about 2007-11-29 06:42 great 2007-11-29 06:43 does zumastor prefer any particular filesystem? 2007-11-29 06:43 most of our testing has been ext3 2007-11-29 06:44 I tend to use ext3 too, but now that I have BBWC I wanted to test XFS. Is there any known issue with it? 2007-11-29 06:45 actually there is an issue with mounting snapshots with xfs 2007-11-29 06:46 the uuid's collide with the origin device because it is a block level snapshot 2007-11-29 06:46 so they need to be mounted with nouuid.. and there currently isn't an interface to specify that option with zumastor 2007-11-29 06:47 i was planning on putting up a patch or at least proposed solution for specifying mountpoints and mount options to the list today 2007-11-29 06:48 drake just noticed that issue yesterday, no one had tested xfs in a while before that ;) 2007-11-29 06:48 I feel like a movie specialist :-) 2007-11-29 06:49 btw, i committed the fix for the patch gen script 2007-11-29 06:49 I've seen it 2007-11-29 06:49 thanks for opening an issue, its a good was of tracking things we've fixed 2007-11-29 06:49 I had modified the files locally, anyway 2007-11-29 06:49 my pleasure 2007-11-29 06:51 yeah, I figured 2007-11-29 09:52 -!- cbsmith(~cbsmith@adsl-75-56-194-53.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-29 11:29 hmm i wonder how the install went for pgquiles 2007-11-29 11:30 late night... /me heads in to the office 2007-11-29 11:31 He logged out a short while ago without saying anything, which hopefully means good things. 2007-11-29 11:34 he usually comes back a bit later, probably on his way home :) 2007-11-29 12:03 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-29 12:08 failed to code my revert before lunch I think 2007-11-29 12:09 got involved in writing up an idea for making replicated snapshots usefully writable 2007-11-29 12:09 oh? 2007-11-29 12:11 fell out of the dreamworks chat 2007-11-29 12:11 user writes a file on a replicated filesystem, it basically gets rsynced back to the central server 2007-11-29 12:11 easy huh? no. 2007-11-29 12:11 but doable 2007-11-29 12:13 yeah, no 2007-11-29 12:14 part of it is working a block difference list back into a file difference list 2007-11-29 12:14 interesting food for thought, but waaay in the future 2007-11-29 12:14 oh yes 2007-11-29 12:14 i know you're just dying to come up with a reason to code the block->file mapping :) 2007-11-29 12:15 however the intellectual question is of immediate interest to people who like us 2007-11-29 12:15 a reason to write it up anyway 2007-11-29 12:15 yeah, i'm going to make a page that linked to the "Best of" the mailing list 2007-11-29 12:15 coding it will be quite some effort 2007-11-29 12:16 posts with good technical discussions, otherwise they get buried 2007-11-29 12:16 there's a best of? 2007-11-29 12:16 well, basically all the posts where you spew design for 3 pages until everyone elses head spins and no one replies 2007-11-29 12:16 heh 2007-11-29 12:16 yes, try to widen the circle of spinning heads 2007-11-29 12:16 I love it when every head in sight is spinning 2007-11-29 12:17 pgquiles was asking if all the ideas for future designs we had were recorded somewhere 2007-11-29 12:17 and the answer was sadly, no, theres not 2007-11-29 12:17 mostly in peoples heads, and occassional mailing list posts 2007-11-29 12:17 we should dredge some up from internal mail 2007-11-29 12:17 my outbox would be a good place to start looking 2007-11-29 12:18 jiayingz has a couple too 2007-11-29 12:18 and you 2007-11-29 12:18 a wiki page pointing to googlegroups posts would be a good start which would require little effort 2007-11-29 12:18 that would likely get some views 2007-11-29 12:18 views might turn into code 2007-11-29 12:18 stranger things have happened 2007-11-29 12:18 my thoughts exactly 2007-11-29 12:18 and we really, really have to integrate the snapshot mounting patch 2007-11-29 12:19 or even working code attached to research papers 2007-11-29 12:19 :) 2007-11-29 12:19 we've been good about that so far 2007-11-29 12:19 though the zumastor paper barely qualifies as research 2007-11-29 12:20 i was thinking the mountable delta files 2007-11-29 12:20 oh right 2007-11-29 12:21 we should encourage him to present his paper 2007-11-29 12:21 fast! 2007-11-29 12:21 in fact, and the FAST conference in february 2007-11-29 12:21 I will ping him 2007-11-29 12:21 s/and/at/ 2007-11-29 12:22 Paper submissions due: September 12, 2007, 9:00 p.m. EDT (firm deadline) 2007-11-29 12:22 thats for FAST '08 2007-11-29 12:23 really 2007-11-29 12:23 yeah 2007-11-29 12:23 I just put in a proposal last week 2007-11-29 12:23 not a paper 2007-11-29 12:23 well 2007-11-29 12:23 OLS then 2007-11-29 12:23 better exposure there anyway 2007-11-29 12:24 pgquiles: howd the install go? 2007-11-29 12:26 shapor: I installed trunk in a VMWare Workstation virtual machine, then got abducted into a meeting 2007-11-29 12:26 I had no time to test but at least it boots :-) 2007-11-29 12:26 most excellent! 2007-11-29 12:26 we had trouble booting on vmware before 2007-11-29 12:26 tomorrow I'd like to test zumastor a bit, then build a kernel mixing zumastor and xen patches and run that 2007-11-29 12:26 believed to be due to the 2.6.21.1 kernel we were running though 2007-11-29 12:27 I tried with 2.6.22.14 2007-11-29 12:27 aha 2007-11-29 12:27 so throwing out the vm wankery (replaced with bio wankery) is likely the reason it now works 2007-11-29 12:27 a couple of (maybe stupid) questions 2007-11-29 12:28 is zumastor able to replicate and take snapshots of files in use? 2007-11-29 12:28 of course 2007-11-29 12:28 yes 2007-11-29 12:28 great 2007-11-29 12:28 with the current state of our documentation there are *no* stupid questions 2007-11-29 12:28 I should not have implied it was a stupid question 2007-11-29 12:29 in fact the questions you have will help us write better docs, so fire away 2007-11-29 12:29 just that this project would not have been worth doing if it were not so 2007-11-29 12:29 and how do I mount snapshots to make data accessible? I mean, something like /data/200711292100, /data/200711302100, etc 2007-11-29 12:29 zumastor does it for you 2007-11-29 12:30 if you want to do it at the ddsnap level, you get to work out some gory details involving dmsetup 2007-11-29 12:30 actually, what I want is data to be available through Samba or NFS and users seeing /data/current, /data/20071128, /data/20071127, etc 2007-11-29 12:31 is that what zumastor does for me? 2007-11-29 12:31 essentially, but currently the mountpoints are not as nice as you specify 2007-11-29 12:32 more like /data/volumename, /data/volumename(2), /data/volumename(3) 2007-11-29 12:32 and those numbers are snapshot nubmers rather than timestamps as you would like 2007-11-29 12:32 but that is relatively easy to change 2007-11-29 12:33 its specified in a bash script 2007-11-29 12:33 which does the mounting 2007-11-29 12:33 and is related to the current discussion on the list of allowing the user to specify their own mountpoints and filesystem mount options 2007-11-29 12:34 user can of course hack the script 2007-11-29 12:34 no problem with ugly names, I can show what I want by using symlinks or the Samba share name 2007-11-29 12:35 then push back proposals based on that 2007-11-29 12:35 zumastor looks fucking awesome :-) 2007-11-29 12:35 pgquiles, you know that the zumastor layer (not ddsnap) is a bash script? 2007-11-29 12:35 flips: no, I had not had the time to look at the source 2007-11-29 12:36 heh 2007-11-29 12:36 thinking about making it python, but inertia has kept it bash for now 2007-11-29 12:36 it's really easy to work with 2007-11-29 12:36 and fast enough for the high level stuff 2007-11-29 12:38 -!- dkegel(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-29 12:38 and nobody else like python :-P 2007-11-29 12:38 likes* 2007-11-29 12:39 flips: pgquiles had an interesting request to have an api he could call from c++ rather than shelling 2007-11-29 12:40 which makes python seem like a bad choice 2007-11-29 12:40 c=lowest common denominator 2007-11-29 12:40 can't comment without seeing the specific api calls wanted 2007-11-29 12:40 basic things, creating volumes, setting up snapshots, etc 2007-11-29 12:41 and exec won't work? 2007-11-29 12:41 flips: I'll develop a webapp interface to zumastor as soon as it's working and I'm going to write it in C++ using Wt 2007-11-29 12:41 that's why I prefer C or C++ 2007-11-29 12:41 it will, but shelling and parsing output sucks 2007-11-29 12:41 c++ is my favorite target for a compiled-language rewrite 2007-11-29 12:41 Wt, hmm 2007-11-29 12:42 let me see what that is 2007-11-29 12:42 Inertia isn't the only thing keeping it from being rewritten in python 2007-11-29 12:42 anyway, you could just work up a class based on execing zumastor 2007-11-29 12:42 I believe my words were "zumastor needs to go places python can't follow" 2007-11-29 12:42 flips: www.webtoolkit.eu 2007-11-29 12:43 think of ruby on rails in c++with the Qt API 2007-11-29 12:43 I prefer C as system language because I've seen too much of the C++ sausage factory to trust it yet 2007-11-29 12:44 it's more about trusting the programmers writing the code I think 2007-11-29 12:44 but the web interface can happily be C++ 2007-11-29 12:44 No, C++ itself has some very rough bits that can bite you even if you are aware of them. 2007-11-29 12:44 Things like ABI compatibility are not yet nailed down in g++. 2007-11-29 12:44 c = simple api which can easily be wrapped in py, php, perl, pick your poison 2007-11-29 12:45 Indeed. If you need to hear a litany of complaints about C++ and ABI problems, just ask Mike Hearn (mr. autopackage) 2007-11-29 12:45 "write in the most powerful language you possibly can" 2007-11-29 12:46 let me find the link 2007-11-29 12:46 C is good enough for all other system software, i dont see why it wouldnt be good enough for us 2007-11-29 12:46 as I said earlier, zumastor needs to go places python can't follow, and it's probably true for C++ as well; some poor sod is going to want to invoke us from an initrd 2007-11-29 12:47 actually, ddsnap needs to go to such places, not zumastor 2007-11-29 12:47 zumastor is a high level manager 2007-11-29 12:47 But I bet you $50 people will want to run zumastor there, too. 2007-11-29 12:47 flips: not to anyone but us :) 2007-11-29 12:48 it would be a high crime to put zumastor in an initrd for example 2007-11-29 12:48 never mind that such high crimes are commonplace 2007-11-29 12:48 yeah, but you believe initrd itself is a high crime. 2007-11-29 12:48 know it 2007-11-29 12:48 only way I have ever lost a hard disk on linux 2007-11-29 12:48 by a simple system upgrade 2007-11-29 12:49 So we're both curmudgeons, but on different axes 2007-11-29 12:49 indeed 2007-11-29 12:49 axes at dawn, behind the googleplex 2007-11-29 12:49 oh, shapor is a curmudgeon too 2007-11-29 12:49 closet one 2007-11-29 12:50 can't we throw our axes at the people we all agree to disagree with together? :-) 2007-11-29 12:51 lunchtime 2007-11-29 12:53 I'll be over 2007-11-29 12:53 I'll think about who should be axe-target for the day 2007-11-29 12:54 hmm 2007-11-29 12:54 looks like I'll be over as lunch cools down 2007-11-29 12:54 finishing up my mail to scotty 2007-11-29 13:03 flips: Have you looked at the commentary patch yet? 2007-11-29 13:03 not yet, sorry, still doing followup from yesterday 2007-11-29 13:04 Hoping to get it in this week... 2007-11-29 13:04 right after lunch 2007-11-29 13:04 OK 2007-11-29 13:04 Thanks. 2007-11-29 13:05 no problem 2007-11-29 13:06 cc me on the mail to scotty? 2007-11-29 13:06 yes 2007-11-29 13:18 -!- SEJeff(~jeff@office4.tmcs.net) has joined #zumastor 2007-11-29 13:19 -!- MaZe(~MaZe@216-239-45-4.google.com) has left #zumastor 2007-11-29 13:23 -!- mitchg(~mitchg@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-29 13:30 http://www.paulgraham.com/avg.html <- "write in the most powerful language you possibly can" 2007-11-29 13:46 So you're going to be programming in Lisp now? :-P 2007-11-29 15:40 I'm not arguing with that, just saying that the most powerful language we can allow ourselves for Zumastor probably happens to be C. 2007-11-29 15:53 -!- cbsmith(~cbsmith@adsl-76-243-138-218.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-29 16:13 As long as we're having language wars, we ought to go way out there and talk about OCaml and Haskell. ;-) 2007-11-29 16:26 cbsmith: shut up ;) 2007-11-29 16:26 shapor: Coq? 2007-11-29 16:26 :-) 2007-11-29 16:27 as soon as i typed that i heard laughter coming from frank's general direction ;) 2007-11-29 16:27 ACTION is shocked... okay, not really. :-) 2007-11-29 16:28 I can actually make a half-baked argument for OCaml and Haskell though: the Linspire people actually use them for system management and init scripts. 2007-11-29 16:29 Which would seem to at least partially fly in the face of the "we must go places where only C can go" argument. 2007-11-29 18:36 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor 2007-11-29 18:37 Golly. Well, firstly, is linspire still doing that since they switched to basing off Ubuntu? And secondly, do you really want to hold up Linspire as an example of good system management? I thought not... 2007-11-29 19:53 dank: Well, that's exactly why I referred to it as half-baked. 2007-11-29 19:53 dank: But, AFAIK, they are still doing it since they switched to Ubuntu. 2007-11-29 23:39 -!- pgquiles(~pgquiles@62.43.226.52.static.user.ono.com) has joined #zumastor irc.oftc.net #zumastor log beginning Fri Nov 30 00:00:01 PST 2007 2007-11-30 01:13 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-30 02:45 ACTION is alive 2007-11-30 04:21 -!- erwan_taf(~erwan@81.80.43.67) has joined #zumastor 2007-11-30 10:05 good morning 2007-11-30 10:17 hi flips 2007-11-30 10:18 ok, it's finish coding revert this morning or bust! 2007-11-30 10:25 ok, zumastor man page description is still broken 2007-11-30 10:25 is that all you did with the man page? 2007-11-30 10:26 yeah its just a copy of ddsnap, with the options and name changed 2007-11-30 10:26 :( 2007-11-30 10:26 no, I read through the description and thought how to write it so it doesn't sound like it thinks it is a ddsnap man page 2007-11-30 10:26 but just now I wanted to clear out the title diff without creating yet another out of tree patch for myself 2007-11-30 10:27 so I can concentrate on revert 2007-11-30 10:27 i have them getting htmlized and put on zumastor.org/man 2007-11-30 10:27 but dont link to them off the main page since the zumastor one sucks so badly ;) 2007-11-30 10:27 I see, so the warts are magnified 2007-11-30 10:27 later today 2007-11-30 10:27 I'm not in creative writing mode 2007-11-30 10:28 creative revert mood? 2007-11-30 10:28 right brain switched off 2007-11-30 10:28 left brain has a mandake to be completely obsessive for now 2007-11-30 10:28 mandate 2007-11-30 10:29 hm i have the man pages pulling from the 0.4 branch 2007-11-30 10:29 perhaps thats not right 2007-11-30 10:29 seems right 2007-11-30 10:30 most recent stable version sounds right 2007-11-30 10:30 yeah thats what i was thinking 2007-11-30 10:30 but no one seems to be using the stable branch ;) 2007-11-30 10:30 anyone willing to try zumastor doesn't seem to mind being on the head 2007-11-30 10:36 hm what about the "future addition" in man ddsnap 2007-11-30 10:36 seems a bit dated 2007-11-30 10:37 TERMINOLOGY: ... chunk - a user-definable binary multiple of 4K block size. 2007-11-30 10:38 that makes no sense 2007-11-30 10:43 not much indeed 2007-11-30 10:43 I'm still a coder and not a tech writer this morning ;-) 2007-11-30 10:44 ;) 2007-11-30 10:44 we should release an 0.5 early, with docs 2007-11-30 10:44 and not get too obsessed with that for 0.4, which should have been released a week ago 2007-11-30 10:44 we should officially release 0.4, like right away 2007-11-30 10:44 yes 2007-11-30 10:45 just need to update the main page and freshmeat 2007-11-30 10:49 -!- pgquiles(~pgquiles@81.202.65.108.dyn.user.ono.com) has joined #zumastor 2007-11-30 10:50 hi pgquiles 2007-11-30 11:04 hi 2007-11-30 11:05 I'm preparing linux-*-2.6.22-14-xen+zumastor for Ubuntu Gutsy 2007-11-30 11:06 Gutsy's 2.6.22-14 with zumastor boots fine in the HP DL320s server 2007-11-30 11:06 s 2007-11-30 11:09 good to hear 2007-11-30 11:28 so you patched the gutsy kernel with all the zumastor patches? 2007-11-30 11:48 yes, with zumastor-trunk 2.6.22.10 2007-11-30 11:48 did it apply cleanly? 2007-11-30 11:48 100%, there were no changes in the files which needed patching 2007-11-30 11:49 thats great news 2007-11-30 11:49 unfortunately I won't work on that again until monday 2007-11-30 11:50 I want to implement perl regular expressions in cmake this weekend 2007-11-30 11:54 ACTION blinks 2007-11-30 11:54 Perl REs in cmake?? 2007-11-30 12:02 PCRE, yes 2007-11-30 12:02 regexps in cmake suck 2007-11-30 12:02 and thanks to a google guy PCRE is very easy to use in C++ now 2007-11-30 12:03 Well, I would tend to be surprised that REs would be present in any flavor of make in general... But I guess if you have to have REs, PCRE is probably the most capable. 2007-11-30 12:07 cmake != make 2007-11-30 12:07 I know. 2007-11-30 12:08 Build system. I still think that REs are overkill. But whatever, I'm a kernel guy, living and dying by BSD make... 2007-11-30 12:10 regexps are very useful for the FIND_PACKAGE features in cmake 2007-11-30 13:05 -!- cbsmith(~xman@adsl-76-243-138-218.dsl.lsan03.sbcglobal.net) has joined #zumastor 2007-11-30 13:43 -!- dank(~chatzilla@cpe-76-90-56-73.socal.res.rr.com) has joined #zumastor 2007-11-30 13:43 Shapor? 2007-11-30 13:43 -!- phillips_(~phillips@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-30 13:43 Hi Dan 2007-11-30 13:43 hiya 2007-11-30 13:44 hi dank 2007-11-30 13:44 I just sent email about uploading 0.5 snapshot 2007-11-30 13:44 hi 2007-11-30 13:44 sounds good 2007-11-30 13:45 just dump all the files in /downloads/snapshots, or have a subdir for a specific revision? 2007-11-30 13:46 Oh, same subdir name as you have before, so /downloads/snapshots/0.5-r1002 or whatever. 2007-11-30 13:46 yeah thats what i was thinking 2007-11-30 13:47 So all you see in the snapshots dir is a bunch of directories, one per rev we cared to upload 2007-11-30 13:48 I guess I'll finish the vmware player walkthrough, and link to the current 0.5 snapshot packages. 2007-11-30 13:49 Can you give me an ETA on when you'll have the 0.5 snapshot uploaded and the directories renamed? 2007-11-30 13:51 -!- maximadelman(~maxim@mail.violintech.net) has joined #zumastor 2007-11-30 13:52 hello everyone 2007-11-30 13:54 hi maxim 2007-11-30 13:56 hi maximmadelman 2007-11-30 13:56 short nicks on irc are good 2007-11-30 13:57 i am with violin memory 2007-11-30 13:58 we know :) 2007-11-30 13:58 great 2007-11-30 13:58 I'm Daniel 2007-11-30 13:58 so, how can we be of service 2007-11-30 13:58 shapor is shapor, and jiayingz is jiaying 2007-11-30 13:58 hi maxim 2007-11-30 13:58 cool 2007-11-30 13:58 ok where were we 2007-11-30 13:59 we finally figured out you have 4x1 gige ports in your box over there 2007-11-30 13:59 and that is just fine to establish the "this is a badass server" point 2007-11-30 13:59 we should be able to generate a little over 300 MB/sec NFS traffic with that 2007-11-30 14:00 which gets your machine firmly in badass server territory 2007-11-30 14:00 suitable for making a bar chart 2007-11-30 14:00 shapor, can you name it 'releases' rather than 'stable'? 2007-11-30 14:00 afk for 5 minutes... 2007-11-30 14:01 -!- Brad(~Brad@c-24-3-205-220.hsd1.mn.comcast.net) has joined #zumastor 2007-11-30 14:01 :) well, one would be dedicated to the uplink. we can do that experiment next week if you want 2007-11-30 14:01 still, the capacity of our device would be barely tapped - 1500MB/s+ 2007-11-30 14:02 -!- maximadelman(~maxim@mail.violintech.net) has left #zumastor 2007-11-30 14:02 Daniel is it something that makes as much sense to publish 2007-11-30 14:02 -!- maximadelman(~maxim@mail.violintech.net) has joined #zumastor 2007-11-30 14:02 sorry, got dropped for a second 2007-11-30 14:03 maximadelman: what kind of testing are you doing? testing servers before placing them in production or just testing how good zumastor is? 2007-11-30 14:04 pgquiles: flipz & co are doing the testing, we are merely providing the hardware 2007-11-30 14:05 oh 2007-11-30 14:05 maximadelman: what hardware? 2007-11-30 14:07 violin 1010? 2007-11-30 14:07 yep 2007-11-30 14:07 as well as a server and several client machines 2007-11-30 14:08 what software does it currently run? 2007-11-30 14:09 best ask flipz - its a standard centos5 AFAIK with custom kernel 2007-11-30 14:12 mmm violin 1010 looks interesting although I cannot think of any application for such a large amount of RAM right now in my company or our clients 2007-11-30 14:13 10TB RAM geez 2007-11-30 14:13 well, 1/2 TB per 2U 2007-11-30 14:14 pgquiles: it presents itself as a block device 2007-11-30 14:14 not system memory 2007-11-30 14:14 so basically a very large, fast disk with virtually no seek time 2007-11-30 14:15 oh, I thought it was not intendef for storage 2007-11-30 14:15 intended 2007-11-30 14:15 Nah, they'd have much less use for us if they weren't. 2007-11-30 14:15 what happens when it shutdowns? 2007-11-30 14:18 I mean, is data preserved as in hard disks or destroyed as in RAM? 2007-11-30 14:19 its ram :) so yes, destroyed 2007-11-30 14:19 in april or so they will have automatic save-to-flash 2007-11-30 14:19 yep 2007-11-30 14:19 Essentially a monstrously huge cache. 2007-11-30 14:19 but until then, if you need persistence, need to mirror to disk in one way or another 2007-11-30 14:19 fmayhar: for now 2007-11-30 14:20 although right now, with the right UPS, it's not that dangerous :-) 2007-11-30 14:20 2007-11-30 14:20 There are 40 -60 block replication packages that I have cataloged so far 2007-11-30 14:21 Holy crap! That many? Did you put them on a web page? 2007-11-30 14:21 when I was in my sophomore I built a payphone phonecard using a static RAM to keep data even without power 2007-11-30 14:21 I don't know if they have been replaced by NVRAM and so these days 2007-11-30 14:21 3 most popular, veritas, steeleye and R1soft - may are open source 2007-11-30 14:21 many 2007-11-30 14:22 I'm waiting to test some before we publish 2007-11-30 14:23 the 3 popular ones support windows too, but veritas supports 6 os 2007-11-30 14:24 Who is 'we' (and for that matter, 'you')? Yeah, we're somewhat aware of veritas and the like, but weren't aware of any good free ones for Linux. 2007-11-30 14:28 Thats what We (violin) are trying to sort out. but that will take time - thats why the 3 top ones are higher on the list 2007-11-30 14:28 ah. Yeah, you totally need to evaluate the popular ones running on your box. 2007-11-30 14:28 but even the cheapest one is $600 a server 2007-11-30 14:28 Time for some good free competition, that's what I say :-) 2007-11-30 14:30 exactly, there needs to be a shakeout of these. I really liked this guy who is now at microsoft did - but now he's tat microsoft so no more updates 2007-11-30 14:31 ACTION back 2007-11-30 14:33 What did he do? 2007-11-30 14:34 he's the guy that started pratima 2007-11-30 14:35 hasn't been touched in years 2007-11-30 14:36 then r1soft was just bought out - so up goes the prices 2007-11-30 14:38 i think windows server supports soft mirroring 2007-11-30 14:38 and ldm under linux might support it too 2007-11-30 14:38 brad, 40-60 packages of what? 2007-11-30 14:38 not sure if the writes are delayed, but even if they aren't, it might be a good enough solution if only read performance matters 2007-11-30 14:39 replicators, etc 2007-11-30 14:40 replicators of what to what? 2007-11-30 14:40 rsync type things don't count 2007-11-30 14:42 The first theought was checkpoints 2007-11-30 14:44 the majority of customers though have databases, so within those there are features for table replication and transaction logs 2007-11-30 14:45 but these can also work with disk replicators too, but people haven't fully gotten their teeth into them yet 2007-11-30 14:46 what is the total number of disk replicators you know of then? 2007-11-30 14:46 database replicators don't count either ;) 2007-11-30 14:50 ACTION wonders whether 'flips' is a zumastor-block-level-snapshot of 'flipz' or just a hard-link 2007-11-30 14:50 :-P 2007-11-30 14:51 pgquiles: Mostly a hard-link, although every now and then it appears there are a few blocks still in flight. ;-) 2007-11-30 14:54 goodnight all, tty monday 2007-11-30 14:55 -!- maximadelman(~maxim@mail.violintech.net) has left #zumastor 2007-11-30 14:57 we didn't count the db becuase that is a function of the db, mainly these are derivatives of backup software 2007-11-30 14:59 we also didn't include on our list the possible compatibility with trucopy, snapmirror, SRDF that are hardware dependent but possible solutions 2007-11-30 15:15 -!- fmayhar_(~chatzilla@207.47.98.129.static.nextweb.net) has joined #zumastor 2007-11-30 15:15 -!- fmayhar_(~chatzilla@207.47.98.129.static.nextweb.net) has left #zumastor 2007-11-30 16:51 -!- tnh(~tnh@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor 2007-11-30 16:54 hello 2007-11-30 17:08 -!- tnh(~tnh@cpe-76-90-128-140.socal.res.rr.com) has left #zumastor 2007-11-30 18:58 -!- tnh(~tnh@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor 2007-11-30 19:24 -!- tnh(~tnh@cpe-76-90-128-140.socal.res.rr.com) has left #zumastor 2007-11-30 19:30 -!- Tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor 2007-11-30 19:30 mo betta 2007-11-30 22:15 -!- Tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has joined #zumastor 2007-11-30 22:16 -!- Tim_vimm(~Tim@cpe-76-90-128-140.socal.res.rr.com) has left #zumastor