| [2010/07/19 00:28:10] @ Log started by gepetto | ||
| [2010/07/19 00:28:10] @ Quit: jmccune: Ping timeout: 260 seconds | ||
| [2010/07/19 00:33:13] @ jmccune joined channel #puppet-dev | ||
| [2010/07/19 00:34:43] @ DavidS1 joined channel #puppet-dev | ||
| [2010/07/19 00:35:57] @ DavidS1 is now known as DavidS | ||
| [2010/07/19 01:17:51] @ Quit: jmccune: Ping timeout: 258 seconds | ||
| [2010/07/19 01:28:35] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 01:29:26] @ Quit: DavidS: Disconnected by services | ||
| [2010/07/19 01:29:27] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 01:35:38] <DavidS> any ideas where the user provider for windows is? It was referenced on wiki:Puppet_Windows | ||
| [2010/07/19 01:35:39] <gepetto> DavidS: wiki:Puppet_Windows is http://projects.puppetlabs.com/projects/puppet/wiki/Puppet_Windows "Puppet - Puppet Windows - Puppet Labs" | ||
| [2010/07/19 01:37:04] @ Quit: stevenjenkins: Ping timeout: 265 seconds | ||
| [2010/07/19 01:37:09] <sejo> will it ever be possible to use a wildcard it a filepath? | ||
| [2010/07/19 01:37:29] <sejo> usecase: set all /home/* folders to 0700 | ||
| [2010/07/19 01:42:23] @ Quit: DavidS: Ping timeout: 265 seconds | ||
| [2010/07/19 01:53:27] @ stevenjenkins joined channel #puppet-dev | ||
| [2010/07/19 01:54:10] <masterzen> sejo: there is a feature request for this I believe. But nobody seems to care enough. | ||
| [2010/07/19 01:56:56] <sejo> masterzen: ok couldn't find the feature request | ||
| [2010/07/19 01:58:16] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 01:58:22] <sejo> found it thx! | ||
| [2010/07/19 01:58:30] <DavidS> sejo: file { "/home": recurse => 1, mode => 0700 } | ||
| [2010/07/19 01:59:29] <masterzen> DavidS: recurse => number is deprecated, prefer recurse => true, recurselimit => 1 | ||
| [2010/07/19 01:59:41] <sejo> cool thanks! | ||
| [2010/07/19 02:00:43] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 02:02:05] @ Quit: DavidS: Disconnected by services | ||
| [2010/07/19 02:02:08] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 02:02:34] <DavidS> masterzen: thx | ||
| [2010/07/19 02:07:48] <DavidS> any ideas where the user provider for windows is? It was referenced on wiki:Puppet_Windows (missed any answers thanks to my shitty dsl) | ||
| [2010/07/19 02:07:48] <gepetto> DavidS: wiki:Puppet_Windows is http://projects.puppetlabs.com/projects/puppet/wiki/Puppet_Windows "Puppet - Puppet Windows - Puppet Labs" | ||
| [2010/07/19 02:18:49] <masterzen> DavidS: I thought you were supposed to do the windows sutff? | ||
| [2010/07/19 02:24:45] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 02:26:07] @ Quit: DavidS: Ping timeout: 245 seconds | ||
| [2010/07/19 02:26:55] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 02:27:29] <DavidS> masterzen: I am, but that page is from before my times... | ||
| [2010/07/19 02:28:23] <masterzen> DavidS: oh, I see. | ||
| [2010/07/19 02:28:50] <DavidS> and since i haven't found that code in master, I'm beginning to wonder ... | ||
| [2010/07/19 02:29:21] <masterzen> DavidS: check this: http://github.com/finalprefix/puppet/tree/win | ||
| [2010/07/19 02:29:42] <masterzen> DavidS: I found the url from an old discussion in the puppet-dev list. I thought we merged-in the code though | ||
| [2010/07/19 02:30:08] <DavidS> I'll take a look | ||
| [2010/07/19 02:30:14] @ Quit: DavidS2: Ping timeout: 265 seconds | ||
| [2010/07/19 02:32:27] <masterzen> DavidS: check this: http://www.mail-archive.com/puppet-dev@googlegroups.com/msg03391.html | ||
| [2010/07/19 02:33:52] <DavidS> Yeah, that's definitely not merged | ||
| [2010/07/19 02:40:15] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 02:40:42] @ Quit: DavidS: Disconnected by services | ||
| [2010/07/19 02:40:44] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 03:03:07] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 03:04:21] @ Quit: DavidS: Disconnected by services | ||
| [2010/07/19 03:04:29] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 04:07:50] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 04:10:21] @ Quit: DavidS: Ping timeout: 252 seconds | ||
| [2010/07/19 04:41:42] @ Quit: DavidS2: Ping timeout: 265 seconds | ||
| [2010/07/19 04:41:48] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 05:04:38] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 05:06:57] @ Quit: DavidS: Ping timeout: 245 seconds | ||
| [2010/07/19 05:09:53] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 05:12:47] @ Quit: DavidS2: Ping timeout: 245 seconds | ||
| [2010/07/19 05:14:26] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 05:16:57] @ Quit: DavidS: Ping timeout: 245 seconds | ||
| [2010/07/19 05:37:37] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 05:46:20] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 05:48:16] @ Quit: DavidS: Ping timeout: 265 seconds | ||
| [2010/07/19 05:48:35] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 05:51:47] @ Quit: DavidS2: Ping timeout: 265 seconds | ||
| [2010/07/19 05:54:06] @ Quit: DavidS: Read error: Connection reset by peer | ||
| [2010/07/19 05:54:42] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 06:16:22] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 06:19:00] @ Quit: DavidS: Ping timeout: 240 seconds | ||
| [2010/07/19 06:19:26] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 06:22:09] @ Quit: DavidS2: Ping timeout: 240 seconds | ||
| [2010/07/19 06:23:07] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 06:26:03] @ Quit: DavidS: Ping timeout: 258 seconds | ||
| [2010/07/19 06:38:15] @ DavidS joined channel #puppet-dev | ||
| [2010/07/19 06:41:23] @ Quit: DavidS2: Ping timeout: 276 seconds | ||
| [2010/07/19 06:44:34] @ rcrowley joined channel #puppet-dev | ||
| [2010/07/19 06:44:59] @ DavidS2 joined channel #puppet-dev | ||
| [2010/07/19 06:48:16] @ Quit: DavidS: Ping timeout: 276 seconds | ||
| [2010/07/19 08:18:10] @ jmccune joined channel #puppet-dev | ||
| [2010/07/19 08:19:37] @ notbrien joined channel #puppet-dev | ||
| [2010/07/19 08:29:23] @ Quit: jmccune: Ping timeout: 260 seconds | ||
| [2010/07/19 09:30:13] @ lak joined channel #puppet-dev | ||
| [2010/07/19 09:31:10] @ DavidS2 is now known as DavidS | ||
| [2010/07/19 09:31:23] @ Quit: rcrowley: Quit: rcrowley | ||
| [2010/07/19 09:44:28] @ jmccune joined channel #puppet-dev | ||
| [2010/07/19 09:52:29] @ jmccune_ joined channel #puppet-dev | ||
| [2010/07/19 09:53:47] <duritong> jamesturnbull: just to let you know: so far I only tested a rather small 0.25.5 client against 2.6 and I will test the rest when problems have worked out in that small setup | ||
| [2010/07/19 09:54:35] <jamesturnbull> duritong: thanks | ||
| [2010/07/19 10:00:15] <duritong> jamesturnbull: one of the issues that imho should be fixed before .0 is that undef variables are not anymore empty strings | ||
| [2010/07/19 10:00:33] <jamesturnbull> duritong: which ticket is that? | ||
| [2010/07/19 10:01:30] @ rcrowley joined channel #puppet-dev | ||
| [2010/07/19 10:01:54] <duritong> jamesturnbull: #4269 | ||
| [2010/07/19 10:01:54] <gepetto> duritong: jamesturnbull: #4269 is http://projects.puppetlabs.com/issues/show/4269 "Puppet - Bug #4269: behavior change in 2.6: undef variables are not anymore empty strings - Puppet Labs" | ||
| [2010/07/19 10:03:54] <jamesturnbull> duritong: thanks will look | ||
| [2010/07/19 10:05:28] <duritong> (I had the feeling that you guys are going for a final release and wanted to clarify that I couldn't yet test everything) | ||
| [2010/07/19 10:05:42] @ Quit: lak: Quit: lak | ||
| [2010/07/19 10:06:33] <jamesturnbull> it's okay | ||
| [2010/07/19 10:06:38] <jamesturnbull> if we do .1 RSN after | ||
| [2010/07/19 10:06:43] <jamesturnbull> then I am okay with that | ||
| [2010/07/19 10:06:50] <duritong> yeah | ||
| [2010/07/19 10:07:29] <DavidS> jamesturnbull: as i just said on #puppet, I'm currently preparing a patchset to fix a few things on the windows side (and fix approx 400 of 600 specs on windows) | ||
| [2010/07/19 10:07:40] <DavidS> I'll expect to send it within the next few hours | ||
| [2010/07/19 10:07:47] <DavidS> s/'ll// | ||
| [2010/07/19 10:07:50] <jamesturnbull> it won't make 2.6.0 | ||
| [2010/07/19 10:07:59] <jamesturnbull> just so you know | ||
| [2010/07/19 10:10:30] <DavidS> ok | ||
| [2010/07/19 10:30:18] @ plathrop-away is now known as plathrop | ||
| [2010/07/19 10:45:11] @ Quit: rcrowley: Ping timeout: 260 seconds | ||
| [2010/07/19 10:46:54] @ rcrowley joined channel #puppet-dev | ||
| [2010/07/19 10:57:49] @ dcarley joined channel #puppet-dev | ||
| [2010/07/19 11:00:31] <dcarley> hi | ||
| [2010/07/19 11:00:38] <dcarley> jeff> are you around? | ||
| [2010/07/19 11:02:32] <jamesturnbull> dcarley: I'll get him | ||
| [2010/07/19 11:02:59] <dcarley> thanks | ||
| [2010/07/19 11:03:24] <jamesturnbull> dcarley: okay he stepped out but he's apparently replicated your problem | ||
| [2010/07/19 11:03:50] <dcarley> ah, good. | ||
| [2010/07/19 11:04:48] <dcarley> i noticed he asked for the config i was testing with. it's nothing out of the ordinary, but here if still required - http://pastie.org/1050755 | ||
| [2010/07/19 11:05:33] <jamesturnbull> dcarley: thanks! | ||
| [2010/07/19 11:06:31] <dcarley> i also noticed that it seems to auto-import a module from the prod env, if available, as if the modulepath had been prefixed. | ||
| [2010/07/19 11:07:15] <jamesturnbull> yeah I think Jeff saw that when he replicated | ||
| [2010/07/19 11:08:35] @ nanliu joined channel #puppet-dev | ||
| [2010/07/19 11:08:37] <dcarley> cool. | ||
| [2010/07/19 11:10:41] <jamesturnbull> duritong: re 4232 - so ALL clients have NTP - failing and not? | ||
| [2010/07/19 11:12:51] <jamesturnbull> duritong: I am sensing a Webrick SSL bug | ||
| [2010/07/19 11:13:29] <duritong> duritong: yes all clocks are synced | ||
| [2010/07/19 11:13:41] <duritong> jamesturnbull: that was meant for you | ||
| [2010/07/19 11:13:50] <duritong> jamesturnbull: yes I'm also smelling such a bug | ||
| [2010/07/19 11:14:00] * jamesturnbull hates webrick | ||
| [2010/07/19 11:14:06] * duritong food | ||
| [2010/07/19 11:26:02] @ Quit: DavidS: Read error: Connection reset by peer | ||
| [2010/07/19 11:26:41] @ lak joined channel #puppet-dev | ||
| [2010/07/19 12:45:39] <jamesturnbull> duritong: ping | ||
| [2010/07/19 13:10:10] @ joe-mac joined channel #puppet-dev | ||
| [2010/07/19 13:10:24] <joe-mac> having a hell of a time getting my custom function to work, how can i debug it to see exactly what'sg oing on? | ||
| [2010/07/19 13:12:16] <Volcane> hard to debug :( | ||
| [2010/07/19 13:12:26] <Volcane> add Puppet.debug("foo") style lines | ||
| [2010/07/19 13:12:35] <Volcane> actually Puppet.notice() probably | ||
| [2010/07/19 13:13:16] <joe-mac> ugh and you have to restart the master everything | ||
| [2010/07/19 13:13:17] <joe-mac> totally suck | ||
| [2010/07/19 13:14:47] <Volcane> you can debug the stuff with puppet | ||
| [2010/07/19 13:14:57] <Volcane> and a simple test.pp | ||
| [2010/07/19 13:15:09] <Volcane> it supports calling parser functions | ||
| [2010/07/19 13:15:34] <joe-mac> so if i run that it will automatically find the plugin in modules? | ||
| [2010/07/19 13:15:53] <Volcane> mkdir foo/puppet/parser/functions -p | ||
| [2010/07/19 13:15:59] <Volcane> put the file in there | ||
| [2010/07/19 13:16:03] <Volcane> export RUBYLIB=foo | ||
| [2010/07/19 13:16:11] <Volcane> and it should just work roughly | ||
| [2010/07/19 13:16:13] <joe-mac> that wow awesome | ||
| [2010/07/19 13:16:23] <joe-mac> it worked immediately Volcane | ||
| [2010/07/19 13:16:25] <joe-mac> saw tehj actual error | ||
| [2010/07/19 13:16:26] <Volcane> been a while since i did it but i think thats it | ||
| [2010/07/19 13:16:29] <Volcane> heh nice | ||
| [2010/07/19 13:17:46] <joe-mac> hm, looks like this erb template i am compiling is not coming together correctly | ||
| [2010/07/19 13:21:19] <joe-mac> you rock | ||
| [2010/07/19 13:21:30] <joe-mac> wish i had known that puppet test.pp wopuld ebven test customf unctions without reloading | ||
| [2010/07/19 13:21:39] <joe-mac> maybe i will write one to query ad now lol | ||
| [2010/07/19 13:21:42] <Volcane> puppet standalone can do most things | ||
| [2010/07/19 13:21:50] <Volcane> even external nodes and exported resources | ||
| [2010/07/19 13:22:53] <joe-mac> LEET DUDE | ||
| [2010/07/19 13:23:05] <joe-mac> i now have a function that automagically writes out node info to our wiki | ||
| [2010/07/19 13:23:15] <joe-mac> gotta expand on it to make the formatting right and show all the facts | ||
| [2010/07/19 13:23:16] <Volcane> nice | ||
| [2010/07/19 13:35:37] <duritong> jamesturnbull: pong | ||
| [2010/07/19 13:35:44] <duritong> jamesturnbull: but on the phone | ||
| [2010/07/19 13:36:25] <jamesturnbull> duritong: there is a fix for #4269 on the -dev list if you have a chance to look please | ||
| [2010/07/19 13:36:25] <gepetto> jamesturnbull: duritong: #4269 is http://projects.puppetlabs.com/issues/show/4269 "Puppet - Bug #4269: behavior change in 2.6: undef variables are not anymore empty strings - Puppet Labs" | ||
| [2010/07/19 13:36:59] <duritong> jamesturnbull: yeah will soon | ||
| [2010/07/19 13:37:35] <dcarley> wow, i didn't know about ruby-debug. | ||
| [2010/07/19 13:37:49] <dcarley> that's much easier than splatterings of Puppet.debug() | ||
| [2010/07/19 14:10:20] <joe-mac> hm | ||
| [2010/07/19 14:10:25] <joe-mac> anybody know why this wouldn't work http://www.pastie.org/1051052 | ||
| [2010/07/19 14:10:45] <joe-mac> says undefined local variable 'iface' | ||
| [2010/07/19 14:10:53] <joe-mac> yet iface is created by the .each do block | ||
| [2010/07/19 14:25:02] <eric0> i'm not sure that erb preserves loop scope across <% %> blocks | ||
| [2010/07/19 14:25:33] <Volcane> it does | ||
| [2010/07/19 14:37:23] <joe-mac> yea, i could have sworn i've used that before | ||
| [2010/07/19 14:42:15] <duritong> jamesturnbull: looks good | ||
| [2010/07/19 14:42:32] <jamesturnbull> duritong: thanks - are all your issues solved for now? | ||
| [2010/07/19 14:42:47] <duritong> except the autoimport I think so | ||
| [2010/07/19 14:42:51] <duritong> letme check | ||
| [2010/07/19 14:45:26] <duritong> #4253 got postponed, which is fine as I think (or hope) it will not affect masters started by daemon processes | ||
| [2010/07/19 14:45:26] <gepetto> duritong: #4253 is http://projects.puppetlabs.com/issues/show/4253 "Puppet - Bug #4253: puppetmaster started in a non accessible directory for the puppet user causing problems - Puppet Labs" | ||
| [2010/07/19 14:46:29] <duritong> jamesturnbull: #4242 is not really a bug and the bug report is already quite some documentation with even the proper solutions | ||
| [2010/07/19 14:46:29] <gepetto> duritong: jamesturnbull: #4242 is http://projects.puppetlabs.com/issues/show/4242 "Puppet - Bug #4242: behavior change in recursive inclusions - Puppet Labs" | ||
| [2010/07/19 14:47:21] <duritong> jamesturnbull: so what remains is #4268 which is a bit ugly but the first round patch already reduced the problem a lot | ||
| [2010/07/19 14:47:22] <gepetto> duritong: jamesturnbull: #4268 is http://projects.puppetlabs.com/issues/show/4268 "Puppet - Bug #4268: puppet master 2.6 always imports manifests newly - Puppet Labs" | ||
| [2010/07/19 14:48:22] <duritong> jamesturnbull: I'm probably now to switch my master to 2.6 and start testing the other clients | ||
| [2010/07/19 14:48:34] <duritong> and then starting to test with 2.6 clients | ||
| [2010/07/19 14:48:55] <jamesturnbull> okay sounds good | ||
| [2010/07/19 14:49:24] <duritong> will see how much these autoimport will affect performance | ||
| [2010/07/19 14:51:06] <duritong> but currently a client with 1.2K resources take 40s to compile | ||
| [2010/07/19 14:51:11] <duritong> which used to have 4s | ||
| [2010/07/19 14:51:19] <duritong> (after everything have been imported) | ||
| [2010/07/19 15:00:07] <jamesturnbull> duritong: *sighs* that's not good | ||
| [2010/07/19 15:00:13] <jamesturnbull> performance sprint for 2.6.1 I think | ||
| [2010/07/19 15:01:11] * duritong runs the first 0.25.5 client agains mongrel master | ||
| [2010/07/19 15:02:11] <duritong> hmm | ||
| [2010/07/19 15:03:07] <duritong> first it couldn't find a class which later it could | ||
| [2010/07/19 15:03:16] <duritong> strange | ||
| [2010/07/19 15:10:01] @ Quit: notbrien: Quit: notbrien | ||
| [2010/07/19 15:11:38] <duritong> jamesturnbull: yeah performance of compiling degraded in range of 7-10 | ||
| [2010/07/19 15:11:56] <duritong> but this might be because things are always re-imported | ||
| [2010/07/19 15:12:05] <duritong> I'm curious why I'm the only person seeing that | ||
| [2010/07/19 15:14:01] <duritong> #4242 hits me now as well in other place | ||
| [2010/07/19 15:14:01] <gepetto> duritong: #4242 is http://projects.puppetlabs.com/issues/show/4242 "Puppet - Bug #4242: behavior change in recursive inclusions - Puppet Labs" | ||
| [2010/07/19 15:22:28] <jamesturnbull> duritong: so any joy at tracking which method is causing this? | ||
| [2010/07/19 15:23:02] <jamesturnbull> duritong: obviously #4268 helped a bit - did it expose in your discussions with Markus where else it might be? | ||
| [2010/07/19 15:23:02] <gepetto> jamesturnbull: duritong: #4268 is http://projects.puppetlabs.com/issues/show/4268 "Puppet - Bug #4268: puppet master 2.6 always imports manifests newly - Puppet Labs" | ||
| [2010/07/19 15:26:15] <markus> duritong: I think the remaining performance is due to the reimporting, and that seems to be related to it having multiple copies of the "production" environment (one per thread). Does that seem to describe it? | ||
| [2010/07/19 15:29:37] <jamesturnbull> duritong: you use mongrel right? | ||
| [2010/07/19 15:30:38] <duritong> jamesturnbull: yes using mongrel, but I can observe the problem as well with webrick | ||
| [2010/07/19 15:30:50] <joe-mac> hey Volcane | ||
| [2010/07/19 15:30:55] <duritong> markus: yeah I also think it's due to the importing | ||
| [2010/07/19 15:31:06] <joe-mac> when i try to use extlookup from my custom function, it returns a caharcter as some kind of int | ||
| [2010/07/19 15:31:14] <duritong> markus: but I don't fully understand what you mean with per thread | ||
| [2010/07/19 15:31:19] <joe-mac> what method do i call on it to get it back into a regular string, .to_s? | ||
| [2010/07/19 15:33:23] <duritong> markus: I use the config I sent you within that modules package | ||
| [2010/07/19 15:33:49] <markus> (looking) | ||
| [2010/07/19 15:34:15] <duritong> and ran it with webrick | ||
| [2010/07/19 15:35:07] <duritong> and all the environments are commented out | ||
| [2010/07/19 15:35:38] * lak wakes up and starts looking at this | ||
| [2010/07/19 15:35:57] <lak> duritong: if this example code isn't on the ticket, can you send it to me, too? | ||
| [2010/07/19 15:36:13] <lak> the code that's doing this is most likely my fault, so i should look at it, too | ||
| [2010/07/19 15:36:58] <markus> I can send it to you. | ||
| [2010/07/19 15:37:09] <duritong> lak: I pasted the url into the query | ||
| [2010/07/19 15:37:26] <duritong> lak: and you can comment out the stored configs stuff | ||
| [2010/07/19 15:37:36] <duritong> it works also without it | ||
| [2010/07/19 15:38:00] <markus> The key aspect is that each run gets a new environment object and thus a new type_collection and thus reimports everything. | ||
| [2010/07/19 15:38:39] <markus> I believe this is because the first & second request are running in separate threads, but have not traced it down further. | ||
| [2010/07/19 15:39:03] <joe-mac> it's returning a FixNum Volcane, you ever done this before? | ||
| [2010/07/19 15:39:19] <markus> I'm going to email you the output from when duritong ran it with the diagnostics on the ticket. | ||
| [2010/07/19 15:39:54] <duritong> markus: are things also running in a seperate thread even with webrick? | ||
| [2010/07/19 15:40:05] <markus> hmmm. | ||
| [2010/07/19 15:40:21] <duritong> this is what confuses we me with webrick | ||
| [2010/07/19 15:40:28] <markus> lak: can you propose and alternate mechanism then? | ||
| [2010/07/19 15:40:46] <duritong> for mongrel it would be clear, as each instance is even an own process | ||
| [2010/07/19 15:41:02] <lak> markus: looking at it now, i haven't quite gotten it running | ||
| [2010/07/19 15:41:03] <markus> Yeah, I see your confusion and raise a befuddlement: if it is, why don't more people see it? | ||
| [2010/07/19 15:41:36] <markus> lak: If you can reproduce the failure, there will be general rejoicing. | ||
| [2010/07/19 15:42:02] <Volcane> joe-mac: if you're using it in a template then just .to_s | ||
| [2010/07/19 15:42:19] <joe-mac> yea it';s in a template inside a puppet function | ||
| [2010/07/19 15:42:49] <Volcane> joe-mac: or just quote your data in the csv | ||
| [2010/07/19 15:42:53] <lak> it's just slow, right, it's not actually a failure? | ||
| [2010/07/19 15:43:00] <joe-mac> o fuck that's probably causing it Volcane | ||
| [2010/07/19 15:43:16] <duritong> lak: no not a failure, it's "just" 10 times slower | ||
| [2010/07/19 15:43:22] <joe-mac> hmm no it is quoted | ||
| [2010/07/19 15:43:39] <lak> trust me - i consider that a show-stopper, i just mean that i'm not looking for an exception | ||
| [2010/07/19 15:44:22] <duritong> lak: yeah, I understood it that way | ||
| [2010/07/19 15:44:38] <joe-mac> http://www.pastie.org/1051212 | ||
| [2010/07/19 15:44:56] <markus> duritong: An additional test if you're up for it; rather than two runs, try 20+ and see if it every stops reimporting (to shoot a hole in my per thread theory). | ||
| [2010/07/19 15:45:12] <markus> s/every/ever/ | ||
| [2010/07/19 15:45:13] <joe-mac> any idea why it's returning fixnums from Volcane's extlookup function? | ||
| [2010/07/19 15:45:45] <Volcane> joe-mac: if you're going to call it direct from another function or a template it will do that | ||
| [2010/07/19 15:46:12] <joe-mac> how do you get around it? | ||
| [2010/07/19 15:46:26] <Volcane> joe-mac: its just the csv's behavior, i guess i could coerce every result into a string in the extlookup code though | ||
| [2010/07/19 15:46:38] <joe-mac> hmm | ||
| [2010/07/19 15:47:01] <joe-mac> that would entitle adding .to_s to each returned result in your code/ | ||
| [2010/07/19 15:47:11] <duritong> markus: I can do that | ||
| [2010/07/19 15:47:21] <joe-mac> entail even is the word i was looking for | ||
| [2010/07/19 15:47:28] <Volcane> joe-mac: just do that in your function | ||
| [2010/07/19 15:47:35] <joe-mac> i do, still gives me numbers | ||
| [2010/07/19 15:47:49] <Volcane> then you're doing it wrong :) | ||
| [2010/07/19 15:47:58] <lak> i'm having another problem - the client side is trying to look up some of the type information it's getting, and that's throwing an exception | ||
| [2010/07/19 15:48:05] <lak> i ran into this problem once before and thought it was fixed | ||
| [2010/07/19 15:48:13] <markus> ah. | ||
| [2010/07/19 15:48:13] <lak> but i'll focus on the reimporting issues on the server for now | ||
| [2010/07/19 15:48:15] <joe-mac> Volcane: | ||
| [2010/07/19 15:48:19] <joe-mac> shaking fist! | ||
| [2010/07/19 15:48:34] <joe-mac> | Location \ | <%= function_extlookup("address").to_s %>, <%= function_extlookup("citystate").to_s %> \ | | ||
| [2010/07/19 15:48:41] <joe-mac> this is what it looks like in confluence when i do that | ||
| [2010/07/19 15:48:45] <markus> lak: how would you characterize it? (e.g. "this problem" is...") | ||
| [2010/07/19 15:48:50] <joe-mac> Location 100, 105 | ||
| [2010/07/19 15:49:19] <Volcane> joe-mac: so whats wrong? | ||
| [2010/07/19 15:49:30] <lak> markus: the problem is that the client should be treating types as simple strings, but instead is trying to load them from disk (e.g., defined resource types and classes) | ||
| [2010/07/19 15:49:58] <markus> lak: ?? The client? | ||
| [2010/07/19 15:50:58] <markus> I agree that we may have / probably have a problem there, but isn't the present problem entirely server side? | ||
| [2010/07/19 15:51:15] <joe-mac> Volcane: the address and citystate in extdata are street address and city, state | ||
| [2010/07/19 15:51:18] <joe-mac> like, real strings | ||
| [2010/07/19 15:51:23] <joe-mac> not numbers pretending to be strings :-P | ||
| [2010/07/19 15:51:40] <Volcane> joe-mac: then i suspect you have something else going on :) | ||
| [2010/07/19 15:51:59] <lak> markus: this is for a succesful catalog compile on the server (i.e., no exceptions on the server): http://pastie.org/1051229 | ||
| [2010/07/19 15:52:02] <Volcane> joe-mac: weird data somewhere or something like that, or , in your csv file | ||
| [2010/07/19 15:52:10] <Volcane> joe-mac: address,some,address | ||
| [2010/07/19 15:52:12] <lak> notice that the stack trace leads to showing the client trying to load types via import | ||
| [2010/07/19 15:52:15] <Volcane> joe-mac: that returns [some, address] | ||
| [2010/07/19 15:52:27] <Volcane> joe-mac: address,"some,address" returns a string | ||
| [2010/07/19 15:52:36] <lak> markus: like i said, focusing on the other bits for now, but i will look at that next | ||
| [2010/07/19 15:52:56] <joe-mac> hm Volcane would doing .flatten.to_s maybe help then? | ||
| [2010/07/19 15:53:08] <joe-mac> i do have a couple of multi entry variables ni there | ||
| [2010/07/19 15:53:14] <joe-mac> i loop through them to build fw rules and such | ||
| [2010/07/19 15:53:22] <Volcane> joe-mac: quote it like i showed | ||
| [2010/07/19 15:53:39] <markus> lak: interesting. Jesse had seen that at one point when we were working on something else, but when we went back we couldn't reproduce it. | ||
| [2010/07/19 15:54:07] <lak> like i said, i've seen it and fixed it before | ||
| [2010/07/19 15:54:29] <markus> But how could that be causing the problem server side? Or is it an unrelated observation? | ||
| [2010/07/19 15:54:32] <joe-mac> hmm idk Volcane, look citystate,"Boston MA" | ||
| [2010/07/19 15:54:38] <Volcane> joe-mac: but really, not something for #puppet-dev, should ask this kind of thing in #puppet | ||
| [2010/07/19 15:54:38] <joe-mac> i am quoting around that | ||
| [2010/07/19 15:54:47] <joe-mac> ok, let me move it over there, sorry guys | ||
| [2010/07/19 15:55:16] <lak> but clearly actually not :/ | ||
| [2010/07/19 15:55:23] <lak> markus: it's unrelated | ||
| [2010/07/19 15:56:20] <markus> lak: the remnants of my sanity thank you. I was going to be very worried if you said they were related. | ||
| [2010/07/19 15:56:29] <lak> nope | ||
| [2010/07/19 15:57:28] <lak> duritong: what are you seeing as compile times for those test manifests? | ||
| [2010/07/19 16:07:39] <lak> markus: ok, initial testing shows that classes, at least, are actually getting reloaded, and lost between runs | ||
| [2010/07/19 16:08:29] <duritong> lak: 2s for that example | ||
| [2010/07/19 16:09:16] <duritong> lak: but for a bigger example with 1.2k resources (which a modest example in our environment) I get 40s | ||
| [2010/07/19 16:09:28] <duritong> and the bigger ones even up to 110s | ||
| [2010/07/19 16:09:59] <markus> lak: So what we saw with the diagnostics was that they weren't getting lost so much as a new "producion" environment was being created (new object_id) and thus there was a new (empty) type_collection. | ||
| [2010/07/19 16:10:01] <lak> duritong: hmm, so that sample doesn't really show the performance degradation? | ||
| [2010/07/19 16:10:08] <lak> markus: meh, same diff :) | ||
| [2010/07/19 16:10:20] <markus> Also, it wasn't the "stale" test that was doing it. | ||
| [2010/07/19 16:10:33] <lak> duritong: are you seeing drastic performance differences between the first runs in 0.25 and 2.6rc4? | ||
| [2010/07/19 16:10:36] <duritong> markus: you might be right with hitting more than ones cool down things, as I'm getting nodes on the mongrel environemnt with 0.02s comile time | ||
| [2010/07/19 16:10:42] <lak> that is, the ones where they theoretically do the initial import | ||
| [2010/07/19 16:10:44] <markus> lak: it's the same as long as you have plenty of ram. | ||
| [2010/07/19 16:11:10] <markus> duritong: Cool! | ||
| [2010/07/19 16:11:28] <duritong> lak: well, the perfomance degradation can be seen as it doesn't improve over the second run and I still get the autoimport messages | ||
| [2010/07/19 16:12:06] <markus> duritong: how many runs in do you start seeing the faster / no-re-import cases? | ||
| [2010/07/19 16:12:23] <lak> duritong: right, but my question is, is the initial import *also* worse in 2.6? | ||
| [2010/07/19 16:12:34] <lak> or just the later runs which shouldn't be doing imports? | ||
| [2010/07/19 16:13:10] <duritong> lak: comparing to 0.25.5 I see a degradation of around 6 times on the initial compile | ||
| [2010/07/19 16:13:16] <lak> awesome :/ | ||
| [2010/07/19 16:13:26] <lak> so slower on initial compile, and then really bad on later compiles | ||
| [2010/07/19 16:13:48] <markus> dratsab! | ||
| [2010/07/19 16:13:50] <duritong> markus: I need to go through the log, but first help a bit in the kitchen | ||
| [2010/07/19 16:14:06] <markus> duritong: I understand. | ||
| [2010/07/19 16:14:18] <markus> What TZ are you in at present? | ||
| [2010/07/19 16:17:36] <duritong> markus: I'm in NYC currently | ||
| [2010/07/19 16:17:54] <duritong> so +3 to portland I think | ||
| [2010/07/19 16:18:13] <duritong> EDT UTC+5 | ||
| [2010/07/19 16:18:39] <duritong> I'll be awake some time | ||
| [2010/07/19 16:18:44] <jhelwig> Yeah, that's PDX +3. | ||
| [2010/07/19 16:18:54] <lak> duritong: thanks for the help, i'll check back as i have results | ||
| [2010/07/19 16:19:15] <duritong> lak: cool, thank you. and if I can do anything, letme know | ||
| [2010/07/19 16:27:05] <lak> duritong: will do | ||
| [2010/07/19 16:57:41] @ plathrop is now known as plathrop-away | ||
| [2010/07/19 17:20:02] @ plathrop-away is now known as plathrop | ||
| [2010/07/19 17:35:02] @ Quit: jmccune: Ping timeout: 252 seconds | ||
| [2010/07/19 17:35:38] @ Quit: jmccune_: Ping timeout: 240 seconds | ||
| [2010/07/19 17:39:46] @ beata joined channel #puppet-dev | ||
| [2010/07/19 17:51:42] @ plathrop is now known as plathrop-away | ||
| [2010/07/19 17:54:02] @ plathrop-away is now known as plathrop | ||
| [2010/07/19 17:59:18] <markus> duritong: feel like another round of diagnostics? | ||
| [2010/07/19 17:59:55] <duritong> markus: in 30 min.? | ||
| [2010/07/19 18:00:02] <duritong> we're currently eating | ||
| [2010/07/19 18:00:11] <markus> Cool. I'll work up a patch to send you. | ||
| [2010/07/19 18:00:14] <markus> Enjoy! | ||
| [2010/07/19 18:00:25] <duritong> ok, thanks | ||
| [2010/07/19 18:04:31] @ Quit: rcrowley: Quit: rcrowley | ||
| [2010/07/19 18:10:17] <markus> duritong: we got further; luke has a patch on the puppet-dev list that should just fix it. | ||
| [2010/07/19 18:11:16] <markus> Could you a) try that, to see if it does, and b) see how much (if any) of the performance problems remain after the repeated import problem is gone? | ||
| [2010/07/19 18:11:55] <markus> Of course, if it _doesn't_ fix it we want to know that too. We may not want to hear it, but we want to know it. | ||
| [2010/07/19 18:17:48] @ jmccune joined channel #puppet-dev | ||
| [2010/07/19 18:17:52] @ jmccune__ joined channel #puppet-dev | ||
| [2010/07/19 18:21:12] @ plathrop is now known as plathrop-away | ||
| [2010/07/19 18:30:02] @ plathrop-away is now known as plathrop | ||
| [2010/07/19 18:56:28] @ rcrowley joined channel #puppet-dev | ||
| [2010/07/19 18:58:39] <markus> duritong: I'm about to build an rc5 with Luke's fix. Have you had a chance to test it | ||
| [2010/07/19 18:58:40] <markus> ? | ||
| [2010/07/19 19:05:34] @ Quit: lak: Quit: lak | ||
| [2010/07/19 19:07:18] @ Quit: nanliu: Quit: nanliu | ||
| [2010/07/19 19:24:33] @ plathrop is now known as plathrop-away | ||
| [2010/07/19 19:29:29] @ Quit: rcrowley: Quit: rcrowley | ||
| [2010/07/19 19:33:30] @ Quit: jmccune__: Quit: leaving | ||
| [2010/07/19 19:35:09] <duritong> markus: testing | ||
| [2010/07/19 19:35:17] <duritong> sorry, took a bit longer | ||
| [2010/07/19 19:35:25] <markus> duritong: Thanks! | ||
| [2010/07/19 19:45:54] <duritong> markus: the second autoimport message went away and compiling is much faster | ||
| [2010/07/19 19:46:06] <markus> Woot! | ||
| [2010/07/19 19:46:08] <duritong> the second compiling | ||
| [2010/07/19 19:46:17] <markus> As fast as 0.25.x? | ||
| [2010/07/19 19:46:34] <duritong> but the first one stills is about 6 times slower than it was in 0.25.5 | ||
| [2010/07/19 19:47:10] <duritong> so we got rid off the re-import which slowed things down | ||
| [2010/07/19 19:47:19] <duritong> for every further runs | ||
| [2010/07/19 19:51:18] <duritong> I put the infos in the bug report | ||
| [2010/07/19 19:57:13] <markus> duritong: thanks. | ||
| [2010/07/19 19:57:37] <markus> I'll look at it shortly | ||
| [2010/07/19 20:08:50] @ rcrowley joined channel #puppet-dev | ||
| [2010/07/19 20:25:11] <markus> duritong: If you're still up, could you try again with reductivelabs/master? It includes two changes that may provide speed-up | ||
| [2010/07/19 20:34:58] <duritong> markus: ok, also luke's patch? | ||
| [2010/07/19 20:36:00] <jamesturnbull> duritong: that's in master too | ||
| [2010/07/19 20:36:54] <duritong> ok | ||
| [2010/07/19 20:43:22] @ lak joined channel #puppet-dev | ||
| [2010/07/19 20:58:35] <duritong> markus: which changes? | ||
| [2010/07/19 20:58:49] <duritong> markus: because I don't see any speedup I didn't yet include | ||
| [2010/07/19 20:58:55] <jamesturnbull> duritong: okay thanks | ||
| [2010/07/19 20:59:08] <jamesturnbull> duritong: markus will be BRB I think :) | ||
| [2010/07/19 21:00:12] @ Quit: jmccune: Ping timeout: 252 seconds | ||
| [2010/07/19 21:20:23] <jamesturnbull> duritong: thanks for this - we've decided to release with HEAD and do more testing on the slow-down | ||
| [2010/07/19 21:21:46] <duritong> well I can imagine that these issues are hard to track down | ||
| [2010/07/19 21:23:33] @ github joined channel #puppet-dev | ||
| [2010/07/19 21:23:33] <github> puppet: master Markus Roberts * 40e6f02 (1 files in 1 dirs): Tweak to fix for #4233 -- only accept word chars in types ... | ||
| [2010/07/19 21:23:33] <gepetto> github: puppet: #4233 is http://projects.puppetlabs.com/issues/show/4233 "Puppet - Bug #4233: Catalogs fail to deserialize if resources contain a newline in the title - Puppet Labs" | ||
| [2010/07/19 21:23:33] <github> puppet: master Markus Roberts * bbc07f2 (1 files in 1 dirs): Bandaid for #4285 -- :name vs <namevar> ... | ||
| [2010/07/19 21:23:34] <gepetto> github: puppet: #4285 is http://projects.puppetlabs.com/issues/show/4285 "Puppet - Bug #4285: ArgumentError: Cannot alias File[mytitle] to [nil] - Puppet Labs" | ||
| [2010/07/19 21:23:34] <github> puppet: master Markus Roberts * d2da1d4 (1 files in 1 dirs): Tweak to tweak to fix for #4233 -- ":" is valid in type names ... | ||
| [2010/07/19 21:23:34] <gepetto> github: puppet: #4233 is http://projects.puppetlabs.com/issues/show/4233 "Puppet - Bug #4233: Catalogs fail to deserialize if resources contain a newline in the title - Puppet Labs" | ||
| [2010/07/19 21:23:34] <github> puppet: master Matt Robinson * 539d716 (4 files in 4 dirs): [#4287] Fix the undefined evaluate_match error when comparing functions ... | ||
| [2010/07/19 21:23:34] <github> puppet: master Jesse Wolfe * 1288f8c (6 files in 4 dirs): [#4270] Force inherited classes to load into the correct environment ... | ||
| [2010/07/19 21:23:34] <github> puppet: master Nick Lewis * 06fc40c (2 files in 2 dirs): [#4269] Undef variables interpolate to empty string ... | ||
| [2010/07/19 21:23:35] <github> puppet: master Luke Kanies * 42a475e (4 files in 4 dirs): Fixing #4268 - manifests always imported ... | ||
| [2010/07/19 21:23:35] <gepetto> github: puppet: #4268 is http://projects.puppetlabs.com/issues/show/4268 "Puppet - Bug #4268: puppet master 2.6 always imports manifests newly - Puppet Labs" | ||
| [2010/07/19 21:23:35] <github> puppet: master James Turnbull * db0b30d (1 files in 1 dirs): Updated CHANGELOG for 2.6.0 | ||
| [2010/07/19 21:23:35] <github> puppet: master commits d87a2e3...db0b30d - http://bit.ly/ajMjNC | ||
| [2010/07/19 21:23:36] @ github left channel #puppet-dev () | ||
| [2010/07/19 21:37:47] @ jmccune joined channel #puppet-dev | ||
| [2010/07/19 22:02:28] @ github joined channel #puppet-dev | ||
| [2010/07/19 22:02:28] <github> puppet: master James Turnbull * 64386cf (1 files in 1 dirs): Fixed Indirection reference - http://bit.ly/bHjk4l | ||
| [2010/07/19 22:02:28] @ github left channel #puppet-dev () | ||
| [2010/07/19 22:04:39] <markus> jamesturnbull: I can't log into my reductivelabs gmail account at the moment. Any idea why? | ||
| [2010/07/19 22:06:02] <markus> hmm. Now gmail is Ok, but IM is not. | ||
| [2010/07/19 22:25:42] @ github joined channel #puppet-dev | ||
| [2010/07/19 22:25:42] <github> puppet: master James Turnbull * a0a63c3 (2 files in 2 dirs): Fixed network and indirection reference - http://bit.ly/c1bmaO | ||
| [2010/07/19 22:25:42] @ github left channel #puppet-dev () | ||
| [2010/07/19 22:25:55] <jamesturnbull> markus: those commits are fixed post the 2.6.0 tag for puppetdoc | ||
| [2010/07/19 22:26:03] <jamesturnbull> markus: which wouldn't generate reports | ||
| [2010/07/19 22:26:59] <markus> jamesturnbull: what? | ||
| [2010/07/19 22:27:45] <jamesturnbull> markus: -reports +references | ||
| [2010/07/19 22:27:58] <jamesturnbull> markus: the network and indirection REFERENCES | ||
| [2010/07/19 22:28:01] <jamesturnbull> markus: were broken | ||
| [2010/07/19 22:28:06] * jamesturnbull is sorry for making you panic | ||
| [2010/07/19 22:28:14] <markus> Ah, ok. | ||
| [2010/07/19 22:28:59] <markus> jamesturnbull: I'm trying to figure out why I can't get on IM with my reductivelabs account. | ||
| [2010/07/19 22:30:15] @ Quit: markus: Read error: Connection reset by peer | ||
| [2010/07/19 22:30:15] @ mqr joined channel #puppet-dev | ||
| [2010/07/19 22:30:16] @ mqr is now known as markus | ||
| [2010/07/19 22:30:39] <jamesturnbull> markus: not sure? works for me? | ||
| [2010/07/19 22:31:35] <markus> I think I am having network issues. | ||
| [2010/07/19 22:33:02] <jamesturnbull> markus: are your kids near the cables? | ||
| [2010/07/19 22:34:25] <markus> jamesturnbull: To the extent that I'm using wifi, yes. Maybe I should check if Max snuck any 2+GHz equipment up to his bedroom. | ||
| [2010/07/19 22:35:59] <jamesturnbull> markus: I thought they might have figured they'd get dad back if they severed some cables | ||
| [2010/07/19 22:36:02] <jamesturnbull> or fingers | ||
| [2010/07/19 22:36:04] <jamesturnbull> cables is better | ||
| [2010/07/19 22:36:41] <markus> Yeah. The bigger two are in bed & Walt is helping me type. | ||
| [2010/07/19 22:36:59] <markus> But they've really been in wanting dad back mode. | ||
| [2010/07/19 22:37:13] <lak> yeah, i've found my kids don't really make me type better, for some reason :/ | ||
| [2010/07/19 22:38:09] <markus> When I got back from Texas, Walter wouldn't speak to me for a day or so. Just glared at me with his lips quivering and turned away. | ||
| [2010/07/19 22:38:22] <lak> heh | ||
| [2010/07/19 22:38:32] <lak> my kids still think i fly to work every day :/ | ||
| [2010/07/19 22:38:34] <markus> But we're buds again. | ||
| [2010/07/19 22:38:43] <markus> lak: you do. | ||
| [2010/07/19 22:39:00] <lak> magic bus, magic bus? | ||
| [2010/07/19 22:39:13] <markus> Have you worked out your average daily commute for the past few months? | ||
| [2010/07/19 22:39:26] <lak> no thank you | ||
| [2010/07/19 22:39:35] <markus> I suspect you've got me beat by a factor of 5 at least. | ||
| [2010/07/19 22:41:29] <lak> well i've got to do *something* :) | ||
| [2010/07/19 22:41:34] <lak> tty tomorrow | ||
| [2010/07/19 22:41:40] <lak> good job getting the release done | ||
| [2010/07/19 22:42:04] @ Quit: lak: Quit: lak | ||
| [2010/07/19 22:47:44] @ mqr joined channel #puppet-dev | ||
| [2010/07/19 22:47:45] @ Quit: markus: Read error: Connection reset by peer | ||
| [2010/07/19 22:47:46] @ mqr is now known as markus | ||
| [2010/07/19 22:48:25] @ Quit: markus: Client Quit | ||
| [2010/07/19 22:57:45] @ Quit: plathrop-away: Read error: Operation timed out | ||
| [2010/07/19 23:00:44] @ plathrop-away joined channel #puppet-dev | ||
| [2010/07/19 23:01:37] @ markus joined channel #puppet-dev | ||
| [2010/07/19 23:07:55] @ plathrop-away is now known as plathrop | ||
| [2010/07/19 23:13:55] @ nanliu joined channel #puppet-dev | ||
| [2010/07/19 23:15:00] @ Quit: rcrowley: Quit: rcrowley | ||
| [2010/07/19 23:45:29] @ Quit: nanliu: Quit: nanliu | ||
| [2010/07/19 23:56:15] @ Quit: jmccune: Ping timeout: 265 seconds |
Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!