Thursday, 2009-04-09

[2009/04/09 00:03:47] @ Log started by gepetto
[2009/04/09 00:03:47] @ PhabX_ joined channel #puppet
[2009/04/09 00:09:28] @ WALoeIII joined channel #puppet
[2009/04/09 00:12:57] @ Quit: fujin:
[2009/04/09 00:13:07] @ machpo joined channel #puppet
[2009/04/09 00:19:27] @ Quit: PhabX: Read error: 110 (Connection timed out)
[2009/04/09 00:21:32] @ thatch_ joined channel #puppet
[2009/04/09 00:21:47] @ Quit: ethan_rowe: "Lack of interest wins out."
[2009/04/09 00:23:20] @ edwardam|AFK is now known as edwardam
[2009/04/09 00:25:31] @ Quit: justdave: Remote closed the connection
[2009/04/09 00:28:33] @ justdave joined channel #puppet
[2009/04/09 00:30:23] @ PaulWay joined channel #puppet
[2009/04/09 00:30:23] @ Quit: d3vilb0x_:
[2009/04/09 00:47:06] @ Quit: lak:
[2009/04/09 00:47:33] @ mfoster left channel #puppet ()
[2009/04/09 00:52:43] @ Quit: thatch_: Read error: 110 (Connection timed out)
[2009/04/09 01:02:42] @ WALoeIII_ joined channel #puppet
[2009/04/09 01:03:56] @ Quit: cwebber:
[2009/04/09 01:04:34] @ WALoeIII__ joined channel #puppet
[2009/04/09 01:04:45] @ Quit: WALoeIII_: Read error: 104 (Connection reset by peer)
[2009/04/09 01:05:29] @ Quit: WALoeIII__: Client Quit
[2009/04/09 01:13:38] <PaulWay> I'm writing a nagios check on for puppet, and I've been told (here and there) to check /var/lib/puppet/localconfig.yaml and /var/lib/puppet/state/state.yaml.
[2009/04/09 01:14:12] <PaulWay> AFAIK the first is the local machine's configuration state, and the latter is the state as downloaded from the server.
[2009/04/09 01:14:15] <PaulWay> Can anyone confirm?
[2009/04/09 01:18:29] @ Quit: WALoeIII: Read error: 110 (Connection timed out)
[2009/04/09 01:23:36] @ Quit: ezekiel-v3: Read error: 110 (Connection timed out)
[2009/04/09 01:31:36] @ fujin joined channel #puppet
[2009/04/09 01:36:24] @ Quit: fujin: Client Quit
[2009/04/09 01:42:45] @ Quit: jtimberman_: "Quitters never win [at irc]."
[2009/04/09 01:44:39] <PaulWay> And can I just double-check here - I can add a tag to the list in a node definition with '$tag += "awesome"'
[2009/04/09 01:53:45] @ Quit: PhabX_: "Leaving..."
[2009/04/09 02:00:03] @ philipp_ joined channel #puppet
[2009/04/09 02:02:28] <PaulWay> Aaahhhhh, my love for Puppet knows no bounds right this minute.
[2009/04/09 02:03:07] <PaulWay> I just taught our 2IC linux sysadmin how to write puppet rules and check them in with Subversion.
[2009/04/09 02:03:30] @ Quit: Djelibeybi: "Leaving"
[2009/04/09 02:05:05] @ nasrat joined channel #puppet
[2009/04/09 02:06:03] @ ohadlevy joined channel #puppet
[2009/04/09 02:12:09] <ohadlevy> anyone here knows how to convert a STDIN/OUT to socket in ruby?
[2009/04/09 02:21:13] @ pleemans joined channel #puppet
[2009/04/09 02:23:34] @ Quit: nasrat:
[2009/04/09 02:26:32] @ nasrat joined channel #puppet
[2009/04/09 02:30:43] @ PsychoSid joined channel #puppet
[2009/04/09 02:43:04] @ edwardam is now known as edwardam|zZzZz
[2009/04/09 02:43:06] @ Quit: alban2: Read error: 113 (No route to host)
[2009/04/09 02:52:04] @ Quit: claymation: Read error: 110 (Connection timed out)
[2009/04/09 02:56:34] @ PaulWay left channel #puppet ()
[2009/04/09 03:00:35] @ PaulWay joined channel #puppet
[2009/04/09 03:01:02] <PaulWay> What's the correct way of having a list of tags?
[2009/04/09 03:01:38] <PaulWay> I want to do $tag = ["tag1", "tag2"] but it's saying "Cannot append, variable tag is defined in this scope"
[2009/04/09 03:01:55] <rabbit7> seems to be difficult to report a bug in the puppet project !
[2009/04/09 03:02:23] @ Quit: pleemans: Read error: 110 (Connection timed out)
[2009/04/09 03:05:49] @ Quit: mconigliaro1: Read error: 104 (Connection reset by peer)
[2009/04/09 03:09:27] @ Innocenti joined channel #puppet
[2009/04/09 03:11:18] @ MattyM joined channel #puppet
[2009/04/09 03:11:24] @ PaulWay left channel #puppet ()
[2009/04/09 03:12:26] @ DerekW joined channel #puppet
[2009/04/09 03:19:08] @ stijnbe joined channel #puppet
[2009/04/09 03:20:48] @ Quit: nasrat:
[2009/04/09 03:22:26] <rabbit7> anybody knows how to do such a construct in a template <%= variable || defaultvalue %>
[2009/04/09 03:22:31] <rabbit7> this would be veeery neat
[2009/04/09 03:23:41] <ohadlevy> prio to 0.24-5 i think you need to do it with a begin and rescue blocks
[2009/04/09 03:23:58] <ohadlevy> in newer version there is a method which checks if a variable is empty (if i remeber correctly)
[2009/04/09 03:24:26] <rabbit7> ohh that would be great.. do you know it by any chance ?
[2009/04/09 03:35:20] @ tuf joined channel #puppet
[2009/04/09 03:35:53] @ nasrat joined channel #puppet
[2009/04/09 03:36:08] @ thegcat joined channel #puppet
[2009/04/09 03:36:24] <nasrat> rabbit7: http://projects.reductivelabs.com/projects/puppet/issues/
[2009/04/09 03:36:46] <rabbit7> nasrat: thanks was finally able to do it
[2009/04/09 03:37:23] @ Quit: nasrat: Client Quit
[2009/04/09 03:42:30] @ tripoux joined channel #puppet
[2009/04/09 03:45:10] @ bajan joined channel #puppet
[2009/04/09 03:45:18] @ andrewcshafer joined channel #puppet
[2009/04/09 03:46:52] @ LostyJai joined channel #puppet
[2009/04/09 03:46:54] @ MarlondB joined channel #puppet
[2009/04/09 03:48:32] @ Quit: ssm: Read error: 101 (Network is unreachable)
[2009/04/09 03:58:08] @ keithlard joined channel #puppet
[2009/04/09 04:02:18] @ jlamsens joined channel #puppet
[2009/04/09 04:03:42] <jlamsens> anyone knows how to let a server execute MySQL statements (creating an initial DB with a table), other than copying a statements.sql to the server and doing an exec of "mysql -u root -ppasword < statements.sql" ?
[2009/04/09 04:04:13] @ LostyJai left channel #puppet ()
[2009/04/09 04:12:13] <JD> jlamsens: write a mysql type?
[2009/04/09 04:13:01] <jlamsens> JD: could you give me an example or maybe a link to more info ?
[2009/04/09 04:18:07] <JD> jlamsens: it involves learning ruby
[2009/04/09 04:18:37] @ bajan left channel #puppet ()
[2009/04/09 04:18:52] @ Quit: ghenry: Read error: 110 (Connection timed out)
[2009/04/09 04:19:30] @ ghenry joined channel #puppet
[2009/04/09 04:19:54] <keithlard> jlamsens: what problem are you trying to solve?
[2009/04/09 04:23:46] <jlamsens> keithlard: I'm trying to create a mysql database
[2009/04/09 04:23:56] <jlamsens> and I want to automagically do it via Puppet
[2009/04/09 04:24:10] <jlamsens> very easy: create a database and execute some MySQL statements
[2009/04/09 04:25:47] <jlamsens> I think I know a possibility: file {"/tmp/mysqlstatements.sql": ensure => present} and then do an exec of mysql -u root -ppass < /tmp/mysqlstatements.sql
[2009/04/09 04:26:03] <jlamsens> but I wanted to know if there is a 'more elegant' solution
[2009/04/09 04:27:19] <keithlard> not that i know of - that's the way i do it :)
[2009/04/09 04:29:49] <jlamsens> ok thanks
[2009/04/09 04:30:36] @ Quit: ohadlevy: Read error: 110 (Connection timed out)
[2009/04/09 04:37:19] @ Quit: mikepea:
[2009/04/09 04:40:33] @ mikepea joined channel #puppet
[2009/04/09 04:42:44] <z00dax> jlamsens: depending on the db type, you could just ship the db files under mysql :)
[2009/04/09 04:43:37] @ verwilst joined channel #puppet
[2009/04/09 04:49:11] @ Quit: thegcat:
[2009/04/09 04:50:00] @ thegcat joined channel #puppet
[2009/04/09 04:58:25] @ alban2 joined channel #puppet
[2009/04/09 04:59:11] @ Quit: bolt: "telnet irc.freenode.net 6667"
[2009/04/09 05:09:57] @ pleemans joined channel #puppet
[2009/04/09 05:16:56] @ glaw joined channel #puppet
[2009/04/09 05:17:17] @ dene left channel #puppet ()
[2009/04/09 05:22:14] @ markl____ joined channel #puppet
[2009/04/09 05:22:15] @ Quit: markl___: Read error: 104 (Connection reset by peer)
[2009/04/09 05:26:23] @ Quit: f3ew: Read error: 104 (Connection reset by peer)
[2009/04/09 05:27:07] @ Quit: keithlard:
[2009/04/09 05:27:24] @ f3ew joined channel #puppet
[2009/04/09 05:28:45] @ aymerick joined channel #puppet
[2009/04/09 05:34:07] @ Quit: markl____: Read error: 60 (Operation timed out)
[2009/04/09 05:41:10] <glaw> /eave
[2009/04/09 05:41:14] @ glaw left channel #puppet ()
[2009/04/09 05:45:20] @ bolt joined channel #puppet
[2009/04/09 05:46:13] @ Quit: PsychoSid: "Leaving."
[2009/04/09 05:47:57] @ Quit: markl__: Read error: 110 (Connection timed out)
[2009/04/09 05:49:06] @ Quit: markl_: Read error: 110 (Connection timed out)
[2009/04/09 05:50:50] @ Quit: bolt: "telnet irc.freenode.net 6667"
[2009/04/09 06:05:10] @ Quit: pleemans: Read error: 110 (Connection timed out)
[2009/04/09 06:10:57] @ keithlard joined channel #puppet
[2009/04/09 06:25:28] @ d3vilb0x joined channel #puppet
[2009/04/09 06:36:43] @ Quit: drmikecrowe: Read error: 104 (Connection reset by peer)
[2009/04/09 06:37:28] @ drmikecrowe joined channel #puppet
[2009/04/09 06:38:39] * Volcane thinks 0.24.8 has some weird issue making home dirs on centos
[2009/04/09 06:39:45] <Volcane> or some defaults changed
[2009/04/09 06:41:06] @ Quit: mikepea:
[2009/04/09 06:55:31] @ Quit: d3vilb0x:
[2009/04/09 07:02:05] @ d3vilb0x_ joined channel #puppet
[2009/04/09 07:03:55] @ Quit: Bass10: Read error: 110 (Connection timed out)
[2009/04/09 07:04:15] @ Quit: sc0ttB: Read error: 60 (Operation timed out)
[2009/04/09 07:07:42] @ mikepea joined channel #puppet
[2009/04/09 07:08:37] @ elementai joined channel #puppet
[2009/04/09 07:18:41] @ Quit: falsch: Read error: 110 (Connection timed out)
[2009/04/09 07:19:00] @ Quit: ijcd:
[2009/04/09 07:32:22] @ Quit: d3vilb0x_:
[2009/04/09 07:36:20] @ d3vilb0x joined channel #puppet
[2009/04/09 07:54:09] @ nakano_ is now known as nakano
[2009/04/09 08:08:04] <phantez> hey, is puppet fully compatible with ruby 1.9 ?
[2009/04/09 08:11:25] @ Quit: barn: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:26] @ Quit: tim|imac: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:26] @ Quit: arjanoosting: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:26] @ Quit: madduck: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:27] @ Quit: _nono_: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:27] @ Quit: alban2: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:28] @ Quit: ribo: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:28] @ Quit: FiXion: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:29] @ Quit: johnf1911: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:29] @ Quit: LapTop006: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:29] @ Quit: idimmu: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:30] @ Quit: riddley: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:30] @ Quit: fsweetser: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:30] @ Quit: duritong: kubrick.freenode.net irc.freenode.net
[2009/04/09 08:11:52] @ alban2 joined channel #puppet
[2009/04/09 08:11:52] @ barn joined channel #puppet
[2009/04/09 08:11:52] @ ribo joined channel #puppet
[2009/04/09 08:11:52] @ FiXion joined channel #puppet
[2009/04/09 08:11:53] @ duritong joined channel #puppet
[2009/04/09 08:11:53] @ _nono_ joined channel #puppet
[2009/04/09 08:11:53] @ johnf1911 joined channel #puppet
[2009/04/09 08:11:53] @ LapTop006 joined channel #puppet
[2009/04/09 08:11:53] @ riddley joined channel #puppet
[2009/04/09 08:11:53] @ fsweetser joined channel #puppet
[2009/04/09 08:11:53] @ idimmu joined channel #puppet
[2009/04/09 08:11:53] @ tim|imac joined channel #puppet
[2009/04/09 08:11:53] @ madduck joined channel #puppet
[2009/04/09 08:11:53] @ arjanoosting joined channel #puppet
[2009/04/09 08:23:06] @ bolt joined channel #puppet
[2009/04/09 08:24:46] @ Quit: bolt: Client Quit
[2009/04/09 08:27:11] @ bajan joined channel #puppet
[2009/04/09 08:38:19] @ flashn joined channel #puppet
[2009/04/09 08:40:09] @ Quit: slap_stick: "leaving"
[2009/04/09 08:41:23] <flashn> puppet seems more straight-ahead and learning curve is alot shorter than cfengine.. trying to decide what would be easiest to spread administrational knowledge etc for roughly 80 people
[2009/04/09 08:42:47] @ drmikecrowe_ joined channel #puppet
[2009/04/09 08:45:36] @ hexasof1 joined channel #puppet
[2009/04/09 08:45:40] <hexasof1> hello
[2009/04/09 08:47:12] <hexasof1> a question about facter: I created a very simple facter which is a wrapper for an internal command
[2009/04/09 08:47:54] <hexasof1> it works fine, but I what to set a specific PATH before running it. someone knows how to do that or a link to a related doc?
[2009/04/09 08:48:16] <hexasof1> -> %x(/afs/in2p3.fr/common/uss/machine_status).chomp
[2009/04/09 08:49:18] <Volcane> ENV['PATH'] is a normal ruby array that you can manipulate like any array
[2009/04/09 08:49:55] <hexasof1> ok, thanks, I will check (I know very little about ruby, in fact)
[2009/04/09 08:50:23] <Volcane> if you just want to add to it then something like: ENV['PATH'] << "/afs/in2p3.fr/common/uss"
[2009/04/09 08:50:25] <Volcane> should do it
[2009/04/09 08:50:35] <Volcane> oh actually i am being stupid
[2009/04/09 08:50:39] <Volcane> its not an array obviously
[2009/04/09 08:50:57] <hexasof1> I will try
[2009/04/09 08:51:16] <Volcane> ENV['PATH'] += "/afs/in2p3.fr/common/uss"
[2009/04/09 08:52:04] @ ssm joined channel #puppet
[2009/04/09 08:52:07] <hexasof1> (as soon as I will be out of this meeting)
[2009/04/09 08:52:25] <Volcane> arg, and put a : infront of that
[2009/04/09 08:52:30] <Volcane> sheesh, you'd swear its friday or something
[2009/04/09 08:54:27] <huggie> :)
[2009/04/09 08:56:38] @ Quit: drmikecrowe: Nick collision from services.
[2009/04/09 08:56:38] @ drmikecrowe_ is now known as drmikecrowe
[2009/04/09 09:00:39] @ joe-mac joined channel #puppet
[2009/04/09 09:04:15] @ bajan left channel #puppet ()
[2009/04/09 09:10:29] @ Quit: philipp_: Remote closed the connection
[2009/04/09 09:10:49] @ Quit: thegcat:
[2009/04/09 09:11:27] @ bolt joined channel #puppet
[2009/04/09 09:13:07] @ erm_ joined channel #puppet
[2009/04/09 09:16:35] @ Quit: egerlach: "leaving"
[2009/04/09 09:19:43] @ Quit: sfeehan: Read error: 60 (Operation timed out)
[2009/04/09 09:21:07] @ sfeehan joined channel #puppet
[2009/04/09 09:21:38] @ PsychoSid joined channel #puppet
[2009/04/09 09:22:50] @ claymation joined channel #puppet
[2009/04/09 09:23:37] @ Quit: jmeeuwen: Remote closed the connection
[2009/04/09 09:23:56] @ Quit: maxagaz: Read error: 110 (Connection timed out)
[2009/04/09 09:24:02] @ Quit: tobias-: Read error: 60 (Operation timed out)
[2009/04/09 09:24:10] @ jmeeuwen joined channel #puppet
[2009/04/09 09:27:52] @ tobias- joined channel #puppet
[2009/04/09 09:28:09] @ schubi2 joined channel #puppet
[2009/04/09 09:30:17] @ avleen_ joined channel #puppet
[2009/04/09 09:30:21] @ Quit: jbk: Read error: 104 (Connection reset by peer)
[2009/04/09 09:30:24] @ jbk joined channel #puppet
[2009/04/09 09:30:29] @ Nigel_ joined channel #puppet
[2009/04/09 09:31:00] @ Quit: avleen: Read error: 104 (Connection reset by peer)
[2009/04/09 09:31:34] @ Quit: schubi2: Client Quit
[2009/04/09 09:37:28] @ Bass10 joined channel #puppet
[2009/04/09 09:42:09] @ Quit: G: Read error: 110 (Connection timed out)
[2009/04/09 09:44:28] @ mfoster joined channel #puppet
[2009/04/09 09:45:15] @ maxagaz joined channel #puppet
[2009/04/09 09:46:24] @ philipp_ joined channel #puppet
[2009/04/09 09:53:07] @ jtimberman joined channel #puppet
[2009/04/09 09:56:35] @ lak joined channel #puppet
[2009/04/09 10:00:07] <elementai> hi, folks, if anyone tried to implement UsingMongrelNginx, didn't your puppetmaster+Mongrel denied clients with "notice: Denying unauthenticated client localhost(127.0.0.1) access to puppetmaster.getconfig", although Webrick works just fine ?
[2009/04/09 10:00:58] <Volcane> double check your ssl certs and that all those proxy_set_header stuff is correct
[2009/04/09 10:01:04] @ cwebber joined channel #puppet
[2009/04/09 10:01:45] @ blubb joined channel #puppet
[2009/04/09 10:03:06] <blubb> Does anybody know of some useful puppet documentation? Especially some "step by step" guide would be highly appreciated.
[2009/04/09 10:03:43] <Volcane> blubb: try http://www.devco.net/pubwiki/Puppet/GettingStarted should get you to the point where the docs on the wiki is useful
[2009/04/09 10:05:24] <huggie> Can anyone see what's wrong with http://pastebin.com/m1207c363 ? Does it not like uppercase in class names?
[2009/04/09 10:05:36] <huggie> I found: http://reductivelabs.com/trac/puppet/wiki/FrequentlyAskedQuestions#what-characters-are-permitted-in-a-class-name
[2009/04/09 10:05:49] <blubb> Thanks, but this site just links to the "documentation" on reductive labs site. I have already read that, but it always stops when things start to get interesting.
[2009/04/09 10:05:52] <huggie> The other feature classes (all lowercase letters) all work.
[2009/04/09 10:05:59] <Volcane> huggie: yeah afaik no camelcasing
[2009/04/09 10:06:25] <huggie> heh. utf8 it is then, ta.
[2009/04/09 10:06:25] <Volcane> blubb: it doesn't, its a thorough step by step from zero to copying files out introduction to puppet
[2009/04/09 10:06:30] <elementai> blubb: you can also try to "Pulling Strings With Puppet" book
[2009/04/09 10:06:42] <elementai> blubb: * to read
[2009/04/09 10:07:06] <Volcane> blubb: but its also specifically there to show you what always-up-to-date docs on the reductive labs are useful and to get yo to the point where you can use those
[2009/04/09 10:07:32] @ Quit: blubb: "http://www.mibbit.com ajax IRC Client"
[2009/04/09 10:08:30] * Volcane suspect the problem is the "read" part in documentation for some people :P
[2009/04/09 10:09:03] @ unxfrek joined channel #puppet
[2009/04/09 10:13:44] @ Quit: bolt: Remote closed the connection
[2009/04/09 10:15:44] @ blubb joined channel #puppet
[2009/04/09 10:15:47] @ egerlach joined channel #puppet
[2009/04/09 10:16:31] @ Quit: lak:
[2009/04/09 10:17:53] @ fujin joined channel #puppet
[2009/04/09 10:18:12] <joe-mac> the most useful puppet documentation IMO is in the recipes section
[2009/04/09 10:18:19] <joe-mac> shows you how people are actaully doing things
[2009/04/09 10:18:27] <Volcane> yeah, when you have the basics down they're great
[2009/04/09 10:20:13] <blubb> Exactly the basics are the problem. I do not get the big picture. Had a look at the recipes, but they do not even state whether some file needs to live on the client or on the server...
[2009/04/09 10:20:35] <Volcane> blubb: did you read the link I sent you? it really is quite thorough in getting you past the basic problems
[2009/04/09 10:21:05] <blubb> Another example from http://reductivelabs.com/trac/puppet/wiki/BigPicture: "When configuration needs to be "pushed" to the clients, the Server can be asked to attempt to trigger each client to request "pull" a new host configuration." Sounds cool, but everybody I showed this, immediately asked "ok, and HOW do you do that?"
[2009/04/09 10:21:44] <philipp_> QUESTION: Puppet doesn't like site.pp~ . Am I using kate for editing and as a default he does some backups during editing. I removed this option. But is there a way that puppetmaster ignores all *.pp~ files ?
[2009/04/09 10:22:30] <joe-mac> philip__: what are you importing?
[2009/04/09 10:22:53] <joe-mac> well, nm i am stupid, i think site.pp might be a hardcoded thing
[2009/04/09 10:23:08] <joe-mac> i know if it's missing or puppet can't read it, it won't start
[2009/04/09 10:23:37] <philipp_> site.pp is default. But for me it look like puppet also accept site.pp~ .
[2009/04/09 10:24:19] <joe-mac> that's odd, i have no idea on that one. my first inclination would be that it shouldn't, but who knows
[2009/04/09 10:26:33] <Volcane> philip__: I suspect you have a import "/path/to/manifests/*"
[2009/04/09 10:26:50] <Volcane> rather than *.pp
[2009/04/09 10:27:35] @ edwardam|zZzZz is now known as edwardam
[2009/04/09 10:28:32] <philipp_> Sorry I check it again. I was no correct in my statment ...
[2009/04/09 10:29:37] @ Quit: fujin:
[2009/04/09 10:29:39] <philipp_> But I had issues with some "*.pp~" . I think it was when I used import /path/to/manifests/*
[2009/04/09 10:30:54] @ chillitom joined channel #puppet
[2009/04/09 10:32:15] <jbooth> philipp_: Did you get my last message yesterday?
[2009/04/09 10:33:57] <jbooth> philipp_: about not needing to import custom, modules main init.pp get auto-imported these days, so if you have defines there it should be good.
[2009/04/09 10:34:22] @ ezekiel-v3 joined channel #puppet
[2009/04/09 10:35:11] <philipp_> Yes I got. Manwhile it is working. But it will only work If import custom in site.pp. My custom/manifest/init.pp looks like: http://pastebin.com/m655d0141
[2009/04/09 10:36:18] <philipp_> As soon as I remove "import custom" from site.pp. I got error messages can't find exec-remote-script.
[2009/04/09 10:37:31] @ bolt joined channel #puppet
[2009/04/09 10:38:13] <philipp_> I noticed in your code "http://pastie.org/440797" that you included custom inside of the define "concat_file"? I was surprised about this?
[2009/04/09 10:39:51] <blubb> joe-mac: Yes, I tried various receipts. But they seem to be incomplete. Tried the ResolvConf example, but is is unclear where the posted files should be stored. After putting all of them into the same dir, they were found but it simply resulted in syntax errors.
[2009/04/09 10:40:27] <joe-mac> blubb: care to paste a link to waht you're talking about?
[2009/04/09 10:40:27] @ lak joined channel #puppet
[2009/04/09 10:40:35] <joe-mac> i and others will surely help with what spare time we have
[2009/04/09 10:40:52] <blubb> http://reductivelabs.com/trac/puppet/wiki/Recipes/ResolvConf
[2009/04/09 10:40:53] <philipp_> jbooth: Sorry again a mistake.... As soon as I removed "include custom" from site.pp. I got error messages can't find exec-remote-script.
[2009/04/09 10:41:29] <joe-mac> as someone else mentionemd i was really foggy on puppet until i read jamesturnbull's book and then started trying stuff out, i'm alright at it now, not pro or anything, but it took a lot of tinkering
[2009/04/09 10:41:32] @ markl_ joined channel #puppet
[2009/04/09 10:41:56] <joe-mac> ok blubb you notice wherer it says content => template("filename.erb")
[2009/04/09 10:43:03] <joe-mac> the template function looks in templatedir for an erb template to compile. templatedir is defined in /etc/puppet/puppet.conf. the template function then will compile the erb template and it will serve out the file to the path specified. in this case the path is the name of the resource "/etc/resolv.conf"
[2009/04/09 10:43:53] <Volcane> as before, u need to learn the basics, then these things would be obvious
[2009/04/09 10:44:02] <Volcane> recipes are only useful if you already understand puppet
[2009/04/09 10:44:13] <Volcane> but thusfar you're unprepared to read the docs that will get you to the point
[2009/04/09 10:44:24] <Volcane> crawl, walk, run, etc
[2009/04/09 10:44:46] <joe-mac> yea, Volcane is right here, i really *really* recommend getting the book pulling strings with puppet. front to back it's tiny, but after you read it you will have a vague sense of what is going on
[2009/04/09 10:44:53] @ edwardam is now known as edwardam|walk
[2009/04/09 10:44:57] <philipp_> Volcane : Ok than I will use in future "import /path/to/manifest/*.pp". Anyway Im not using them anymore so often, because most of the stuff is located in own modules ...
[2009/04/09 10:44:59] <joe-mac> it's also not expensive, so well worth the investment
[2009/04/09 10:45:41] <Volcane> philipp_: yeah, also check for the same where you load the nodes though
[2009/04/09 10:45:49] <Volcane> unless you use a node tool
[2009/04/09 10:45:54] <blubb> Volcane: I perfectly understand. This is why I asked _WHERE_ I can learn the basics. The documentation on reductive labs site is only useful for people who are already puppet experts. All the basic information for getting started is missing or incomplete. Seems I will need to buy the book you're talked about.
[2009/04/09 10:46:11] <Volcane> blubb: I pasted you such a url
[2009/04/09 10:46:30] <philipp_> OK, I have to go. My train. Thanks again for the help!
[2009/04/09 10:46:35] <joe-mac> Volcane: "node tool" i am putting that on our whiteboard as word of the day cause it sounds filthy
[2009/04/09 10:46:48] @ Quit: philipp_: "Ex-Chat"
[2009/04/09 10:46:48] <Volcane> joe-mac: lol
[2009/04/09 10:49:40] <Volcane> blubb: but in case you missed it, here it is again http://www.devco.net/pubwiki/Puppet/GettingStarted
[2009/04/09 10:49:42] <Volcane> the page says
[2009/04/09 10:49:44] <Volcane> "It is often the case that a new user joins the irc channel and ask questions regarding a complex problem but without the background knowledge of doing simple things first with Puppet they do not understand the answers completely. This is frustrating for everyone involved.
[2009/04/09 10:49:52] <Volcane> I've written this guide for newcomers, it will guide you from starting with a blank system to a complete server and 1 or more clients that will have a centrally managed /etc/hosts file. "!
[2009/04/09 10:49:55] <Volcane> sounds familiar...
[2009/04/09 10:51:47] <blubb> joe-mac: Thanks for pointing me to "templatedir". This does not exist in my example puppet.conf (from version0.24.7).
[2009/04/09 10:52:02] <joe-mac> np
[2009/04/09 10:52:29] <blubb> This is exactly the problem: Basic things are undocumented. I'm using linux since 1992 and never needed to buy some book only to understand the basic things of a subsystem. Sigh :(
[2009/04/09 10:53:07] <Volcane> blubb: amazingly - and i do mean amazingly considering the current conversion - the document and i pasted 3 times to you now does cover templates, and setting up the template dir
[2009/04/09 10:53:16] <Volcane> blubb: so pretty please, go read it.
[2009/04/09 10:54:32] * jbooth can give a thumbs up to Volcane's getting started docs. "I had to learn the hard way before he wrote that."
[2009/04/09 11:02:25] @ fujin joined channel #puppet
[2009/04/09 11:03:57] @ Quit: fujin: Remote closed the connection
[2009/04/09 11:03:59] @ fujin joined channel #puppet
[2009/04/09 11:04:13] <blubb> Volcane: Sorry, for some reason I really missed the link you posted. At first glance, this looks very promising! Exactly the kind of documentation I would have expected to find on the main page. Will go to read and no longer bother you. Thanks a lot!
[2009/04/09 11:09:42] <epsas> /win/win 2
[2009/04/09 11:09:42] <epsas> /win/wi/win 2
[2009/04/09 11:10:08] <jbooth> blubb: the biggest thing that's trouble, even after Volcane's docs is that puppet's language is declaritive, not imperative. If you've had enough CS to make that statement make sense you'll be answering questions here in no time. ;-)
[2009/04/09 11:10:20] <hexasof1> (16:49:40) Volcane: blubb: but in case you missed it, here it is again http://www.devco.net/pubwiki/Puppet/GettingStarted
[2009/04/09 11:10:38] @ Quit: unxfrek: "Leaving"
[2009/04/09 11:10:44] <hexasof1> blubb: this one?
[2009/04/09 11:11:57] @ hexasof1 left channel #puppet ()
[2009/04/09 11:15:52] * Volcane thinks blubb needs a new irc client if he missed it 3 times :)
[2009/04/09 11:16:08] @ jmarki joined channel #puppet
[2009/04/09 11:16:08] @ fdz joined channel #puppet
[2009/04/09 11:16:45] @ Quit: jmarki: Client Quit
[2009/04/09 11:16:56] @ jmarki joined channel #puppet
[2009/04/09 11:17:31] @ Quit: jmarki: Client Quit
[2009/04/09 11:17:39] @ unxfrek joined channel #puppet
[2009/04/09 11:18:16] <blubb> blubb has never used irc before. He's an old man and thus is used to newsgroups and mailing lists :)
[2009/04/09 11:20:00] @ jmarki joined channel #puppet
[2009/04/09 11:20:48] @ Quit: PsychoSid: "Leaving."
[2009/04/09 11:21:27] <claymation> welcome to the 90s, blubb ... irc is a revolution (still) waiting to happen :)
[2009/04/09 11:22:35] <blubb> jbooth: Yes, I do have a little bit of experience with various languages. As usual, the biggest hurdle is starting. The recipes certainly will help a lot once one has understood the basic concepts. Maybe Volcane can persuade the reductive labs people to integrate his beginners' documentation. I know others had similar problems. Again: Big thank you!
[2009/04/09 11:24:11] <Volcane> the reductivelabs website is a wiki, so anyone can contribute
[2009/04/09 11:24:39] <jbooth> blubb: I've had plenty of imperative and functional programming background. Declaritive still messes with my head though.
[2009/04/09 11:24:46] <Volcane> I'll contribute my doc once i am 100% done with it, but also when the RL wiki is based on more tolerable software, cos I'd rather pull my toe nails than write trac wiki
[2009/04/09 11:27:24] <blubb> jbooth: You're not alone :)
[2009/04/09 11:28:43] <joe-mac> fyi, "lsddistcodename" is *not* a fact
[2009/04/09 11:29:04] <blubb> claymation: many people (including my kids) tried to convince me that irc is great. But I still prefer mail...
[2009/04/09 11:29:25] <claymation> blubb: irc is the twitter of email
[2009/04/09 11:29:30] <joe-mac> just pushed out a bunch of files all named wrong due to that typo :-(
[2009/04/09 11:29:38] <claymation> make of that what you will :)
[2009/04/09 11:29:43] <Volcane> joe-mac: lol
[2009/04/09 11:31:16] <blubb> Happy Easter to all of you. Especially to Volcane :)
[2009/04/09 11:31:19] @ snafu joined channel #puppet
[2009/04/09 11:31:34] <joe-mac> yea my pidgin lights up a different color and makes a noise whens omeone says my name in chat
[2009/04/09 11:31:40] <joe-mac> essential to ever responding to anything pretty much
[2009/04/09 11:32:30] <jtimberman> heh, i've been using IRC for 15 years.
[2009/04/09 11:33:07] <jtimberman> actually started using irc before email on my first 'net account :)
[2009/04/09 11:33:57] @ Quit: verwilst: "Ex-Chat"
[2009/04/09 11:34:13] <claymation> jtimberman: same here, actually
[2009/04/09 11:34:20] <claymation> jtimberman: back when I thought anonymous FTP was "hacking" :)
[2009/04/09 11:34:23] @ edwardam|walk is now known as edwardam
[2009/04/09 11:35:46] @ Quit: blubb: "http://www.mibbit.com ajax IRC Client"
[2009/04/09 11:36:18] <jtimberman> claymation: EFnet, ah how I don't miss it.
[2009/04/09 11:38:29] @ Quit: bolt: "telnet irc.freenode.net 6667"
[2009/04/09 11:38:43] @ bolt joined channel #puppet
[2009/04/09 11:38:57] <joe-mac> anybody managing their apt repos via puppet like the recipe shown on the wiki?
[2009/04/09 11:39:31] <joe-mac> nm just solved my own problem, cookie for me
[2009/04/09 11:41:24] @ Quit: andrewcshafer:
[2009/04/09 11:42:18] @ Quit: MarlondB:
[2009/04/09 11:42:42] @ Quit: raphink: Read error: 104 (Connection reset by peer)
[2009/04/09 11:43:09] @ raphink joined channel #puppet
[2009/04/09 11:45:24] @ nigelk joined channel #puppet
[2009/04/09 11:54:54] @ Quit: bolt: Remote closed the connection
[2009/04/09 11:55:57] @ Quit: jlamsens: "Leaving"
[2009/04/09 11:57:57] @ andrewcshafer joined channel #puppet
[2009/04/09 11:59:59] @ Quit: Innocenti: Client Quit
[2009/04/09 12:00:26] @ ethan_rowe joined channel #puppet
[2009/04/09 12:01:20] @ bolt joined channel #puppet
[2009/04/09 12:02:56] @ Quit: stijnbe: "Leaving..."
[2009/04/09 12:03:02] @ Quit: bolt: Client Quit
[2009/04/09 12:07:11] <nigelk> yo git experts :)
[2009/04/09 12:07:25] <nDuff> mrr?
[2009/04/09 12:07:50] <nigelk> looking for some general git advice as we're moving the debian facter stuff into a debian git repo the same way Puppet is being done...
[2009/04/09 12:07:57] <goozbach> nigelk: http://video.google.com/videoplay?docid=-3999952944619245780
[2009/04/09 12:08:06] <goozbach> talk by Randall Swartz about git
[2009/04/09 12:08:09] <nigelk> well not quite at that level :)
[2009/04/09 12:08:28] <nigelk> I do know how to use it, just have an issue looking for some advice on
[2009/04/09 12:08:39] * f3ew need to talk to nigelk about Google's programming "process"
[2009/04/09 12:08:47] * nDuff is waiting for a question in response to which general advice can be offered.
[2009/04/09 12:08:52] <nigelk> particularly if anyone has used the debian git-import-dsc stuff
[2009/04/09 12:09:06] <nigelk> ok, so here's the issue
[2009/04/09 12:09:16] <nigelk> git-import-dsc creates a local git repo based upon the dsc contents
[2009/04/09 12:09:20] <jbk> speaking of facter, is there any special process to get a new module(?) for facter contributed?
[2009/04/09 12:09:31] <gepetto> ::trac:: Downloading Puppet edited by jbarratt @ http://reductivelabs.com/trac/puppet/wiki/DownloadingPuppet (by jbarratt@serialized.net)
[2009/04/09 12:09:32] <nigelk> I then want to merge in the reductivelabs facter repo to bring it up to 1.5.4
[2009/04/09 12:09:37] @ Quit: MattyM: "ta ta"
[2009/04/09 12:09:53] <nigelk> The only way I've been able to do this and preserve the git log from reductivelabs is to:
[2009/04/09 12:10:14] <nigelk> * add remote for reductive
[2009/04/09 12:10:27] <nigelk> * git checkout -b workingmaster master
[2009/04/09 12:10:35] <nigelk> (master is the branch git-import-dsc creates)
[2009/04/09 12:10:40] <nigelk> * git rebase 1.5.4
[2009/04/09 12:11:12] <nigelk> this seems to do what I want, but leaves the git log with " Imported Debian patch 1.5.1-0.1" as the most recent change, with all the 1.5.4 stuff before it
[2009/04/09 12:11:16] <nigelk> which feels like the wrong way around
[2009/04/09 12:11:31] <nigelk> (even given that I'll be committing more changes after this step)
[2009/04/09 12:11:34] <nigelk> make sense?
[2009/04/09 12:11:45] <nDuff> it *is* what the operations you did do -- rebase puts the branch you're on *on top of* the thing you're rebasing it to
[2009/04/09 12:12:03] @ Quit: philip__: Read error: 145 (Connection timed out)
[2009/04/09 12:12:17] <nigelk> yes. I don't seem to be able to do this the other way around though, ie branch 1.5.4 and rebase master
[2009/04/09 12:13:19] @ masterzen joined channel #puppet
[2009/04/09 12:13:34] <nDuff> I'm not familiar enough with git-import-dsc to know what it does, and thus how the history of the branch it creates connects with everything underneath; however, if any problems are along those lines, have you tried an *interactive* rebase in the reverse direction?
[2009/04/09 12:13:42] @ Quit: chillitom: Remote closed the connection
[2009/04/09 12:13:42] <nDuff> this might be more a question for #git, btw
[2009/04/09 12:14:03] <nigelk> yeah, that's the next step
[2009/04/09 12:14:07] <nigelk> I just happened to be here already
[2009/04/09 12:14:17] <masterzen> lak: Beside finishing the REST auth layer, is there anything you'd like me to do on 0.25?
[2009/04/09 12:14:37] @ Quit: bje: Read error: 60 (Operation timed out)
[2009/04/09 12:14:56] <gepetto> ::trac:: Puppet Debian edited by jbarratt @ http://reductivelabs.com/trac/puppet/wiki/PuppetDebian (by jbarratt@serialized.net)
[2009/04/09 12:15:57] @ stijnbe joined channel #puppet
[2009/04/09 12:16:11] @ MarlondB joined channel #puppet
[2009/04/09 12:16:23] <lak> masterzen: #1909 would be great
[2009/04/09 12:16:24] <gepetto> lak: masterzen: #1909 is http://projects.reductivelabs.com/issues/show/1909 "Puppet - Refactor #1909: Add 'rack' http handler for RESTful support - ReductiveLabs.com"
[2009/04/09 12:16:44] <masterzen> lak: ouch, looks like a tough one, no?
[2009/04/09 12:16:46] <lak> we can't actually replace all of the handlers, because we don't want to add rack as a requirement for puppetmaster
[2009/04/09 12:16:50] <lak> i don't think so
[2009/04/09 12:17:04] <lak> it's only two classes, and you can use the existing mongrel/webrick stuff as a good example
[2009/04/09 12:17:16] <lak> we'll keep webrick around, but we could then get rid of mongrel if we wanted
[2009/04/09 12:17:43] <lak> i'm bumping #1885
[2009/04/09 12:17:44] <gepetto> lak: #1885 is http://projects.reductivelabs.com/issues/show/1885 "Puppet - Bug #1885: Requires are passed down to defines and exported with resources - ReductiveLabs.com"
[2009/04/09 12:18:02] <masterzen> I'll study rack to see how it works. I'm really ignorant on this topic. I'm not even sure to know what rack is :-)
[2009/04/09 12:18:40] <lak> it's already used in 0.24.x
[2009/04/09 12:19:52] <lak> masterzen: http://github.com/lak/puppet/blob/88aa1bc2cfef4517fca9e467c86f70ae356be883/lib/puppet/network/http_server/rack.rb
[2009/04/09 12:20:40] <lak> compared to http://github.com/lak/puppet/blob/88aa1bc2cfef4517fca9e467c86f70ae356be883/lib/puppet/network/http/mongrel.rb
[2009/04/09 12:21:13] <masterzen> ah ok, that should be finger in the nose, then :-)
[2009/04/09 12:21:35] @ Quit: DerekW: "Leaving"
[2009/04/09 12:22:23] <masterzen> lak: I'll post the patch before the end of the week-end (yes I'm on a short deserved vacation starting tonight after several 15h days at the office :-))
[2009/04/09 12:23:10] <lak> awesome on both counts
[2009/04/09 12:23:14] <lak> vacation would be nice
[2009/04/09 12:23:20] <lak> i hope to get all of those bugs fixed this week
[2009/04/09 12:23:23] <lak> hmm
[2009/04/09 12:23:40] <masterzen> what's your planned release date for 0.25?
[2009/04/09 12:23:43] <lak> i've been thinking of adding some kind of routing support
[2009/04/09 12:23:47] <masterzen> loosely I mean
[2009/04/09 12:23:48] <lak> as soon as that bug list is fixed :/
[2009/04/09 12:24:10] <lak> and then some node/fact modification i need to make
[2009/04/09 12:24:13] <masterzen> can you elaborate on the routing thing?
[2009/04/09 12:24:34] <lak> gimme a sec, on the phone
[2009/04/09 12:24:48] @ markl__ joined channel #puppet
[2009/04/09 12:25:40] <masterzen> yep sure
[2009/04/09 12:27:02] <lak> basically, we've got hard-coded settings in puppetd, puppetmasterd, puppet, etc. that determine what terminus classes are used by what indirected classes
[2009/04/09 12:27:51] <lak> i was thinking of an external yaml file that could be used to configure this
[2009/04/09 12:27:54] <masterzen> yes I noticed that. I understand what you want to achieve.
[2009/04/09 12:28:55] <Volcane> in addition to the other config file?
[2009/04/09 12:29:05] <nigelk> lak: with http://projects.reductivelabs.com/issues/2142 is there any point me patching 0.24.x and 0.25 ?
[2009/04/09 12:29:40] <nigelk> note the NDA^H^H^H use of the word "hypothetically"
[2009/04/09 12:30:21] @ Quit: alban2: Read error: 110 (Connection timed out)
[2009/04/09 12:31:00] @ Quit: masterzen: "Leaving"
[2009/04/09 12:31:50] @ WALoeIII joined channel #puppet
[2009/04/09 12:32:48] <lak> nigelk: we'll only have another 0.24.x release if something is drastically wrong with 0.24.8
[2009/04/09 12:32:53] <lak> which so far does not appear to be the case
[2009/04/09 12:32:56] <nigelk> yeah that's what I assumed
[2009/04/09 12:33:00] <lak> Volcane: yeah, it would have to be an additional config file
[2009/04/09 12:33:07] <lak> because the current one doesn't support rich data structures
[2009/04/09 12:33:13] <nigelk> I may end up having to distribute a patched version after OS X LolCat is released
[2009/04/09 12:33:14] <lak> which is essentially what we need
[2009/04/09 12:33:20] <nigelk> as otherwise puppet is completely broken on it
[2009/04/09 12:33:20] <Volcane> oh :(
[2009/04/09 12:33:22] <lak> LolCat?
[2009/04/09 12:33:31] <nigelk> Snow Leopard
[2009/04/09 12:33:35] <lak> heh
[2009/04/09 12:33:36] <lak> nice
[2009/04/09 12:33:57] <nigelk> I'm trying to get a meme started for the next unreleased version of OS X always being referred to as LolCat
[2009/04/09 12:34:03] @ Quit: fujin: Read error: 110 (Connection timed out)
[2009/04/09 12:34:07] <nigelk> must post more on 4chan
[2009/04/09 12:37:17] <f3ew> nigelk sounds interesting
[2009/04/09 12:37:23] <f3ew> The LolCat OX
[2009/04/09 12:37:34] <nigelk> f3ew: what was the question you wanted to ask?
[2009/04/09 12:38:15] @ claymation_ joined channel #puppet
[2009/04/09 12:39:02] <f3ew> nigelk, I have a developer here with no operations experience who thinks that doping what Google does is always right, regardless of the local ops expertise or problem to be solved
[2009/04/09 12:39:20] * f3ew was wondering about the whole initial release thing with sysadminjs and developers working together
[2009/04/09 12:39:31] <f3ew> A blog entry or something about it would be nice
[2009/04/09 12:39:49] <nigelk> what initial release thing? I'm a bit confused
[2009/04/09 12:39:51] <f3ew> If only so that I could tell people like that "See why you can't do what Google does"
[2009/04/09 12:40:52] @ bolt joined channel #puppet
[2009/04/09 12:41:58] <f3ew> nigelk, the Site Reliability Engineering stuff
[2009/04/09 12:42:03] <nigelk> ah
[2009/04/09 12:42:21] <nigelk> so SREs blur the line between sysadmin and programmer more than in most places I think
[2009/04/09 12:42:47] <f3ew> Right
[2009/04/09 12:43:11] <nigelk> so if your point is that developers with no operations experience shouldn't override operational expertise, I agree with you
[2009/04/09 12:43:23] * f3ew has been trying to point out that you have a SRE team, which we don't (I play that role)
[2009/04/09 12:43:40] <nigelk> the flip side is that our SWEs (more traditional programmers) are also expected to rotate through SRE training
[2009/04/09 12:43:43] <f3ew> But the culture here is development overrides ops
[2009/04/09 12:43:53] @ ezmob joined channel #puppet
[2009/04/09 12:43:54] <nigelk> so our developers are forced into operational experience
[2009/04/09 12:43:57] <f3ew> which does not happen
[2009/04/09 12:44:02] <nigelk> f3ew: like many places unfortunately
[2009/04/09 12:44:08] <f3ew> Yup
[2009/04/09 12:44:33] <nigelk> I mean we still have the problem of developers who are smart people who think they know all about operational considerations
[2009/04/09 12:44:37] <tim|imac> we actually adopt a "sure, we'll do it your way, but then you're responsible"-thingy with such attitudes
[2009/04/09 12:44:48] <nigelk> when they have little experience, but the culture tries to change that
[2009/04/09 12:45:01] <nigelk> tim|imac: that's kind of one thing we do
[2009/04/09 12:45:21] <nigelk> if you develop something, you're on call for it until it passes the service tests such that SREs will take it oncall
[2009/04/09 12:45:44] <nigelk> and even after that if it pages too much, or is flaky, or whatever, pages will get bounced back to the developers until they fix it
[2009/04/09 12:45:44] <tim|imac> usually people back off and actually listen to reason then... since our core product is "managed operations", in the end they usually understand they outsource that part because of expertise issues
[2009/04/09 12:46:24] <tim|imac> yeah, and "mongrel sucks, switch to passenger" is *not* a proper response
[2009/04/09 12:47:04] <f3ew> Mind if I distribute this conversation internally?
[2009/04/09 12:48:21] <nigelk> I'd rather not be directly quoted but feel free to convey the gist of it :)
[2009/04/09 12:48:35] <f3ew> Right
[2009/04/09 12:48:38] @ Quit: mikepea: Read error: 110 (Connection timed out)
[2009/04/09 12:48:52] @ Quit: siezer: "[BX] *PHEAR MY BITCHX-NESS!* BOW TO ME!"
[2009/04/09 12:48:59] <gwar9999> where do I need to place an erb function that I'd like to be able to access from a number of different puppet modules? And then how would I reference it from erb? Currently, the function is embedded in a single erb but I'd like to use it elsewhere w/o copying/pasting.
[2009/04/09 12:49:46] <Volcane> I think normal ruby search paths would apply and std "require 'foo'" would work inside a erb
[2009/04/09 12:49:50] <Volcane> but not tried it
[2009/04/09 12:50:26] <gwar9999> Volcane: thanks. I'll give that a try. Always good to have a decent starting point of where to look.
[2009/04/09 12:53:05] @ PhabX joined channel #puppet
[2009/04/09 12:53:18] @ Quit: claymation: Read error: 110 (Connection timed out)
[2009/04/09 12:53:24] @ philip__ joined channel #puppet
[2009/04/09 12:54:52] <Volcane> yup, looks like it works fine
[2009/04/09 12:55:42] @ Quit: drmikecrowe: Success
[2009/04/09 12:58:30] @ jsm joined channel #puppet
[2009/04/09 13:01:16] @ claymation_ is now known as claymation
[2009/04/09 13:04:05] @ Quit: fdz: "Leaving."
[2009/04/09 13:07:13] @ jtimberman left channel #puppet ()
[2009/04/09 13:13:14] @ ed209 joined channel #puppet
[2009/04/09 13:13:26] <scooby2> Does openvpn support any "2 factor" authentication such as SecurID, Vasco, etc?
[2009/04/09 13:13:53] <Volcane> the next version has a really sweet system to do that, in the current versions there are some hacky solutions
[2009/04/09 13:15:10] <scooby2> Anything better out there for us open source types? SecurID is ridiculously expensive.
[2009/04/09 13:15:34] <Volcane> wikid seems nice
[2009/04/09 13:15:51] <scooby2> Volcane: thanks!
[2009/04/09 13:16:05] <Volcane> I've not used it, but really keen to give it a go some time
[2009/04/09 13:17:03] <scooby2> looks like it would do the job
[2009/04/09 13:17:30] <Volcane> http://www.openvpn.net/index.php/documentation/howto.html#pkcs11 thats for 2.1
[2009/04/09 13:17:41] <Volcane> which will support stuff like smartcards and all that
[2009/04/09 13:18:24] <nDuff> I've got a configuration that was successfully retrieved and applied on one run, but which on the next run -- with no changes -- reported a massive set of circular dependencies (with a 457,000 character line listing them!). This happened with puppet 0.24.8; is there a way it could actually be a configuration bug, or is it certainly a puppet thing? If the latter, anyone know if it's been fixed since?
[2009/04/09 13:18:28] @ Quit: markl__: "Lost terminal"
[2009/04/09 13:18:52] <scooby2> i see
[2009/04/09 13:18:56] @ Quit: andrewcshafer:
[2009/04/09 13:20:28] @ Quit: alfism: "Connection reset by beer"
[2009/04/09 13:20:32] <Volcane> nDuff: ouch, reading those things are a pain :(
[2009/04/09 13:21:03] @ Quit: ezmob: "Bye!"
[2009/04/09 13:21:10] @ ezmob_ joined channel #puppet
[2009/04/09 13:21:10] <nDuff> Volcane, indeed. We have a script to put them into graphviz-parsable form, but for that one I doubt that even having a graph will do much good.
[2009/04/09 13:21:15] <Volcane> i remember someone made soething to create a graphviz
[2009/04/09 13:21:17] <Volcane> ah there u go heh
[2009/04/09 13:22:51] @ Quit: madrescher: Read error: 110 (Connection timed out)
[2009/04/09 13:23:59] @ Quit: stijnbe: "Leaving..."
[2009/04/09 13:24:30] <nDuff> ...btw, has anyone given though to applying David Schmitt's patch sorting entries by .to_s before running the topo sort? Having the same inputs only occasionally result in an error is... disconcerting, especially when I have my QA automation hat on. :)
[2009/04/09 13:24:58] @ Quit: ezekiel-v3: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:24:58] @ Quit: n0ts_____: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:24:59] @ Quit: jhulten: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:24:59] @ Quit: ivan: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:25:00] @ Quit: auxesis: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:25:00] @ Quit: hamish: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:25:00] @ Quit: Cuchulain: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:25:00] @ Quit: tehhobbit: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:25:01] @ Quit: euclid: kubrick.freenode.net irc.freenode.net
[2009/04/09 13:30:22] @ jhulten joined channel #puppet
[2009/04/09 13:31:23] <joe-mac> i am about to write an openbsd module, has anybody already done this so i don't re-create someone else's effort?
[2009/04/09 13:34:14] <nDuff> hrm -- I don't see a 0.24.8 tag in the puppet git repo, only 0.24.8rc1; are they the same?
[2009/04/09 13:34:42] <Volcane> there were a few rcs
[2009/04/09 13:34:49] <Volcane> so probably not where you're looking :P
[2009/04/09 13:35:27] * nDuff is looking in "git tag -l" after running a fetch from git://reductivelabs.com/puppet
[2009/04/09 13:46:07] @ andrewcshafer joined channel #puppet
[2009/04/09 13:50:04] @ flakrat joined channel #puppet
[2009/04/09 13:57:16] @ freiheit joined channel #puppet
[2009/04/09 13:59:52] @ mccune joined channel #puppet
[2009/04/09 14:01:08] @ Quit: keithlard:
[2009/04/09 14:02:32] @ hamish joined channel #puppet
[2009/04/09 14:03:13] @ auxesis joined channel #puppet
[2009/04/09 14:06:07] @ ivan joined channel #puppet
[2009/04/09 14:08:53] @ msf joined channel #puppet
[2009/04/09 14:08:53] @ mcbride joined channel #puppet
[2009/04/09 14:09:06] @ ezekiel-v3 joined channel #puppet
[2009/04/09 14:09:06] @ n0ts_____ joined channel #puppet
[2009/04/09 14:09:06] @ Cuchulain joined channel #puppet
[2009/04/09 14:09:06] @ euclid joined channel #puppet
[2009/04/09 14:09:06] @ tehhobbit joined channel #puppet
[2009/04/09 14:09:25] @ tuf_ joined channel #puppet
[2009/04/09 14:10:03] @ Quit: tuf: Read error: 104 (Connection reset by peer)
[2009/04/09 14:13:14] @ pleemans joined channel #puppet
[2009/04/09 14:16:34] @ Quit: lak:
[2009/04/09 14:16:59] @ lak joined channel #puppet
[2009/04/09 14:17:16] @ Quit: lak: Client Quit
[2009/04/09 14:19:20] @ alfism joined channel #puppet
[2009/04/09 14:20:49] @ bevans5446 joined channel #puppet
[2009/04/09 14:25:02] <cwebber> has anyone seen this issue? Could not request certificate: Certificate retrieval failed: Certificates were not trusted: SSL_connect returned=1 errno=0 state=SSLv3 read finished A: sslv3 alert bad record mac
[2009/04/09 14:28:29] @ Quit: jmarki: "Leaving"
[2009/04/09 14:30:15] @ rgsteele||work joined channel #puppet
[2009/04/09 14:30:37] <rgsteele||work> Um, does this not hold true anymore? http://reductivelabs.com/trac/puppet/wiki/CommonMisconceptions#class-inheritance-and-variable-scope
[2009/04/09 14:31:00] <rgsteele||work> The hack to circumvent the variable scope by including the base class?
[2009/04/09 14:31:02] <rgsteele||work> After upgrading to 0.24.8, it seems not to.
[2009/04/09 14:32:24] @ Quit: plathrop: "leaving"
[2009/04/09 14:32:41] @ plathrop joined channel #puppet
[2009/04/09 14:33:54] @ Quit: bevans5446:
[2009/04/09 14:35:17] @ lak joined channel #puppet
[2009/04/09 14:35:23] @ Quit: plathrop: Client Quit
[2009/04/09 14:35:49] @ plathrop joined channel #puppet
[2009/04/09 14:36:59] @ Quit: tuf_: Read error: 110 (Connection timed out)
[2009/04/09 14:37:46] @ Quit: PhabX: "Leaving..."
[2009/04/09 14:47:57] <joe-mac> anybody know if you need to set the PKG_PATH yoruself before puppet will workw ith the openbsd package provider?
[2009/04/09 14:48:00] @ coofamani joined channel #puppet
[2009/04/09 14:51:57] @ mikepea joined channel #puppet
[2009/04/09 14:58:29] @ alban2 joined channel #puppet
[2009/04/09 14:59:05] @ nakano is now known as nakano_
[2009/04/09 15:04:23] @ Quit: pleemans: Read error: 110 (Connection timed out)
[2009/04/09 15:09:46] @ Quit: lak:
[2009/04/09 15:11:09] @ Quit: andrewcshafer:
[2009/04/09 15:12:54] <jrojas> cwebber: that looks like something is funky with your cers
[2009/04/09 15:12:57] <jrojas> *certs
[2009/04/09 15:17:51] @ bevans5446 joined channel #puppet
[2009/04/09 15:19:52] @ Quit: mikepea:
[2009/04/09 15:21:41] @ scooby2 left channel #puppet ()
[2009/04/09 15:24:26] <jrojas> anyone around?
[2009/04/09 15:24:38] @ andrewcshafer joined channel #puppet
[2009/04/09 15:36:04] <joe-mac> anybody know how a good way to only serve a file if a user has been realized?
[2009/04/09 15:36:09] <joe-mac> o nm
[2009/04/09 15:39:41] <joe-mac> yea i'm running into an issue where a file goes to be servbed out before a resource is realized
[2009/04/09 15:39:55] <joe-mac> anybody know a good way ? i require the resource but it says cannot find uit
[2009/04/09 15:41:18] <joe-mac> the onmly thing i can think of is put the file inside the define of the virtual resource
[2009/04/09 15:41:23] <joe-mac> hey Volcane you around?
[2009/04/09 15:44:02] <plathrop> joe-mac: that's the way
[2009/04/09 15:44:08] <plathrop> joe-mac: put the file into a define with the user
[2009/04/09 15:44:14] <plathrop> and virtualize the define
[2009/04/09 15:45:12] <joe-mac> yea, only problem is it's a define called 'admin' for people in my group it pushes out our ssh keys and whatnot and i am the only one witha customized bashrc, before today they weren't virtual resources, now that they are i am finding it tough to push out a .bashrc only on a specific user
[2009/04/09 15:45:24] <joe-mac> i guess i can add a parameter to the define
[2009/04/09 15:46:32] @ Quit: andrewcshafer:
[2009/04/09 15:47:17] @ Quit: Bass10: "Leaving"
[2009/04/09 15:48:26] @ tuf joined channel #puppet
[2009/04/09 15:48:38] <coofamani> is there any way to ignore certification errors? I assume it's doing a DNS lookup manually or something since my hosts files agree with the server names. Just in development, dont need good certs
[2009/04/09 15:49:01] @ stijnbe joined channel #puppet
[2009/04/09 15:50:07] <joe-mac> coofamani: afaik they go by certname
[2009/04/09 15:50:18] <coofamani> hmm, ok. thanks
[2009/04/09 15:50:38] <joe-mac> coofamani: i say that because at my DC i don't have a DNS box for those machines
[2009/04/09 15:50:41] <joe-mac> and they still seem to work fine
[2009/04/09 15:51:21] <coofamani> "Could not retrieve catalog: Certificates were not trusted: hostname not match with the server certificate"
[2009/04/09 15:51:29] <coofamani> odd, all the names seem to be correct
[2009/04/09 15:56:06] <coofamani> from the documentation "The default server for puppetd is puppet, so you could just create a CNAME of that to whatever server is running puppetmasterd"
[2009/04/09 15:56:15] <coofamani> so Im guessing this is hardcoded somewhere
[2009/04/09 15:56:25] <cwebber> puppetd and puppetmasterd should be able to run on the same box right?
[2009/04/09 15:57:23] <joe-mac> yea, i set up m,y puppetmasters with puppet on the same box
[2009/04/09 15:57:56] <joe-mac> coofamani: yea that could be coming from yourt puppet box, is it not puppet.$domain?
[2009/04/09 15:59:42] <coofamani> no, its not
[2009/04/09 15:59:54] <coofamani> puppetmaster.*
[2009/04/09 16:00:24] <joe-mac> so if you do host puppetmaster.whatever.com from your node, does it resolve?
[2009/04/09 16:00:37] <joe-mac> set servername=puppetmaster on the CLI if it does
[2009/04/09 16:00:54] <coofamani> yeah
[2009/04/09 16:01:22] <joe-mac> if you run puppet without specifying that in the config or on the command line it defaults to puppet.$domain
[2009/04/09 16:01:28] @ Quit: MarlondB:
[2009/04/09 16:01:41] <coofamani> got it, I was doing --server _ip_
[2009/04/09 16:01:56] <coofamani> using the fqdn works, wasnt thinking thats where the cert comparison was coming from
[2009/04/09 16:01:59] <coofamani> thanks!
[2009/04/09 16:02:04] <joe-mac> np
[2009/04/09 16:04:09] @ Quit: aymerick: "kit mais sage"
[2009/04/09 16:07:03] @ Bass10 joined channel #puppet
[2009/04/09 16:08:31] @ Quit: Bass10: SendQ exceeded
[2009/04/09 16:09:08] @ Bass10 joined channel #puppet
[2009/04/09 16:13:06] @ Quit: rgsteele||work: Remote closed the connection
[2009/04/09 16:13:32] <joe-mac> will ensure present overwrite a non-zero file to be a zero file if you don't specify any content or a source?
[2009/04/09 16:26:56] @ yarihm joined channel #puppet
[2009/04/09 16:27:16] <jrojas> joe-mac: afaik no
[2009/04/09 16:30:00] @ Quit: tuf: Read error: 110 (Connection timed out)
[2009/04/09 16:30:27] @ kibahop joined channel #puppet
[2009/04/09 16:30:36] @ kibahop left channel #puppet ()
[2009/04/09 16:34:52] @ lak joined channel #puppet
[2009/04/09 16:36:00] @ mfoster left channel #puppet ()
[2009/04/09 16:40:50] <jrojas> lak: would template concatenation fail if template X out of templates X Y and Z didnt exist?
[2009/04/09 16:41:04] <lak> yes
[2009/04/09 16:41:10] <lak> all specified templates must exist
[2009/04/09 16:41:16] <jrojas> hmm
[2009/04/09 16:42:33] <joe-mac> jrojas: i've asked for the same thing :-)
[2009/04/09 16:42:38] <jrojas> im trying to figure out a way to effectively generate a sudoers file based on hostnames and include the appropriate groups where need be, but its seems like I run into complications left and right.
[2009/04/09 16:43:05] <joe-mac> well, sudoers has a hostname field, that's not adequate?
[2009/04/09 16:43:23] <Volcane> if file() could support relative paths it would be easier :(
[2009/04/09 16:43:28] <jrojas> joe-mac: note for 2k servers
[2009/04/09 16:43:35] <jrojas> note = not
[2009/04/09 16:43:57] <jrojas> if sudo 1.7.0 supported glob includes it would work
[2009/04/09 16:44:19] <jrojas> it already supports #include /etc/sudoers.d/local_sudo
[2009/04/09 16:44:25] <Volcane> jrojas: building the file from snippets is about the most viable at this point :(
[2009/04/09 16:45:08] <jrojas> jrojas: yes and no, if i can find a way to template the /etc/sudoers file to have an include line for each file in /etc/sudoers.d/* then I would be squared
[2009/04/09 16:45:18] <jrojas> heh im talking to myself now
[2009/04/09 16:46:02] <Volcane> jrojas: there are a few samples of defined types out there that takes a sudoers.d style directory and concat them to make a resulting file
[2009/04/09 16:46:45] @ WALoeIII_ joined channel #puppet
[2009/04/09 16:47:01] <jrojas> Volcane: hmm that may be an idea, but for example, if i have a host that is a php server, I only want my team (group admins) to be able to sudo, and I dont want to include the dba group in the sudoers file
[2009/04/09 16:47:21] <Volcane> nods, so i have a thing for my motd's for example
[2009/04/09 16:47:34] <Volcane> my iptables module does something like : motd::register{"iptables module": }
[2009/04/09 16:47:45] <Volcane> which then puts the text "iptables module" into my motd in the right place
[2009/04/09 16:47:57] <Volcane> what motd::register{} does is just put down a file into the motd.d directory
[2009/04/09 16:47:58] <jrojas> interesting.
[2009/04/09 16:48:05] <Volcane> which then triggers the various foo to build the motd
[2009/04/09 16:48:50] <Volcane> http://nephilim.ml.org/~rip/puppet/concatfile/ thats the code for it, uses a little helper shell script that goes out to the machines
[2009/04/09 16:49:58] @ Quit: egerlach: "leaving"
[2009/04/09 16:51:56] <jrojas> hm..that may have some ideas in it..
[2009/04/09 16:52:09] @ Quit: stijnbe: "Leaving..."
[2009/04/09 16:52:25] <lak> yeah, it's usually a better approach to concatenate on the client than the server
[2009/04/09 16:53:02] @ coofamani left channel #puppet ()
[2009/04/09 16:53:06] <Volcane> the concat in the link there deals with simply removing the include line, it'll go and clean up orphaned files and rebuild etc
[2009/04/09 16:53:22] <Volcane> it works really well, though on machines with *many* snippets, it does involve a performance hits
[2009/04/09 16:53:38] <jrojas> lak: i saw that you had rejected a ticket or two saying that you didnt like the concat idea of templates/source files
[2009/04/09 16:53:39] <Volcane> like most things when overused :P
[2009/04/09 16:54:13] @ philip___ joined channel #puppet
[2009/04/09 16:54:16] <lak> ayup
[2009/04/09 16:54:26] <jrojas> lak: any specific reason why?
[2009/04/09 16:54:32] <bda> Is there a way to see what command() is executing?
[2009/04/09 16:54:41] <lak> because it encourages you to manage snippets of text instead of resources
[2009/04/09 16:54:43] <lak> bda: --debug
[2009/04/09 16:54:55] <bda> Debug isn't showing it in this instance.
[2009/04/09 16:54:56] <bda> Sec.
[2009/04/09 16:55:18] <Volcane> i tend to never just push out a bit of text into those dirs, always wrap them in defines to stick to the base concept of puppetness when i do use snippets
[2009/04/09 16:55:33] <Volcane> bacula::job bacula::client etc which then builds the files needed
[2009/04/09 16:56:28] <jrojas> hmm. does augeas manage a snipplet of text?
[2009/04/09 16:56:31] <Volcane> i imagine one day I could turn it into a proper type though, not sure, never looked at writing those
[2009/04/09 16:56:49] <bda> lak: http://gist.github.com/92732
[2009/04/09 16:57:00] <Volcane> jrojas: yes, if theres a supporting lens, however aegus also have the nasty side effect that you then manage just the snippets you know about rather than the hwole file
[2009/04/09 16:57:20] <Volcane> jrojas: someone could add to the file, and you would be non the wiser, very undesired in my mind
[2009/04/09 16:57:42] <Volcane> not sure if aegus supports purging all unmanaged resources in a particular file yet
[2009/04/09 16:57:47] <Volcane> i guess that would make it more bearable
[2009/04/09 16:57:58] <jrojas> Volcane: yeah i never quite understood the point of augeas.
[2009/04/09 16:58:00] <lak> bda: heh, that's because you're not executing the command
[2009/04/09 16:58:07] <lak> command(:foo) retrieves the path to foo
[2009/04/09 16:58:22] <lak> you just want 'svccfg :import, resource[:manifest]'
[2009/04/09 16:58:27] <lak> make it easy ;)
[2009/04/09 16:58:42] <bda> huh.
[2009/04/09 16:58:48] <danp> bda: you got told!
[2009/04/09 16:59:10] <bda> ;)
[2009/04/09 16:59:16] * bda knows '' Ruby still. :(
[2009/04/09 16:59:57] @ Quit: flakrat: "Leaving"
[2009/04/09 17:00:02] <lak> in this case it's more knowing puppet's idioms and whatnot
[2009/04/09 17:01:44] <jrojas> i guess the ideal solution to my problem would be to migrate to netgroup/ldap sudo
[2009/04/09 17:01:46] <bda> Suggested reading?
[2009/04/09 17:02:18] @ Quit: WALoeIII: No route to host
[2009/04/09 17:02:22] <bda> Also, thanks. :)
[2009/04/09 17:03:06] <Volcane> cost to deposit a check = 1.50
[2009/04/09 17:03:07] <Volcane> oops
[2009/04/09 17:03:23] <jrojas> Volcane: that sucks.
[2009/04/09 17:03:31] <nDuff> Is there a way to get stack traces when puppet reports an exception being thrown?
[2009/04/09 17:03:32] <Volcane> i suck at multi window irc clients :P
[2009/04/09 17:03:41] <Volcane> keep sending messages wrong place
[2009/04/09 17:03:45] <jrojas> nDuff: --trace
[2009/04/09 17:04:07] <jrojas> Volcane: ive been doing it all day as well. I think its the blood alcohol imbalance
[2009/04/09 17:04:21] <Volcane> heh
[2009/04/09 17:04:46] <Volcane> anyway, now that i sent the message i might as well give the context :) tax people sent me a £4.49 check
[2009/04/09 17:05:07] <Volcane> madness, who would go throguh the effort and expense to bother
[2009/04/09 17:05:24] <jrojas> well, it is 4.49 you didnt have before
[2009/04/09 17:05:48] <jrojas> i would sign it over to someone else and let them deal with it.
[2009/04/09 17:06:05] @ Quit: philip__: Read error: 110 (Connection timed out)
[2009/04/09 17:06:09] <jrojas> it would make a funny gift for someone's bday
[2009/04/09 17:06:16] <Volcane> hehe, its to my company as well, they'd probably make some little note and retaliate with a audit for good measure if i dont cash the damn thing :)
[2009/04/09 17:09:58] @ omry joined channel #puppet
[2009/04/09 17:12:08] <jbooth> Anyone tackeled the unix username=>email address problem in puppet?
[2009/04/09 17:12:17] <joe-mac> sigh, anyone at all work with openbsd and puppet?
[2009/04/09 17:12:22] * jbooth is kicking at a function.rb which is defeating him.
[2009/04/09 17:12:30] <joe-mac> jbooth: what do you mean?
[2009/04/09 17:12:33] <joe-mac> which problem i mean
[2009/04/09 17:13:53] <jbooth> List of sysadmin, have mailalias { root: recipient => "list.map { |name| name+"@site.com" } "
[2009/04/09 17:14:11] <jbooth> I'm trying to not repeat myself by having two lines, one for sudo, and one for mailalias.
[2009/04/09 17:14:55] <Volcane> so wrap it in a define, call the define once, define makes the 2 things needed?
[2009/04/09 17:16:28] <ch> if I'm using exported resources, should I be able to collect those on the same node as they were exported from?
[2009/04/09 17:17:02] <jbooth> Aha. args.map not args[0].map.
[2009/04/09 17:18:36] @ thatch_ joined channel #puppet
[2009/04/09 17:23:28] <jbooth> I need to find time to get back to that iptables type now that I've chewed through a ruby book.
[2009/04/09 17:25:16] @ falsch joined channel #puppet
[2009/04/09 17:26:38] <gwar9999> if one class installs and enables sendmail and another class installs and enables exim.... if both classes are included in the node, can the exim class disable the sendmail service? Or do I need to ensure only 1 class exists (either the one containing sendmail or the one containing exim)?
[2009/04/09 17:28:03] <Volcane> you can do like if defined(Class["sendmail"]) { }
[2009/04/09 17:28:06] <Volcane> i think, something like that
[2009/04/09 17:28:21] <ChoHag> Why would you define a node which installs and configures both sendmail and exim?
[2009/04/09 17:28:24] <Volcane> yeah
[2009/04/09 17:30:02] <gwar9999> currently, there is a base that includes sendmail stuff and I want to override that if an exim class exists. I suppose breaking out the sendmail stuff to another class might make more sense and then include only one, exim or sendmail... is that a better approach?
[2009/04/09 17:30:13] @ Quit: rsquared: "Leaving"
[2009/04/09 17:30:20] @ Quit: falsch: Remote closed the connection
[2009/04/09 17:30:43] @ Quit: ghenry: Remote closed the connection
[2009/04/09 17:32:11] <ChoHag> Almost certainly.
[2009/04/09 17:33:07] <gwar9999> Is it usual then to have tons of classes that just handle one thing (sendmail) rather than a class that handles multiple things (sendmail, ssh, apache, etc).
[2009/04/09 17:33:20] <Volcane> if the base includes sendmail you'll find it hard to disable it
[2009/04/09 17:33:33] <Volcane> cos things like if define() is order dependant
[2009/04/09 17:33:51] * Volcane has a fact on the nodes for this kind of thing
[2009/04/09 17:34:53] <gwar9999> ok, this is educational for me... thanks all. I guess I'll start breaking out things that may or may not be used on all hosts.
[2009/04/09 17:35:13] <Volcane> everything should be a module
[2009/04/09 17:35:17] <Volcane> that'll help already
[2009/04/09 17:35:21] <Volcane> avoid base node and inheritance
[2009/04/09 17:35:41] <Volcane> rather make a module with your roles as classes
[2009/04/09 17:35:49] <nDuff> I'm curious -- why doesn't puppet use the tsort standard library module?
[2009/04/09 17:36:20] <jrojas> Volcane: why avoid the base node?
[2009/04/09 17:36:21] <Volcane> and have those inherit from the base class, so you can use inheritance a bit better and not be bitten by the various ordering and oddness that comes from node inheriting from a base node
[2009/04/09 17:37:16] <Volcane> nDuff: heres a simple case with this exim/sendmail situation, if he just put include exim aboce his include roles::base he could easily use a simple if to prevent sendmail from going in
[2009/04/09 17:37:37] <Volcane> nDuff: lots of things just work a bit better in class inheritance than node inheritance
[2009/04/09 17:38:03] <Volcane> i dont know, maybe my mind just doesnt work in a way compatible with node inheritance, but its always seemed fairly fundemntally flawed
[2009/04/09 17:38:28] <lak> bda: did you email your patches to the list? i didn't see them
[2009/04/09 17:39:02] @ Quit: erm_: Read error: 110 (Connection timed out)
[2009/04/09 17:39:55] @ falsch joined channel #puppet
[2009/04/09 17:39:58] <gwar9999> I'm not using node inheritance at all... they all come from an external script.
[2009/04/09 17:40:45] <Volcane> yup, thats another reason
[2009/04/09 17:40:59] <Volcane> if you ever gonna move to a node classifier, ldap nodes or anything like that
[2009/04/09 17:41:06] <Volcane> best avoid node inheritance from the get go
[2009/04/09 17:41:58] <gwar9999> so the external script should dictate use: base, exim, apache classes right (where base contains some standard stuff that all hosts use)? and then another node might use base, sendmail, eginx classes instead, while others: base, oracle, etc.... Is this more-or-less a better approach?
[2009/04/09 17:43:12] <jrojas> Volcane: I just use the default node and I have a small bit of logic to include the appropriate things based on hostname. Much easier than listing 2000 servers :)
[2009/04/09 17:44:05] <Volcane> gwar9999: i have like roles::foo_server and just include those, for cases where i have tons of the same, otherwise i just have common::linux which include a ton of linuxy things and whatever makes the machine special then included in addition
[2009/04/09 17:44:16] <Volcane> I've a node classifier that let me define arbitrary groupsing of classes
[2009/04/09 17:44:33] <Volcane> so in the webui i can mkae a group called 'johns webserver' and just drop all the classes in there
[2009/04/09 17:44:42] <Volcane> johns webserver has properties etc, variables basically
[2009/04/09 17:45:12] <Volcane> and i can include it on a node, and the node can also have variables, the ordering of the variables are like the node can override whats in 'johns webserver'
[2009/04/09 17:45:33] <Volcane> so i can much easier model the nodes by just fiddling around the web app than muck about with inheritances
[2009/04/09 17:45:44] <bda> lak: For #2145 and #2146? Yeah.
[2009/04/09 17:45:44] <gepetto> bda: lak: #2145 is http://projects.reductivelabs.com/issues/show/2145 "Puppet - Bug #2145: daemontools.rb status function does not catch downed services - ReductiveLabs.com"
[2009/04/09 17:45:45] <gepetto> bda: lak: #2146 is http://projects.reductivelabs.com/issues/show/2146 "Puppet - Bug #2146: daemontools.rb does not need to unlink a service symlink to down a service - ReductiveLabs.com"
[2009/04/09 17:45:50] <bda> lak: http://groups.google.com/group/puppet-dev/browse_thread/thread/b465e318ffcfb76a#
[2009/04/09 17:45:54] <bda> lak: http://groups.google.com/group/puppet-dev/browse_thread/thread/8db151aaa1f9b264#
[2009/04/09 17:45:59] <gwar9999> so less inhertiance > lots of inheritance
[2009/04/09 17:46:12] <Volcane> i also find inheritance causes all sorts of confusion
[2009/04/09 17:46:16] <bda> afk for a bit.
[2009/04/09 17:46:21] <Volcane> you're happily editing a class foo
[2009/04/09 17:46:32] <Volcane> you look at it, you can see whats going to be affected by your changes
[2009/04/09 17:46:55] <gwar9999> I believe you wrt inheritance since we have a complex server farm here.
[2009/04/09 17:46:55] <Volcane> now, its totally not obvious without much grepping and mucking about that someone else inherited from foo to make a module missioncritical
[2009/04/09 17:47:15] <lak> bda: ah - i see; can you use git send-mail to send the patches? look at the 'mail_patches' Rakefile target in the main rakefile
[2009/04/09 17:47:16] <Volcane> and there you go, bang screwed up a whole lot of stuff cos the relationships are basically only exposed through copious amounts of grepping
[2009/04/09 17:47:32] <lak> composition is a big win over inheritance
[2009/04/09 17:47:37] <lak> i just wish i'd known that years ago
[2009/04/09 17:47:49] <lak> and one of puppet's problems is that it only supports inheritance, no composition
[2009/04/09 17:47:52] <lak> but we're working on fixing that
[2009/04/09 17:47:53] <Volcane> yeah, i avoid inheritance as much as possible (obivously :P)
[2009/04/09 17:48:02] <gwar9999> I'm glad I'm learning now in my early stages of puppet 101.
[2009/04/09 17:48:23] <joe-mac> anybody know how i can force openbsd to just install the damned package
[2009/04/09 17:48:30] <joe-mac> with puoppet i mean
[2009/04/09 17:48:35] <plathrop> gwar9999: on the other hand I make heavy use of inheritance
[2009/04/09 17:49:04] <joe-mac> err: //Node[opensores]/openbsd::packages/Package[screen]/ensure: change from absent to present failed: Execution of '/usr/sbin/pkg_add screen' returned 256: Ambiguous: screen could be screen-4.0.3p1 screen-4.0.3p1-shm screen-4.0.3p1-static
[2009/04/09 17:49:57] <gwar9999> plathrop: gulp. competing views. Are you also using external nodes? Is your server collection complex?
[2009/04/09 17:50:00] <jrojas> joe-mac: package{ screen: ensure => "version you want" }
[2009/04/09 17:50:02] <jenza> ensure => "screen-4.0.3p1",
[2009/04/09 17:50:11] <joe-mac> but that's not always going to be the case
[2009/04/09 17:50:23] <jenza> then create seperate packages
[2009/04/09 17:50:35] <plathrop> gwar9999: Yes, we use an external node source and I'd say our system is pretty darn complex
[2009/04/09 17:50:44] <plathrop> gwar9999: I don't use any *node* inheritance
[2009/04/09 17:50:49] <plathrop> gwar9999: That way lies madness
[2009/04/09 17:50:56] <Volcane> heh
[2009/04/09 17:50:58] <joe-mac> jenza: i was hoping this would be version agnostic for our BSD systems
[2009/04/09 17:51:01] <jrojas> plathrop: agreeed
[2009/04/09 17:51:08] <jenza> don't blame puppet
[2009/04/09 17:51:21] <joe-mac> i don't?
[2009/04/09 17:51:21] <jenza> blame your package repo maintainer
[2009/04/09 17:51:23] @ mfoster joined channel #puppet
[2009/04/09 17:51:27] <jenza> also, solaris :(
[2009/04/09 17:51:28] <joe-mac> lol, do you know openbsd?
[2009/04/09 17:51:33] <plathrop> gwar9999: Don't get me wrong, Volcane is super-smart
[2009/04/09 17:51:36] <jenza> oh, openbsd, ok
[2009/04/09 17:51:41] <gwar9999> plathrop: do you have some hosts use one mail server and others another mail server? Are you using inheritance in those cases?
[2009/04/09 17:51:48] <plathrop> gwar9999: Just saying that there are different ways to use the tool
[2009/04/09 17:52:08] <jenza> node "hosta" { include mailserver;}
[2009/04/09 17:52:11] <jenza> node "hostb" { include mailserver;}
[2009/04/09 17:52:17] <plathrop> gwar9999: No, but for example we have some SOLR servers, there are masters and slaves.
[2009/04/09 17:52:18] <Volcane> i think inheritance is very powerfull and often unavoidable, I've i think 5 or 6 modules that use it effectively
[2009/04/09 17:52:27] <Volcane> but my juniors know to leave them the hell alone :P
[2009/04/09 17:52:49] <plathrop> gwar9999: But there is a lot of overlap between the two and, furthermore, I want to be able to easily promote/demote nodes
[2009/04/09 17:53:05] <gwar9999> plathrop: I'm seeing lot's of different ways, but obviously some are better than others. At this point, since I'm new to puppetland, I'm trying to do things the right way early on.
[2009/04/09 17:53:08] <plathrop> gwar9999: So, we have a base "solr" class and a "master" sublass and "slave" subclass
[2009/04/09 17:53:22] <joe-mac> i'm out, openbsd has wore me out prematurely, see ya and thanks for the help
[2009/04/09 17:53:33] <plathrop> gwar9999: There are some definite universal best practices
[2009/04/09 17:53:34] <Volcane> plathrop: nods, ditto for my nameservers...
[2009/04/09 17:53:47] <plathrop> gwar9999: Put *everything* in a module.
[2009/04/09 17:53:54] <plathrop> gwar9999: Keep site.pp as empty as possible
[2009/04/09 17:54:01] <plathrop> Use external nodes
[2009/04/09 17:54:05] <plathrop> Don't use webrick
[2009/04/09 17:54:07] <gwar9999> I guess I just to need to know when inheritance is better than using diff modules/classes.
[2009/04/09 17:54:09] <Volcane> plathrop: bind::master and bind::slave set them up for diferent uses, but the same bind::zones work on either and does the right thing dependant on if its a bind::master or bind::slave
[2009/04/09 17:54:27] <Volcane> both bind::master and bind::slave inherits from a common class
[2009/04/09 17:54:30] <plathrop> Volcane: yeah, when you can write it that way it is really nice and elegant
[2009/04/09 17:54:58] <Volcane> and so i have things like xen::dom0bind (a local caching server for the vms on the dom0) that also inherits from that
[2009/04/09 17:55:02] <Volcane> and a few other bind related stuff
[2009/04/09 17:55:03] @ joe-mac left channel #puppet ()
[2009/04/09 17:55:20] <gwar9999> Right now, everything is in a module (one of about 5 modules) but I think I'm going to start breaking down the big (base) module into smaller pieces.
[2009/04/09 17:55:22] <Volcane> it saves a heck of a lot of time, but I'd hate to think if it breaks and i am not around cos other guys might struggle
[2009/04/09 17:55:23] <plathrop> gwar9999: Hrm. Well it is hard to give a definitive answer. I'd say that using inheritance makes sense when you really need inheritance features.
[2009/04/09 17:55:24] @ Quit: mccune: "Leaving."
[2009/04/09 17:55:54] <Volcane> yeah, use it when needed. not like some people who use it INSTEAD of include
[2009/04/09 17:55:57] <Volcane> sheesh
[2009/04/09 17:56:42] <gwar9999> ah, ok... basically, I guess, if I need inheritance use it... otherwise avoid it. Sendmail and exim are different packages so I should include one or the other.
[2009/04/09 17:56:46] <plathrop> I wish our solr module weren't so site-specific; I'd totally open-source it.
[2009/04/09 17:56:55] <plathrop> gwar9999: well...
[2009/04/09 17:57:06] <Volcane> gwar9999: sticking to the sendmail/exim thing, you could in theory have both of them inherit from mta
[2009/04/09 17:57:20] <Volcane> gwar9999: and then just override some key things - like which package to install and which service to start
[2009/04/09 17:57:22] <plathrop> gwar9999: Can I ask *why* you need both?
[2009/04/09 17:57:31] <plathrop> gwar9999: Different clients or something?
[2009/04/09 17:57:58] <plathrop> gwar9999: Usually standardizing on a set of tools is one way of doing things the right way from the start
[2009/04/09 17:58:05] <Volcane> ++
[2009/04/09 17:58:07] <gwar9999> some of the systems need exim, most sendmail. I think they are used for filtering or something-- not real sure since I didn't configure them.
[2009/04/09 17:59:33] <gwar9999> Volcane: I guess I could inherit from mta, but not sure what that buys me since they both have diff packages and services so I'd essentially have a base that did nothing. I could see if they had somethings in common but in my case they are completely different.
[2009/04/09 18:00:04] <plathrop> gwar9999: Well... sendmail and exim do the same things, so I don't think you can say they inherently need one or the other. If you are saying there are legacy systems layered on top that you need to be sure not to disturb...
[2009/04/09 18:00:11] <Volcane> your based could setup aliases that u need to be the same between both, but yeah its not an awesome example
[2009/04/09 18:00:33] <Volcane> I'd just make 2 modules and devise a way to choose
[2009/04/09 18:00:53] <Volcane> and once you're installing sendmail in your base node, and then do node inheritance
[2009/04/09 18:00:56] <Volcane> its basically game over
[2009/04/09 18:01:16] <plathrop> In order of preference here's what I'd do:
[2009/04/09 18:01:39] <plathrop> 1) Get rid of sendmail entirely (kill it with FIRE!) and just use exim everywhere.
[2009/04/09 18:01:54] <plathrop> 2) If that isn't possible, figure out *why* it isn't possible and fix the underlying issue.
[2009/04/09 18:02:31] @ mikepea joined channel #puppet
[2009/04/09 18:03:00] <plathrop> 3) If that isn't possible/practical, then write separate classes and only include the one you need. Make sure the classes are smart enough to *remove* the stuff from the other class so you can switch back and forth.
[2009/04/09 18:03:21] <gwar9999> plathrop: I can't kill sendmail (as much as I've never liked it-- always preferred qmail fwiw) but my company uses a number of mail servers for different purposes and it's beyond my control to mandate which. I'm sure some of it is legacy, but not positive.
[2009/04/09 18:03:26] <plathrop> (i.e. in the sendmail class do package { "exim": ensure => absent }
[2009/04/09 18:03:52] <plathrop> That sucks.
[2009/04/09 18:04:13] <plathrop> I hate being given responsibility without the requisite authority to meet the obligation.
[2009/04/09 18:04:41] <plathrop> You're in a tough position. Looks like you have to do it the hard way.
[2009/04/09 18:04:55] <Volcane> node foo { $mta = "sendmail" }
[2009/04/09 18:05:08] <Volcane> class mta { include($mta) }
[2009/04/09 18:05:09] <Volcane> :)
[2009/04/09 18:05:20] <plathrop> Volcane: nice
[2009/04/09 18:05:44] <plathrop> brb
[2009/04/09 18:05:46] @ Quit: plathrop: "leaving"
[2009/04/09 18:05:48] <Volcane> which afaik will even work with node inheritance?
[2009/04/09 18:06:07] <Volcane> cos the node foo can override vars set by the base? think so
[2009/04/09 18:06:18] @ plathrop joined channel #puppet
[2009/04/09 18:06:20] <gwar9999> Volcane: nice, but what is the difference of that example and in the external node config, classes: ['base', 'sendmail'] ? Either way, the node defines the parameter (mta = 'sendmail') or it's defined by the classes.
[2009/04/09 18:06:41] <Volcane> gwar9999: yup, just throwing in many different ways
[2009/04/09 18:06:46] @ Quit: plathrop: Client Quit
[2009/04/09 18:07:05] @ Quit: alban2: Read error: 113 (No route to host)
[2009/04/09 18:07:10] <gwar9999> Volcane: ok, just making sure both were fine. Looks like we pissed off plathrop :)
[2009/04/09 18:07:25] <gwar9999> or maybe it was just me.
[2009/04/09 18:07:27] <Volcane> gwar9999: ultimately i think its not so much that you should use the Best Practise, but that you should take from all the many Best Practises and make one that works for your way/style/mental process
[2009/04/09 18:07:46] @ plathrop joined channel #puppet
[2009/04/09 18:07:48] <gwar9999> "mental" being the operative word.
[2009/04/09 18:08:30] <gwar9999> Thanks for the discussion. I have some good ideas from it so hopefully I can put puppet to good use (until my next series of questions that is)
[2009/04/09 18:15:35] @ Quit: mikepea:
[2009/04/09 18:16:59] @ Quit: jsm: "Leaving."
[2009/04/09 18:17:20] @ alban2 joined channel #puppet
[2009/04/09 18:26:03] @ Quit: claymation:
[2009/04/09 18:31:52] @ andrewcshafer joined channel #puppet
[2009/04/09 18:35:28] @ Quit: thatch_: Remote closed the connection
[2009/04/09 19:04:26] @ Quit: philip___: Read error: 104 (Connection reset by peer)
[2009/04/09 19:12:10] @ Quit: unxfrek: Remote closed the connection
[2009/04/09 19:12:24] @ Quit: andrewcshafer: Read error: 110 (Connection timed out)
[2009/04/09 19:22:11] @ keithlard joined channel #puppet
[2009/04/09 19:23:05] @ Quit: WALoeIII_: "Bai."
[2009/04/09 19:45:54] @ andrewcshafer joined channel #puppet
[2009/04/09 19:46:20] <flashn> hmm, does puppet do parallel tasks on nodes?
[2009/04/09 19:46:22] @ jtimberman_ joined channel #puppet
[2009/04/09 19:47:27] <jrojas> flashn: because of dependencies, it really cant
[2009/04/09 19:49:08] @ fujin joined channel #puppet
[2009/04/09 19:50:52] @ Quit: yarihm: "Leaving"
[2009/04/09 19:52:25] @ Quit: ezmob_: "Bye!"
[2009/04/09 19:53:29] @ Quit: f3ew: Remote closed the connection
[2009/04/09 19:53:56] @ silassewell joined channel #puppet
[2009/04/09 19:55:58] <silassewell> Anyone know what restrictions there are on content served as a file via Puppet. So if I have a module with a file in it which I serve to my web set (SSL certificates for example) and I also have a DNS server; is it possible for the DNS puppet instance to access the SSL certificates?
[2009/04/09 19:59:02] @ Quit: andrewcshafer:
[2009/04/09 19:59:56] <flashn> jrojas, so it executes in the order you put them? trying to do log rotations with intervals (sleep) between nodes
[2009/04/09 20:00:47] <jrojas> flashn: no, nothing is in order, it creates its own order based on dependencies unless you specify a whole bunch of useless stuff
[2009/04/09 20:02:18] <nDuff> flashn, wouldn't that be a case where you'd have puppet (say) distributing a cron job with different minute offsets on different nodes?
[2009/04/09 20:03:13] <flashn> Im trying to make a global file, which can handle any kind of, lets say webfarms
[2009/04/09 20:04:41] <flashn> can you make dependencies for the rotations?
[2009/04/09 20:05:04] <flashn> or is it simply not a puppet task?
[2009/04/09 20:05:18] <flashn> just got started
[2009/04/09 20:06:01] <nDuff> flashn, I believe there's a recipe (in the FAQ, maybe?) documented for generating a different hash for each system based on hostname; use that hash when calculating the rotation time; it's not a "global file" that way in that it evaluates to the same thing on each machine, but you still only need to define it once.
[2009/04/09 20:06:27] <flashn> a sleep-function would work
[2009/04/09 20:06:37] <nDuff> flashn, using a puppet task to actually *do* the rotation, rather than schedule or set up the rotation, isn't the right tool for the job
[2009/04/09 20:07:45] <flashn> problem is I cant think of a way to make a crontab that would scale from like 2 to +50 servers
[2009/04/09 20:08:04] <flashn> well, just an idea I had
[2009/04/09 20:08:13] <nDuff> flashn, how *doesn't* using the puppet cron type with a hash-derived minute value scale?
[2009/04/09 20:08:37] @ daMaestro joined channel #puppet
[2009/04/09 20:08:37] <flashn> oh yeah, sorry
[2009/04/09 20:08:50] <daMaestro> is there any way to secure a given resource (file) to a certain set of nodes
[2009/04/09 20:09:42] <nDuff> daMaestro, you're the second person here with that question -- if you find an answer elsewhere, silassewell would be interested to know.
[2009/04/09 20:09:52] <flashn> nDuff, Ill check the recipes
[2009/04/09 20:09:54] <daMaestro> lol
[2009/04/09 20:10:01] <daMaestro> yeah.. he is sitting across the office from me
[2009/04/09 20:10:04] <daMaestro> har har har
[2009/04/09 20:10:11] * daMaestro stabs silassewell
[2009/04/09 20:10:25] <silassewell> I found http://groups.google.com/group/puppet-users/browse_thread/thread/fafa59ae7d5b50b3/06caf471da7e427e?lnk=gst&q=security+puppet+files#06caf471da7e427e
[2009/04/09 20:10:38] <silassewell> Just hoping something had been added since then
[2009/04/09 20:11:57] <fujin> you can use %h in the fileserver mounts to add a hostname-specific fileserver mount
[2009/04/09 20:12:14] <fujin> been a while since I've used puppet (I'm AJ on that thread) but I believe that was what I settled on
[2009/04/09 20:12:15] @ Quit: cwebber: Read error: 110 (Connection timed out)
[2009/04/09 20:13:17] <gwar9999> can I require a File and Package at the same time? Do I need 2 require lines or can they be combined into 1?
[2009/04/09 20:13:55] <nDuff> gwar9999, require => [ File["foo"], Package["bar"] ]
[2009/04/09 20:14:23] <gwar9999> cool, thanks mDuff
[2009/04/09 20:16:47] <flashn> where can I find any benchmark stuff? how many TPS etc
[2009/04/09 20:20:52] <jrojas> flashn: look up reports
[2009/04/09 20:22:41] @ Quit: keithlard:
[2009/04/09 20:27:02] @ unxfrek joined channel #puppet
[2009/04/09 20:29:43] <flashn> jrojas, just wanted some estimates so I know if I will need a cluster or just 2
[2009/04/09 20:30:44] <flashn> just got it running on two centos in vbox >:]
[2009/04/09 20:33:34] <jrojas> flashn: also look into environments
[2009/04/09 20:34:14] <flashn> transactions per server will more or less continously expand, and scalable for +1000 servers
[2009/04/09 20:34:22] <flashn> jrojas, ah, will do that
[2009/04/09 20:34:53] <jrojas> flashn: you also get great performance by putting puppetmaster behind passenger/apache
[2009/04/09 20:35:08] <jrojas> and its a nice lightweight setup
[2009/04/09 20:35:27] <flashn> thats cool
[2009/04/09 20:36:10] <flashn> you can have apache cache queries?
[2009/04/09 20:36:44] <jrojas> i havent tried, i dont really see a need for it
[2009/04/09 20:36:55] <flashn> hmm, I think I read something about internal caching
[2009/04/09 20:37:46] <flashn> but off-shoring the caching part is always a nice option when looking for scalability
[2009/04/09 20:43:46] @ Quit: cmoates: "Leaving"
[2009/04/09 20:44:31] @ spheromak joined channel #puppet
[2009/04/09 20:45:26] <flashn> gonna need git as well
[2009/04/09 20:46:15] @ friendly12345 joined channel #puppet
[2009/04/09 20:49:52] @ Quit: score: Read error: 110 (Connection timed out)
[2009/04/09 20:57:53] @ Quit: vinci: Remote closed the connection
[2009/04/09 21:02:18] @ andrewcshafer joined channel #puppet
[2009/04/09 21:03:13] @ Quit: nigelk:
[2009/04/09 21:08:12] @ jtimberman_ is now known as jtimberman
[2009/04/09 21:14:33] @ vinci joined channel #puppet
[2009/04/09 21:14:56] @ Quit: daMaestro: "Leaving"
[2009/04/09 21:18:18] @ Quit: fujin:
[2009/04/09 21:19:51] @ Quit: unxfrek: "Leaving"
[2009/04/09 21:30:27] @ Quit: alfism: "Connection reset by beer"
[2009/04/09 21:34:51] @ iFujin joined channel #puppet
[2009/04/09 21:36:28] @ Quit: ethan_rowe: "Lack of interest wins out."
[2009/04/09 21:38:09] @ Quit: iFujin: Client Quit
[2009/04/09 21:57:43] @ fujin joined channel #puppet
[2009/04/09 21:58:21] @ fujin_ joined channel #puppet
[2009/04/09 21:58:34] @ Quit: fujin: Client Quit
[2009/04/09 21:59:59] @ fujin_ is now known as fujin
[2009/04/09 22:09:24] @ score joined channel #puppet
[2009/04/09 22:22:35] @ Quit: msf: "leaving"
[2009/04/09 22:22:57] @ msf joined channel #puppet
[2009/04/09 22:31:54] @ nigelk joined channel #puppet
[2009/04/09 22:34:55] @ nigelk_ joined channel #puppet
[2009/04/09 22:35:17] @ maxagaz_ joined channel #puppet
[2009/04/09 22:35:20] @ Quit: maxagaz_: Client Quit
[2009/04/09 22:36:45] @ nigelk__ joined channel #puppet
[2009/04/09 22:40:37] @ Quit: bgupta: Read error: 104 (Connection reset by peer)
[2009/04/09 22:40:55] @ bgupta joined channel #puppet
[2009/04/09 22:51:31] @ mfoster left channel #puppet ()
[2009/04/09 22:52:19] @ Quit: nigelk: Read error: 110 (Connection timed out)
[2009/04/09 22:54:20] @ Quit: nigelk_: Read error: 110 (Connection timed out)
[2009/04/09 23:13:05] @ WALoeIII joined channel #puppet
[2009/04/09 23:13:34] @ Quit: WALoeIII: Client Quit
[2009/04/09 23:13:45] @ WALoeIII joined channel #puppet
[2009/04/09 23:16:32] @ alfism joined channel #puppet
[2009/04/09 23:28:30] @ Quit: kolla: Remote closed the connection
[2009/04/09 23:36:25] @ tuf joined channel #puppet
[2009/04/09 23:51:28] @ kolla joined channel #puppet
[2009/04/09 23:51:54] @ silassew1ll joined channel #puppet
[2009/04/09 23:57:10] @ Quit: maxagaz: Remote closed the connection

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