| [2010/02/01 00:36:26] @ Log started by gepetto | ||
| [2010/02/01 00:36:26] @ github joined channel #puppet-dev | ||
| [2010/02/01 00:36:26] <github> puppet: feature/svn James Turnbull * a1fcf67 (0 files in 0 dirs): Very, very simple Repo type with SVN and Git providers - http://bit.ly/9mfEQT | ||
| [2010/02/01 00:36:26] @ github left channel #puppet-dev () | ||
| [2010/02/01 01:22:51] @ Quit: notbrien: Quit: notbrien | ||
| [2010/02/01 01:32:47] <ohadlevy> how do i debug segfaults with ruby 1.81 and 0.25.4? | ||
| [2010/02/01 01:53:29] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 01:55:17] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 03:24:33] <masterzen> ohadlevy: produce a core, and use gdb to see where it choked in the ruby code... That might help you... | ||
| [2010/02/01 03:31:46] <ohadlevy> masterzen: i guess not much sense to see a pastie without debug info ? | ||
| [2010/02/01 03:32:10] <masterzen> ohadlevy: can you install the symbols for ruby? | ||
| [2010/02/01 03:32:31] <masterzen> ohadlevy: I remember RH usually ships debuginfo packages | ||
| [2010/02/01 03:32:35] <ohadlevy> yeah | ||
| [2010/02/01 03:32:37] <ohadlevy> I'll give it a try | ||
| [2010/02/01 03:34:16] <masterzen> ohadlevy: there's also the possibility to print the ruby stack trace: http://eigenclass.org/hiki/ruby+live+process+introspection | ||
| [2010/02/01 03:34:34] <masterzen> ohadlevy: I already used this once, works most of the time. Not sure after a segfault... | ||
| [2010/02/01 03:35:40] <ohadlevy> masterzen: i'll try first with the debuginfo :) | ||
| [2010/02/01 03:41:29] <ohadlevy> masterzen: http://pastie.org/803887 if you have time :) | ||
| [2010/02/01 03:42:17] <masterzen> ohadlevy: this is a crash in YAML | ||
| [2010/02/01 03:42:35] <ohadlevy> masterzen: how can you tell? | ||
| [2010/02/01 03:43:33] <masterzen> ohadlevy: because I'm sooo good :-) | ||
| [2010/02/01 03:43:45] <masterzen> ohadlevy: seriously: because it calls: syck_emitter_simple_writ | ||
| [2010/02/01 03:43:57] <masterzen> ohadlevy: syck is the other name of YAML | ||
| [2010/02/01 03:44:05] <masterzen> ohadlevy: maybe you get: syck_emitter_simple_writ | ||
| [2010/02/01 03:44:10] <masterzen> ohadlevy: maybe you get: http://redmine.ruby-lang.org/issues/show/990 | ||
| [2010/02/01 03:45:01] <ohadlevy> hmm.. i thought that catalog is transfered in pson.. is it because its writing the catalog in yaml? | ||
| [2010/02/01 03:47:11] <masterzen> certainly | ||
| [2010/02/01 03:47:40] <masterzen> brb, coffee time | ||
| [2010/02/01 03:53:11] <ohadlevy> masterzen: any way to disable yaml completely on a 0.25.4? | ||
| [2010/02/01 04:03:45] @ Quit: smef: Read error: Connection reset by peer | ||
| [2010/02/01 04:14:22] <masterzen> ohadlevy: unfortunately no | ||
| [2010/02/01 04:14:41] <masterzen> ohadlevy: reports are still using yaml, and also the local catalog cache | ||
| [2010/02/01 04:19:27] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 04:19:43] <ohadlevy> masterzen: it fails very quickly, so I assume its the local catalog | ||
| [2010/02/01 04:20:12] <duritong> another + for #2892 | ||
| [2010/02/01 04:20:13] <gepetto> duritong: #2892 is http://projects.reductivelabs.com/issues/show/2892 "Puppet - Bug #2892: Remove unneeded YAML serialization when caching catalog in puppetd - ReductiveLabs.com" | ||
| [2010/02/01 04:20:27] <duritong> ohadlevy: and if you delete the local catalog? does it still fail? | ||
| [2010/02/01 04:20:33] <ohadlevy> hmm let me try | ||
| [2010/02/01 04:22:13] <ohadlevy> doesnt help, it actually segfault right after Caching catalog line.. | ||
| [2010/02/01 04:22:19] <duritong> :( | ||
| [2010/02/01 04:22:47] <ohadlevy> i wonder if there is a way to find out whats in the catalog generates it | ||
| [2010/02/01 04:23:52] <masterzen> ohadlevy: dump the catalog to a file, then try a little bit of ruby to see YAML crash in live | ||
| [2010/02/01 04:24:06] <ohadlevy> hmm. right I forgot about this option :) | ||
| [2010/02/01 04:24:39] <masterzen> ohadlevy: you can remove the yaml local cache: comment out line 40 in lib/puppet/indirector/yaml.rb. | ||
| [2010/02/01 04:24:53] <ohadlevy> thanks | ||
| [2010/02/01 04:24:54] <masterzen> ohadlevy: I don't know what will be the side effects though :-) | ||
| [2010/02/01 04:24:54] <ohadlevy> will try | ||
| [2010/02/01 04:25:13] <masterzen> duritong: yes, another one for #2892. | ||
| [2010/02/01 04:25:14] <gepetto> masterzen: duritong: #2892 is http://projects.reductivelabs.com/issues/show/2892 "Puppet - Bug #2892: Remove unneeded YAML serialization when caching catalog in puppetd - ReductiveLabs.com" | ||
| [2010/02/01 04:25:14] <duritong> at least running against a cache catalog will fail | ||
| [2010/02/01 04:25:51] <duritong> but while debugging #2892 I didn't encounter any other problems | ||
| [2010/02/01 04:25:52] <gepetto> duritong: #2892 is http://projects.reductivelabs.com/issues/show/2892 "Puppet - Bug #2892: Remove unneeded YAML serialization when caching catalog in puppetd - ReductiveLabs.com" | ||
| [2010/02/01 04:25:59] <masterzen> duritong: BTW, I started working on integrating yajl-ruby in streaming mode to deserialiaze the catalog. This should put an end to puppetd memory consumption once 2892 will be fixed. | ||
| [2010/02/01 04:26:10] <duritong> actually one can save a lot of memory with that | ||
| [2010/02/01 04:26:16] <duritong> masterzen: awesome! | ||
| [2010/02/01 04:26:27] <masterzen> that and streaming file source writes on the client... | ||
| [2010/02/01 04:26:39] <masterzen> duritong: I hope to have a patch ready next week. | ||
| [2010/02/01 04:26:58] <masterzen> duritong: so far I have yajl-ruby working, but not the streaming part yet. | ||
| [2010/02/01 04:30:37] <ohadlevy> masterzen: avoiding the write actully works | ||
| [2010/02/01 04:30:40] <duritong> cool :) streaming down files would be a cool improvement as well | ||
| [2010/02/01 04:31:05] <masterzen> ohadlevy: there is certainly something invalid in your yamelized catalog. | ||
| [2010/02/01 04:31:15] <ohadlevy> tell me about it ;) | ||
| [2010/02/01 04:31:38] <ohadlevy> 50% speed improvement to 0.24.8 | ||
| [2010/02/01 04:32:16] <Volcane> hate++ yaml | ||
| [2010/02/01 04:32:21] <ohadlevy> hmm.. I'm not using cached catalogs anyway... maybe I should recreate my own rpms ;) | ||
| [2010/02/01 04:32:35] <ohadlevy> I do it for RHEL3 anyway.. | ||
| [2010/02/01 04:33:09] <Volcane> ohadlevy: do you notice in 0.24.8 that at the end of the run it just sits there for ages? debug+verbose shows me its doing a whole catalog sorting etc again | ||
| [2010/02/01 04:33:32] <ohadlevy> Volcane: not for too long, maybe a second or two | ||
| [2010/02/01 04:33:32] <Volcane> well not sure if it sorts on the client, but it goes through every resource again for sure | ||
| [2010/02/01 04:33:45] <ohadlevy> Volcane: but I have fast machines ;) | ||
| [2010/02/01 04:33:47] <Volcane> you dont use vmware, i can tell :P | ||
| [2010/02/01 04:34:04] <Volcane> 25 doesnt do that | ||
| [2010/02/01 04:34:09] <ohadlevy> I actually do, as KVM has some nasty bugs :( | ||
| [2010/02/01 04:35:48] <Volcane> it takes minutes for me | ||
| [2010/02/01 04:35:55] <Volcane> dont know why, but time to ditch 24 | ||
| [2010/02/01 04:36:46] <ohadlevy> Volcane: if you can get 0.25x. to work ;) | ||
| [2010/02/01 04:37:23] <ohadlevy> masterzen: replace to_yaml to to_pson works nicely too :) | ||
| [2010/02/01 04:37:48] <Volcane> yeah 0.25 working well for me, but have all my clients to upgrade still | ||
| [2010/02/01 04:37:55] <Volcane> i think 0.25.4 is solid | ||
| [2010/02/01 04:38:15] <ohadlevy> masterzen: I'll try to conver the pson to yaml | ||
| [2010/02/01 04:38:51] <ohadlevy> how do I use the pson methods in puppet? whats the equiv of YAML::load ? | ||
| [2010/02/01 04:39:01] <masterzen> ohadlevy: PSON.parse | ||
| [2010/02/01 04:39:13] <masterzen> ohadlevy: and the reverse PSON.generate | ||
| [2010/02/01 04:39:40] <ohadlevy> thanks | ||
| [2010/02/01 04:39:53] <masterzen> ohadlevy: actually converting pson -> yaml is not directly possible, you need to call Catalog.from_pson() then YAML.dump(catalog) | ||
| [2010/02/01 04:42:30] <ohadlevy> masterzen: thanks, I'll try to find out the cause, probably some exec | ||
| [2010/02/01 04:43:01] <duritong> yeah I would suspect them as well | ||
| [2010/02/01 04:43:17] * duritong had some problems with them on 0.25 as well | ||
| [2010/02/01 04:43:50] <duritong> most likely because they're the own using a lot of "special characters", hence need a lot of escaping | ||
| [2010/02/01 04:53:04] <Volcane> using lookupvar() in a function, is it supposed to have access to variables set in the class its called from and also the node? | ||
| [2010/02/01 04:53:08] <Volcane> I am sure this used to work | ||
| [2010/02/01 04:53:14] <Volcane> but might just be wrong | ||
| [2010/02/01 05:15:42] @ nasrat joined channel #puppet-dev | ||
| [2010/02/01 05:31:33] @ Quit: duritong: Remote host closed the connection | ||
| [2010/02/01 05:31:39] @ duritong joined channel #puppet-dev | ||
| [2010/02/01 08:22:52] @ ShiNboi joined channel #puppet-dev | ||
| [2010/02/01 08:23:28] @ ShiNboi_ joined channel #puppet-dev | ||
| [2010/02/01 08:27:56] @ Quit: ShiNboi: Ping timeout: 272 seconds | ||
| [2010/02/01 08:54:11] @ Quit: ShiNboi_: Quit: Verlassend | ||
| [2010/02/01 09:14:32] @ rocket joined channel #puppet-dev | ||
| [2010/02/01 10:14:26] @ notbrien joined channel #puppet-dev | ||
| [2010/02/01 10:15:01] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 10:27:55] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 10:47:06] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 10:56:23] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 11:29:58] @ Quit: fsweetser: Remote host closed the connection | ||
| [2010/02/01 11:56:43] @ fsweetser joined channel #puppet-dev | ||
| [2010/02/01 11:59:54] @ Quit: ppickfor: Ping timeout: 276 seconds | ||
| [2010/02/01 12:01:13] @ mqr joined channel #puppet-dev | ||
| [2010/02/01 12:32:59] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 12:43:17] @ Quit: fsweetser: Remote host closed the connection | ||
| [2010/02/01 12:44:07] <rocket> mqr: Happen to have a chance to look at the code I sent you? I am sure your busy so I am just following up | ||
| [2010/02/01 12:51:09] <mqr> rocket: briefly, but then I got overtaken by events. I'm trying to work through Friday's things to do list this morning. | ||
| [2010/02/01 12:52:15] @ Quit: bje: Ping timeout: 260 seconds | ||
| [2010/02/01 12:54:53] @ bje joined channel #puppet-dev | ||
| [2010/02/01 12:58:30] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 13:00:56] <rocket> mqr: I understand .. I know work. other todo's etc are higher priority .. and I am still trying to figure it out myself .. thanks :) | ||
| [2010/02/01 13:49:02] @ fsweetser joined channel #puppet-dev | ||
| [2010/02/01 14:09:12] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 14:17:28] @ lak joined channel #puppet-dev | ||
| [2010/02/01 14:30:10] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 14:35:39] @ Quit: bje: Read error: Connection reset by peer | ||
| [2010/02/01 14:45:54] @ bje joined channel #puppet-dev | ||
| [2010/02/01 14:46:04] @ Quit: lak: Quit: lak | ||
| [2010/02/01 14:46:31] @ lak joined channel #puppet-dev | ||
| [2010/02/01 16:17:32] @ Quit: lak: Quit: lak | ||
| [2010/02/01 16:34:34] @ lak joined channel #puppet-dev | ||
| [2010/02/01 16:44:53] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 17:03:57] @ jes5 joined channel #puppet-dev | ||
| [2010/02/01 17:09:35] <duritong> wow michael | ||
| [2010/02/01 17:14:34] <duritong> this is definately a cool move | ||
| [2010/02/01 17:14:54] * duritong wondered what big project he is showing up after leaving cobbler ;) | ||
| [2010/02/01 17:18:54] <Volcane> lak: in a function if i use lookupvar() is it supposed to have access to the varaibles in the class thats calling the function, in the node etc? | ||
| [2010/02/01 17:27:11] @ Quit: fsweetser: Remote host closed the connection | ||
| [2010/02/01 17:48:19] <lak> Volcane: yeah, it should | ||
| [2010/02/01 17:48:54] <Volcane> then i have a bug to report - but first test case to write meh | ||
| [2010/02/01 18:10:02] @ Quit: notbrien: Quit: notbrien | ||
| [2010/02/01 18:29:10] @ Quit: nigelk: *.net *.split | ||
| [2010/02/01 18:29:11] @ Quit: lak: *.net *.split | ||
| [2010/02/01 18:33:04] @ lak joined channel #puppet-dev | ||
| [2010/02/01 18:33:27] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 19:25:03] @ Quit: nigelk: Remote host closed the connection | ||
| [2010/02/01 19:25:33] @ nigelk_ joined channel #puppet-dev | ||
| [2010/02/01 19:32:09] @ Quit: lak: Quit: lak | ||
| [2010/02/01 19:38:35] @ Quit: nigelk_: Remote host closed the connection | ||
| [2010/02/01 19:38:39] @ nigelk joined channel #puppet-dev | ||
| [2010/02/01 19:39:13] @ nigelk_ joined channel #puppet-dev | ||
| [2010/02/01 19:39:14] @ Quit: nigelk: Read error: Connection reset by peer | ||
| [2010/02/01 19:39:14] @ nigelk_ is now known as nigelk | ||
| [2010/02/01 19:39:50] @ nigelk_ joined channel #puppet-dev | ||
| [2010/02/01 19:43:41] @ Quit: nigelk: Ping timeout: 260 seconds | ||
| [2010/02/01 19:43:41] @ nigelk_ is now known as nigelk | ||
| [2010/02/01 20:16:23] @ Quit: nigelk: Quit: nigelk | ||
| [2010/02/01 21:00:32] <ohadlevy> mqr: did you have a chance to look at #2604? | ||
| [2010/02/01 21:00:33] <gepetto> ohadlevy: mqr: #2604 is http://projects.reductivelabs.com/issues/show/2604 "Puppet - Bug #2604: Ruby 1.8.1's known bad syck causes core dumps - ReductiveLabs.com" | ||
| [2010/02/01 21:21:36] @ ppickfor joined channel #puppet-dev | ||
| [2010/02/01 21:29:39] @ Quit: ppickfor: Quit: Leaving. | ||
| [2010/02/01 21:40:24] @ ppickfor joined channel #puppet-dev | ||
| [2010/02/01 21:49:49] @ Quit: jes5: Quit: Leaving. | ||
| [2010/02/01 22:14:51] @ Quit: Chiku: Read error: Connection reset by peer | ||
| [2010/02/01 22:15:05] @ Chiku joined channel #puppet-dev | ||
| [2010/02/01 22:40:39] @ notbrien joined channel #puppet-dev |
Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!