Saturday, 2010-07-17

[2010/07/17 09:08:52] @ Log started by gepetto
[2010/07/17 09:08:52] @ rcrowley joined channel #puppet-dev
[2010/07/17 10:01:42] @ Quit: beata-: Read error: Connection reset by peer
[2010/07/17 10:02:05] @ beata- joined channel #puppet-dev
[2010/07/17 10:45:09] <jamesturnbull> blkperl: you still after luke?
[2010/07/17 11:00:44] * nasrat is after jamesturnbull
[2010/07/17 11:08:40] <jamesturnbull> nasrat: is here
[2010/07/17 11:11:53] <nasrat> jamesturnbull: ok the next few days I am at a conference so can get some late night hacking in
[2010/07/17 11:12:41] <nasrat> do you/puppetlabs have a VT/AMD-V enabled box in the test lab that we can setup the huge no of host/guest systems to test against
[2010/07/17 11:13:00] <jamesturnbull> nasrat: it's on my list but not currently no
[2010/07/17 11:13:41] <nasrat> *sigh* I'll be on my work laptop which I can't really blow away, and ec2 is xen, and vmware is only one tech
[2010/07/17 11:50:24] @ jmccune joined channel #puppet-dev
[2010/07/17 12:19:59] <blkperl> jamesturnbull: nope we got him thanks though
[2010/07/17 13:26:33] @ Quit: jmccune: Ping timeout: 240 seconds
[2010/07/17 14:33:41] <duritong> masterzen: did you already compare compilation time between 0.25 and 2.6?
[2010/07/17 14:43:25] @ Quit: rcrowley: Quit: rcrowley
[2010/07/17 14:57:26] @ joevano joined channel #puppet-dev
[2010/07/17 15:04:22] <duritong> with webrick, is puppet master supposed to import things on every run?
[2010/07/17 17:18:04] @ jmccune joined channel #puppet-dev
[2010/07/17 17:23:04] <jamesturnbull> duritong: that's a bug - markus is lookign now
[2010/07/17 17:32:15] <blkperl> hmm the install.rb uses 'which gzip' thats not going to work on windows
[2010/07/17 17:55:39] <duritong> jamesturnbull: thanks
[2010/07/17 17:59:34] <jamesturnbull> duritong: #4268 neither markus nor I can replicate
[2010/07/17 17:59:35] <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"
[2010/07/17 17:59:38] @ markus joined channel #puppet-dev
[2010/07/17 18:01:03] <markus> duritong?
[2010/07/17 18:02:54] <duritong> markus: here
[2010/07/17 18:03:22] <markus> We're trying to reproduce #4268
[2010/07/17 18:03:22] <gepetto> markus: #4268 is http://projects.puppetlabs.com/issues/show/4268 "Puppet - Bug #4268: puppet master 2.6 always imports manifests newly - Puppet Labs"
[2010/07/17 18:03:26] <duritong> ic
[2010/07/17 18:03:31] <markus> No joy so far.
[2010/07/17 18:03:38] <duritong> ok
[2010/07/17 18:04:28] <duritong> so I run my 2.6rc3 (with some patches) with:
[2010/07/17 18:04:30] <duritong> puppet master --config /srv/puppet/etc/puppet.conf --no-daemonize --masterport 8141 --verbose
[2010/07/17 18:05:16] <duritong> and I connect my 0.25.5 client with
[2010/07/17 18:05:30] <duritong> puppetd --test --masterport 8141 --noop --configtimeout 600
[2010/07/17 18:06:04] <duritong> and I get all these info: Automatically imported foo at the first run
[2010/07/17 18:06:32] <duritong> and then when I rerun the client I get these infos as well
[2010/07/17 18:07:20] <markus> Questions: 1) what patches? 2) Does it do it when not in noop (if that's testable)? 3) can you produce a very simple manfest that demonstrates the problem?
[2010/07/17 18:07:42] <duritong> patches:
[2010/07/17 18:07:54] <duritong> starting from: f60b6a0e54bf99e20bfa876b8341a5cc0da74584
[2010/07/17 18:08:00] <duritong> 1cda7c53e340bd95513c973e356cbc90372efe38
[2010/07/17 18:08:03] <duritong> 91185c6c04dc2bb5659db8f7a7d23a41c4d308ea
[2010/07/17 18:08:08] <duritong> 8701cd2fedcb6b149a396fea4820cc669c5aa3c8
[2010/07/17 18:08:19] <duritong> 1589833153d56195d11e32b4146f0f5ce9ddbdd0
[2010/07/17 18:08:26] <duritong> 788c91f3071f24e9e64bed89e7684aa38f3eb9ac
[2010/07/17 18:08:30] <duritong> a0af61d3879840d163accd9538a5c76179f644fe
[2010/07/17 18:09:43] <duritong> due to the undef != '' bug/behavior change I can't test it without noop
[2010/07/17 18:10:33] <duritong> but I can try to setup a more simple manifest
[2010/07/17 18:15:13] @ Quit: jmccune: Ping timeout: 260 seconds
[2010/07/17 18:15:14] <markus> duritong: *sigh* Well, you've got all the ones I was thinking might help in there.
[2010/07/17 18:16:09] * duritong is preparing a simple setup
[2010/07/17 18:16:37] <markus> Let me know if you come up with a simpler demo-rig. I'm going to try with noop, etc. to see if I can find a clue.
[2010/07/17 18:23:32] <duritong> yep
[2010/07/17 18:43:33] <duritong> markus: ok I have a simple setup that I could send you
[2010/07/17 18:43:44] <markus> Cool!
[2010/07/17 18:43:52] <duritong> and it does the autoimport each time
[2010/07/17 18:43:53] <markus> Email or...?
[2010/07/17 18:44:01] <duritong> letme see how big it is
[2010/07/17 18:44:04] <markus> Ok/
[2010/07/17 18:49:11] <markus> duritong: Also, going back over the patch list, I find I don't have 8701cd2fedcb6b149a396fea4820cc669c5aa3c8 anywhere. Could you give me more info on what that is?
[2010/07/17 18:49:53] <duritong> markus: mom
[2010/07/17 18:50:02] <duritong> markus: you got the tarball in your query
[2010/07/17 18:50:56] <duritong> markus: 8701cd2fedcb6b149a396fea4820cc669c5aa3c8 is a patch of brice regarding '' in comparators
[2010/07/17 18:51:07] <duritong> Fix #4238 - if should match undef as ''
[2010/07/17 18:51:07] <gepetto> duritong: #4238 is http://projects.puppetlabs.com/issues/show/4238 "Puppet - Bug #4238: an undef variable is not anymore an empty string - Puppet Labs"
[2010/07/17 18:51:32] <markus> Ah, ok. Wonder why git isn't finding it...?
[2010/07/17 18:52:23] <duritong> markus: yeah strange
[2010/07/17 18:52:24] <markus> duritong: I did not understand the proceeding messages ("mom"/"you got the tarball in your query")
[2010/07/17 18:52:58] <duritong> markus: ah brice updated that patch
[2010/07/17 18:53:10] <markus> Yeah.
[2010/07/17 18:53:17] <duritong> so I'm on the later version
[2010/07/17 18:53:32] <duritong> regarding query: I pasted a link to a tarball in a query to you
[2010/07/17 18:54:27] <duritong> so you should have some other tab/window opened where you'll find the link, but I can also send it to you by mail
[2010/07/17 18:59:46] <markus> duritong: Yes, I see it now. Thanks!
[2010/07/17 18:59:59] <duritong> np
[2010/07/17 19:00:04] * duritong needs to find power
[2010/07/17 19:05:03] <duritong> re
[2010/07/17 19:05:52] <duritong> markus: do you think that brice improved patch might fix something
[2010/07/17 19:06:10] <markus> I don't think so (they seem unrelated).
[2010/07/17 19:06:19] <markus> But anything is possible.
[2010/07/17 19:06:38] <markus> I downloaded your test case and am looking through it.
[2010/07/17 19:06:52] <duritong> ok I'll include that one as well
[2010/07/17 19:07:11] <markus> Will be having to break for dinner & getting my kids to bed here soon.
[2010/07/17 19:07:26] <duritong> np
[2010/07/17 19:07:48] <markus> Should be back for an hour or two in about 2 hours, and then more in about 14 hours from now.
[2010/07/17 19:08:17] <markus> I'll update the ticket as I learn things (or document my failure to learn things).
[2010/07/17 19:08:55] <duritong> I'm currently in NYC so a bit a better timezone to interact, but I'm also at a conference, so I might be distracted from time to time
[2010/07/17 19:09:29] <duritong> but at least tomorrow during the day I'll should be around
[2010/07/17 19:17:01] <markus> duritong: Ok. Using your tarball w. storeconfigs turned off (I don't have mysql set up and wantedto test before dinner) I do _not_ see the "info: Automatically imported _ from _" messages after the first run. So there's something goofy going on. Will investigate more later.
[2010/07/17 19:17:48] <duritong> markus: ok I can turn storeconfigs of
[2010/07/17 19:18:01] <duritong> and I'll rebuild my version with brice's later patch
[2010/07/17 19:18:10] <markus> Yeah, that would be a decent thing to test.
[2010/07/17 19:20:02] <duritong> yep without stored configs enabled it does not re-import manifets
[2010/07/17 19:20:09] <duritong> manifests
[2010/07/17 19:20:10] <jamesturnbull> hmmm
[2010/07/17 19:21:17] <duritong> eh mom
[2010/07/17 19:21:23] <duritong> I might be wrong
[2010/07/17 19:21:28] <duritong> autoflush...
[2010/07/17 19:24:41] <duritong> no it still does
[2010/07/17 19:24:54] <duritong> sorry I was wrong due to not use autoflush = true
[2010/07/17 19:25:22] <duritong> ok I'll rebuild my version with the proper patch of brice
[2010/07/17 19:40:20] <duritong> ok rebuild didn't help
[2010/07/17 19:40:33] <duritong> markus: it's a 0.25.5 client to 2.6 server
[2010/07/17 19:43:44] @ Quit: beata-: Ping timeout: 258 seconds
[2010/07/17 19:45:23] <duritong> markus: if you can publish somewhere the branch from which you are testing I can take that one and build packages from that one and test it
[2010/07/17 19:45:34] <duritong> so we are sure we use the same versions
[2010/07/17 19:45:41] <duritong> but that's fine tomorrow
[2010/07/17 19:49:20] @ beata- joined channel #puppet-dev
[2010/07/17 19:53:11] <markus> duritong: I was most recently using the (ephemeral) reductivelabs "testing" branch which is presently an initial sketch what we will call 2.6.0rc4; it contains 2.6.0rc3 plus the ready for testing / ready for checkin branches from tickets targeted at 2.6.0
[2010/07/17 19:58:24] <duritong> markus: ok
[2010/07/17 19:58:54] <duritong> I give it a test with that one
[2010/07/17 20:08:18] <duritong> markus: I have the same behavior (always importing) with the testing branch
[2010/07/17 20:10:12] @ rcrowley joined channel #puppet-dev
[2010/07/17 20:10:21] <duritong> ruby version: ruby 1.8.6 (2010-02-05 patchlevel 399) [x86_64-linux]
[2010/07/17 20:14:56] @ Quit: rcrowley: Client Quit
[2010/07/17 20:35:36] @ jmccune joined channel #puppet-dev
[2010/07/17 22:30:05] @ Quit: jmccune: Ping timeout: 240 seconds
[2010/07/17 23:16:34] <markus> duritong: No luck on this end. I even tried setting the source file dates in the future, filetimeout to 0, etc. but I can't get anything that looks like what you see.

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