Tuesday, 2010-01-26

[2010/01/26 00:02:53] @ Log started by gepetto
[2010/01/26 00:02:53] @ Quit: nigelk:
[2010/01/26 00:50:58] @ lak joined channel #puppet-dev
[2010/01/26 01:49:54] @ nigelk joined channel #puppet-dev
[2010/01/26 02:27:20] @ ShiNboi joined channel #puppet-dev
[2010/01/26 02:52:23] @ jes5 joined channel #puppet-dev
[2010/01/26 02:58:50] @ Quit: nigelk:
[2010/01/26 03:46:50] @ Quit: lak:
[2010/01/26 05:50:28] @ Quit: jes5: "Leaving."
[2010/01/26 08:59:24] @ Quit: ppickfor: Read error: 110 (Connection timed out)
[2010/01/26 10:12:05] @ Quit: ShiNboi: Client Quit
[2010/01/26 11:40:15] @ lak joined channel #puppet-dev
[2010/01/26 11:52:12] @ Quit: lak:
[2010/01/26 12:07:15] @ lak joined channel #puppet-dev
[2010/01/26 12:24:46] @ Quit: lak:
[2010/01/26 12:40:37] <nasrat> hmm
[2010/01/26 12:42:15] <nasrat> ok anyone from reductive about?
[2010/01/26 12:50:26] <nasrat> so function autoloading has a funky except clause:
[2010/01/26 12:50:29] <nasrat> http://gist.github.com/287034
[2010/01/26 12:50:38] <nasrat> unless detail.to_s =~ /^no such file/i
[2010/01/26 12:56:15] @ jes5 joined channel #puppet-dev
[2010/01/26 13:02:59] <Volcane> hating how different versions of ruby raise different exceptions too :(
[2010/01/26 13:03:08] <Volcane> some will raise for ^C others wont etc, right pita
[2010/01/26 14:55:23] @ Quit: jes5: "Leaving."
[2010/01/26 14:59:08] @ hacim_ is now known as hacim
[2010/01/26 15:21:51] @ rocket is now known as r0ck3t
[2010/01/26 15:22:20] @ r0ck3t is now known as rocket
[2010/01/26 15:55:41] @ jes5 joined channel #puppet-dev
[2010/01/26 17:04:54] <duritong_> anybody ever thought about running the independent graphs of one catalog in multiple threads?
[2010/01/26 17:05:34] <duritong_> for sure certain resource types such as packages would need a global locking, as most package manager don't like it to be started more than one
[2010/01/26 17:06:26] <duritong_> and you would need something like a max thread count to not knock out the machine
[2010/01/26 17:06:29] <fsweetser> or, even nice, aggregated together to just call the package manager once with all packages
[2010/01/26 17:08:48] <duritong_> well I'm talking about a whole set of independent graphs
[2010/01/26 17:09:11] <duritong_> what have already been worked on is to gather all the packages with no dependencies and install them in one shot
[2010/01/26 17:09:43] <fsweetser> I would guess that the two would be at least somewhat related, since you'd have to take all instances of a given type and make sure that the type itself is safe to run in parallel
[2010/01/26 17:10:24] <fsweetser> packages, augeas, etc might still have to be run non-concurrently
[2010/01/26 17:10:54] <fsweetser> at that point, lumping depdency free packages into a single run should be trivial, I would imagine
[2010/01/26 17:11:19] <duritong_> afair there have been a working prototype for that
[2010/01/26 17:11:34] <duritong_> but it never got out of that state
[2010/01/26 17:11:35] <fsweetser> cool
[2010/01/26 17:11:48] <fsweetser> not very high priority, huh?
[2010/01/26 17:12:53] <duritong_> unfortunately
[2010/01/26 17:13:17] <fsweetser> ah, well
[2010/01/26 17:13:39] <fsweetser> performance is nice, but I can see it taking a back seat to bugfixes and new features
[2010/01/26 17:14:15] <duritong_> yeah
[2010/01/26 17:14:43] <duritong_> and if I look at my manifests I rarely have completely independent package resources
[2010/01/26 17:49:12] @ lak joined channel #puppet-dev
[2010/01/26 17:52:28] @ Quit: ohadlevy: calvino.freenode.net irc.freenode.net
[2010/01/26 17:52:31] @ ohadlevy joined channel #puppet-dev
[2010/01/26 18:54:52] @ Quit: lak:
[2010/01/26 20:04:58] @ github joined channel #puppet-dev
[2010/01/26 20:04:58] <github> puppet: Markus Roberts 0.25.x SHA1-0025e13
[2010/01/26 20:04:58] <github> Partial reversion of patch for #3088 to fix #3104 (Exception misreported)
[2010/01/26 20:04:59] <gepetto> github: #3088 is http://projects.reductivelabs.com/issues/show/3088 "Puppet - Bug #3088: Puppetd fails to stop after receiving SIGTERM - ReductiveLabs.com"
[2010/01/26 20:04:59] <gepetto> github: #3104 is http://projects.reductivelabs.com/issues/show/3104 "Puppet - Bug #3104: Test failed: Puppet::Network::XMLRPCClient when performing the rpc call and an exception is raised.should log and raise XMLRPCClientError if Timeout::Error is raised - ReductiveLabs.com"
[2010/01/26 20:04:59] <github> http://bit.ly/a2KBff
[2010/01/26 20:05:01] @ github left channel #puppet-dev ()
[2010/01/26 20:58:08] <jamesturnbull> jes5: sorry clean forgot about the integration tests when I sent first email. :)
[2010/01/26 21:06:26] @ Quit: jes5: "Leaving."
[2010/01/26 21:19:13] @ jes5 joined channel #puppet-dev
[2010/01/26 21:38:52] @ ppickfor joined channel #puppet-dev
[2010/01/26 22:05:34] @ Quit: jes5: "Leaving."
[2010/01/26 22:05:43] @ jes5 joined channel #puppet-dev
[2010/01/26 22:12:41] @ jes51 joined channel #puppet-dev
[2010/01/26 22:20:48] @ Quit: jes51: "Leaving."
[2010/01/26 22:30:10] @ Quit: jes5: Read error: 110 (Connection timed out)
[2010/01/26 23:54:47] @ Quit: jamesturnbull: "leaving"

Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!