| [2010/01/07 00:02:31] @ Log started by gepetto | ||
| [2010/01/07 00:02:31] @ brd joined channel #puppet | ||
| [2010/01/07 00:03:28] @ Boltsky joined channel #puppet | ||
| [2010/01/07 00:18:39] @ Quit: lero: Read error: 60 (Operation timed out) | ||
| [2010/01/07 00:21:54] @ Quit: cwebber: | ||
| [2010/01/07 00:28:18] @ Demosthenes joined channel #puppet | ||
| [2010/01/07 00:33:06] @ Quit: ambroff: "Ex-Chat" | ||
| [2010/01/07 00:35:05] <jb> ohad? | ||
| [2010/01/07 00:42:45] @ Quit: pting: "Ex-Chat" | ||
| [2010/01/07 00:50:54] @ Quit: jcape1: Read error: 60 (Operation timed out) | ||
| [2010/01/07 01:09:01] @ Quit: Djelibeybi: "Leaving" | ||
| [2010/01/07 01:11:31] @ Quit: bodepd: | ||
| [2010/01/07 01:12:24] @ Quit: littleidea: | ||
| [2010/01/07 01:15:12] @ littleidea joined channel #puppet | ||
| [2010/01/07 01:20:59] @ vorlon2261 left channel #puppet () | ||
| [2010/01/07 01:24:38] @ _silver is now known as hy | ||
| [2010/01/07 01:25:50] @ nakano is now known as nakano_ | ||
| [2010/01/07 01:41:35] @ PaulWay[w] left channel #puppet ("I must go, there are too many tangrams on this channel.") | ||
| [2010/01/07 01:46:13] @ Quit: M-: "Leaving" | ||
| [2010/01/07 01:47:13] @ ibolmo joined channel #puppet | ||
| [2010/01/07 01:47:20] @ Quit: ibolmo: Client Quit | ||
| [2010/01/07 01:48:16] @ Quit: tuf: Read error: 60 (Operation timed out) | ||
| [2010/01/07 01:48:45] @ ibolmo joined channel #puppet | ||
| [2010/01/07 01:49:36] <ibolmo> hey guys, as far as I can gleem; puppet is meant unix systems.. is there a solution for setting up private lans, etc., with mixed platforms? windows + mac + linux | ||
| [2010/01/07 01:50:24] @ Quit: littleidea: | ||
| [2010/01/07 01:51:51] @ jcape joined channel #puppet | ||
| [2010/01/07 01:52:39] <SyTonnerre> ibolmo, there ought not be any problems running Puppet on MacOS >=10 | ||
| [2010/01/07 01:53:09] <ibolmo> well maybe i should clarify | ||
| [2010/01/07 01:53:17] <ibolmo> i'm at a small startup | ||
| [2010/01/07 01:53:21] <ibolmo> and we're starting to scale | ||
| [2010/01/07 01:53:27] <ibolmo> we'll be adding new computers for sales persons | ||
| [2010/01/07 01:53:52] <ibolmo> and well i would default to windows for suppor | ||
| [2010/01/07 01:54:00] <ibolmo> installing of programs, etc | ||
| [2010/01/07 01:54:08] <ibolmo> i know can be done through windows server and remote installation | ||
| [2010/01/07 01:54:31] <ibolmo> i'm wondering if puppet would be able to help with all of that | ||
| [2010/01/07 01:54:43] <ibolmo> setting up the ldap, and bunch of other things | ||
| [2010/01/07 01:55:29] <SyTonnerre> I don't think there's a way to make Puppet run on Windows, especially since Windows lacks any package management, but then again I might be wrong | ||
| [2010/01/07 01:56:26] <ibolmo> yeah i admit that i'm green behind the ears in this forefront | ||
| [2010/01/07 01:56:55] <ibolmo> it's pretty easy for me to run my own computer, but when it comes to dealing with other users within my network i haven't done anything related | ||
| [2010/01/07 01:59:18] @ Quit: bug: | ||
| [2010/01/07 02:13:54] @ skiold joined channel #puppet | ||
| [2010/01/07 02:14:35] @ Quit: jfreeman: Read error: 60 (Operation timed out) | ||
| [2010/01/07 02:17:56] @ bug joined channel #puppet | ||
| [2010/01/07 02:33:55] @ Quit: Bass10: Read error: 60 (Operation timed out) | ||
| [2010/01/07 02:40:29] @ jfreeman joined channel #puppet | ||
| [2010/01/07 02:40:40] @ ibolmo left channel #puppet () | ||
| [2010/01/07 02:47:19] @ Quit: bug: | ||
| [2010/01/07 02:50:06] @ tuf joined channel #puppet | ||
| [2010/01/07 02:54:32] @ giskard joined channel #puppet | ||
| [2010/01/07 02:58:11] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/07 03:01:11] <ohadlevy> jb: yes? | ||
| [2010/01/07 03:01:45] @ slevin joined channel #puppet | ||
| [2010/01/07 03:06:31] <nasrat> there is some work towards a windows port but it's been stalled due to me having little time | ||
| [2010/01/07 03:07:01] <nasrat> are you using Group Policy? | ||
| [2010/01/07 03:07:14] @ Quit: Boltsky: | ||
| [2010/01/07 03:10:12] @ poison joined channel #puppet | ||
| [2010/01/07 03:11:47] @ Quit: xerxas: "Leaving." | ||
| [2010/01/07 03:12:27] @ Quit: jfreeman: Read error: 110 (Connection timed out) | ||
| [2010/01/07 03:15:19] @ pacalm joined channel #puppet | ||
| [2010/01/07 03:21:14] @ Quit: n0ts__: Client Quit | ||
| [2010/01/07 03:21:28] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:30:30] @ Quit: n0ts: "Tiarra 0.1+svn-35634: SIGTERM received; exit" | ||
| [2010/01/07 03:30:40] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:31:30] @ Quit: n0ts: Client Quit | ||
| [2010/01/07 03:31:31] @ giskard joined channel #puppet | ||
| [2010/01/07 03:31:39] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:31:42] @ Quit: n0ts: Client Quit | ||
| [2010/01/07 03:31:52] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:32:01] @ pacalm_ joined channel #puppet | ||
| [2010/01/07 03:32:05] @ Quit: n0ts: Client Quit | ||
| [2010/01/07 03:32:40] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:36:12] @ xerxas joined channel #puppet | ||
| [2010/01/07 03:42:03] @ Quit: pacalm: Read error: 110 (Connection timed out) | ||
| [2010/01/07 03:42:03] @ pacalm_ is now known as pacalm | ||
| [2010/01/07 03:44:32] @ jes5 joined channel #puppet | ||
| [2010/01/07 03:49:07] @ Quit: n0ts: "Tiarra 0.1+svn-35634: SIGTERM received; exit" | ||
| [2010/01/07 03:49:20] @ xerxas1 joined channel #puppet | ||
| [2010/01/07 03:53:43] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:53:52] @ Quit: n0ts: Client Quit | ||
| [2010/01/07 03:55:41] @ n0ts joined channel #puppet | ||
| [2010/01/07 03:57:59] @ pacalm_ joined channel #puppet | ||
| [2010/01/07 03:58:09] @ Quit: pacalm: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 03:58:09] @ pacalm_ is now known as pacalm | ||
| [2010/01/07 03:58:29] @ Ramonster joined channel #puppet | ||
| [2010/01/07 03:58:45] @ wg joined channel #puppet | ||
| [2010/01/07 03:58:51] @ wg left channel #puppet () | ||
| [2010/01/07 04:00:18] @ Quit: xerxas: Read error: 113 (No route to host) | ||
| [2010/01/07 04:03:49] @ Quit: slevin: | ||
| [2010/01/07 04:06:42] @ tim|macbook joined channel #puppet | ||
| [2010/01/07 04:08:36] @ pacalm_ joined channel #puppet | ||
| [2010/01/07 04:16:18] @ Quit: pacalm: Read error: 110 (Connection timed out) | ||
| [2010/01/07 04:16:18] @ pacalm_ is now known as pacalm | ||
| [2010/01/07 04:18:13] @ MattyM joined channel #puppet | ||
| [2010/01/07 04:20:07] @ Robbie_ joined channel #puppet | ||
| [2010/01/07 04:35:13] @ jab_doa joined channel #puppet | ||
| [2010/01/07 04:36:32] @ Quit: jes5: "Leaving." | ||
| [2010/01/07 04:42:17] @ [newbie] joined channel #puppet | ||
| [2010/01/07 04:50:12] @ Quit: [newbie]: Client Quit | ||
| [2010/01/07 04:50:42] <laomao> Hi... I'm trying to exec a git init in a particular directory; Since cd is a shell built-in I can't specify a full path, so I can't cd into the dir and then run git init; Likewise I seem not to be able to have GIT_DIR='/path/to/project' either. How do I get around this? | ||
| [2010/01/07 04:54:03] <Volcane> exec has an option to set working dir | ||
| [2010/01/07 04:55:42] <laomao> ah ok great | ||
| [2010/01/07 04:55:46] <laomao> thanks | ||
| [2010/01/07 04:56:55] <laomao> BTW who is duritong? I recognized the name on a comment on that article by the person who says puppet sucks so he's going to write his own | ||
| [2010/01/07 04:57:17] <laomao> which was, i thought, a particularlty twatty article | ||
| [2010/01/07 04:57:19] @ Quit: Eghie: Read error: 110 (Connection timed out) | ||
| [2010/01/07 04:57:55] <Volcane> well | ||
| [2010/01/07 04:57:59] @ Eghie joined channel #puppet | ||
| [2010/01/07 04:58:00] <Volcane> lets not forget how puppet happened | ||
| [2010/01/07 04:58:03] <Volcane> and how chef happened | ||
| [2010/01/07 04:58:13] <Volcane> 'god this cfengine stuff is rubbish, I'll write puppet' | ||
| [2010/01/07 04:58:22] <Volcane> 'god this puppet stuff is rubbish, I'll write chef' | ||
| [2010/01/07 05:00:49] @ fluxdude joined channel #puppet | ||
| [2010/01/07 05:01:19] <nasrat> indeed it's a time honoured way of starting a project | ||
| [2010/01/07 05:03:32] @ Quit: xerxas1: "Leaving." | ||
| [2010/01/07 05:03:33] <laomao> hehe ok fair poinr | ||
| [2010/01/07 05:04:11] <laomao> i just thougt the OP was massively missing the point: ok so in one or two providers tehre's some functionality missing on some less popular platforms | ||
| [2010/01/07 05:04:25] <laomao> fine - fix it - that's got to be easier than reinventing the wheel | ||
| [2010/01/07 05:04:29] <Volcane> i do agree though his gripes seems a bit trivial | ||
| [2010/01/07 05:04:54] <Volcane> like, if u have philisophical problems with a DSL, fair point | ||
| [2010/01/07 05:05:00] <Volcane> or if u want tighter data integration etc | ||
| [2010/01/07 05:05:05] <Volcane> those are harder to fix with puppet | ||
| [2010/01/07 05:05:16] <laomao> right - major design issues - fine, do it a differnet way | ||
| [2010/01/07 05:05:18] <Volcane> but his complaints, he's a bit of a whiney twat | ||
| [2010/01/07 05:05:46] <nasrat> anyone got the link to hand | ||
| [2010/01/07 05:05:49] <Volcane> i suspect he's user of $another_langage and dint want to dig into rby | ||
| [2010/01/07 05:06:32] <Volcane> had it yesterday can dig if laomao doesnt ahve it | ||
| [2010/01/07 05:06:34] <laomao> http://blog.steve.org.uk/and_if_someone_gets_upset_you_say___chill_out__.html | ||
| [2010/01/07 05:06:37] @ Quit: tjoe: "Lost terminal" | ||
| [2010/01/07 05:14:20] @ g0nzo joined channel #puppet | ||
| [2010/01/07 05:16:50] @ erlingre joined channel #puppet | ||
| [2010/01/07 05:16:54] @ Quit: erlingre: Remote closed the connection | ||
| [2010/01/07 05:29:03] @ danielbln joined channel #puppet | ||
| [2010/01/07 05:29:49] @ gebi joined channel #puppet | ||
| [2010/01/07 05:32:47] @ Quit: alban2: "Leaving." | ||
| [2010/01/07 05:37:56] @ alban2 joined channel #puppet | ||
| [2010/01/07 05:43:32] @ marcoecc joined channel #puppet | ||
| [2010/01/07 05:44:15] <duritong> laomao: ping | ||
| [2010/01/07 05:44:26] <duritong> ;) | ||
| [2010/01/07 05:50:41] @ Guest20675 is now known as chah | ||
| [2010/01/07 05:50:43] @ Quit: jab_doa: Remote closed the connection | ||
| [2010/01/07 05:53:53] @ Quit: kolla: Remote closed the connection | ||
| [2010/01/07 05:59:57] @ slevin joined channel #puppet | ||
| [2010/01/07 06:02:45] <slevin> hey guys... i have a really simple question and i feel like an idiot but can someone help me? | ||
| [2010/01/07 06:02:45] <slevin> i'm writing a module that 1) copies a file on the node 2) executes some stuff on that 3) deletes that file again. my problem is now that puppet is always complaining about duplicate definitions. i already tried to alias the file but that's not really changing anything. please give me a hint :) | ||
| [2010/01/07 06:07:49] <Volcane> puppet manages resources in a single state | ||
| [2010/01/07 06:07:58] <Volcane> so asking it to make a file then delete it is just not something it does | ||
| [2010/01/07 06:08:25] <slevin> okay so... forget the plan and live with the old files on the system? | ||
| [2010/01/07 06:09:11] <Volcane> well it depends what u are actually trying to do, but conceptually pupet is all about defining the desired state of sometihng | ||
| [2010/01/07 06:09:18] <Volcane> and conceptually there can be only 1 state of something | ||
| [2010/01/07 06:09:29] <Volcane> because remmeber puppet runs every n minutes | ||
| [2010/01/07 06:09:41] <Volcane> so if u do somehting like you describe its possible you'd be making/deleting every 30 minutes | ||
| [2010/01/07 06:09:57] <slevin> yeah i understand the idea behind that | ||
| [2010/01/07 06:10:06] <slevin> okay - thanks a lot! | ||
| [2010/01/07 06:36:03] @ mattwynne joined channel #puppet | ||
| [2010/01/07 06:36:48] <mattwynne> does puppet have the concept of 'environments' so I can test changes to my manifests on a few machines before rolling them out to production? | ||
| [2010/01/07 06:37:09] <Berge> Yes. | ||
| [2010/01/07 06:37:25] <Berge> mattwynne: http://reductivelabs.com/trac/puppet/wiki/UsingMultipleEnvironments | ||
| [2010/01/07 06:37:47] <mattwynne> awesome, thanks Berge | ||
| [2010/01/07 06:37:58] <Berge> No trouble. | ||
| [2010/01/07 06:43:36] @ Quit: alexine_dsouza: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 06:49:12] <mattwynne> so in practice, how am I going to maintain the puppet config files for each environment? we keep the config in git, and it feels like one git branch per environment would work. | ||
| [2010/01/07 06:49:21] <mattwynne> am I going along the right lines? | ||
| [2010/01/07 06:50:09] <nevyn> use environments if your change control process will allow it? | ||
| [2010/01/07 06:50:13] @ Quit: Robbie_: Read error: 60 (Operation timed out) | ||
| [2010/01/07 06:51:07] <Berge> mattwynne: Yes, that would work. | ||
| [2010/01/07 06:51:55] @ Quit: pacalm: | ||
| [2010/01/07 06:55:29] @ Quit: Eghie: Read error: 60 (Operation timed out) | ||
| [2010/01/07 06:56:10] @ Robbie_ joined channel #puppet | ||
| [2010/01/07 06:56:48] <ohadlevy> mattwynne: some people create a repo per module, and then using git submodules to create an environment, so YMMV | ||
| [2010/01/07 06:58:36] <mattwynne> as far as making the changes to puppet.conf, our file seems to have confdir specified in the [main] section, so I can add another environment and specify a different confdir? | ||
| [2010/01/07 06:58:59] <mattwynne> then checkout my 'development' branch to the new confdir specified for my 'development' environment? | ||
| [2010/01/07 07:00:45] <Volcane> u change moduledir etc per environment | ||
| [2010/01/07 07:00:53] <Volcane> and manifest | ||
| [2010/01/07 07:01:09] <Volcane> as in the wiki | ||
| [2010/01/07 07:01:31] <mattwynne> Volcane: I know that's what it says on the wiki - that's why I'm asking. We have an existing setup here that was built by a contractor and I don't want to change it any more than necessary. | ||
| [2010/01/07 07:02:57] <Volcane> best to pastie your config | ||
| [2010/01/07 07:03:24] @ yarihm joined channel #puppet | ||
| [2010/01/07 07:03:43] <FiXion> anyone have a RHEL 5 at their disposal - and can tell me what is says in /etc/redhat-release file? | ||
| [2010/01/07 07:03:53] <mattwynne> Volcane: good idea: http://gist.github.com/271190 | ||
| [2010/01/07 07:05:03] <Volcane> you dont use modules? | ||
| [2010/01/07 07:05:19] @ Eghie joined channel #puppet | ||
| [2010/01/07 07:05:20] <kjetilho> FiXion: which update version? | ||
| [2010/01/07 07:06:14] <kjetilho> Red Hat Enterprise Linux Server release 5.4 (Tikanga) | ||
| [2010/01/07 07:06:22] <mattwynne> Volcane: the folder at confdir contains the a git tree with the following folders in it: manifests, modules, templates, utils | ||
| [2010/01/07 07:06:46] <Volcane> i guess its defaulting to modules then | ||
| [2010/01/07 07:07:11] <Volcane> on your master do: | ||
| [2010/01/07 07:07:12] @ Quit: yarihm: Read error: 60 (Operation timed out) | ||
| [2010/01/07 07:07:30] <Volcane> (as root) puppetmasterd --genconfig|grep modulepath | ||
| [2010/01/07 07:07:45] <mattwynne> it's mostly manifests I want to change anyway -r rightly or wrongly that's where the interesting stuff is in our setup | ||
| [2010/01/07 07:08:27] <Volcane> environments without modules is pretty painful | ||
| [2010/01/07 07:08:46] <mattwynne> modulepath = /var/lib/scalefactory/puppet/modules:/usr/share/puppet/modules | ||
| [2010/01/07 07:08:59] <mattwynne> so it's picking up $confdir/modules | ||
| [2010/01/07 07:09:02] <Volcane> k | ||
| [2010/01/07 07:09:14] <Volcane> so, for environments you'd set that uniquely per environment | ||
| [2010/01/07 07:09:19] <Volcane> and a site.pp per environment | ||
| [2010/01/07 07:09:32] <Volcane> site.pp presumably imports all your manifests? | ||
| [2010/01/07 07:09:41] <FiXion> kjetilho: thanks. | ||
| [2010/01/07 07:09:49] <FiXion> I'm trying to install hpsmh on centos :) | ||
| [2010/01/07 07:09:54] <FiXion> it checks which os it's running on | ||
| [2010/01/07 07:10:27] @ andreid20 joined channel #puppet | ||
| [2010/01/07 07:10:48] <andreid20> hello everyone | ||
| [2010/01/07 07:11:40] @ yarihm joined channel #puppet | ||
| [2010/01/07 07:11:51] <andreid20> I currently use puppet 0.24.8 and want to upgrade to latest 0.25.2 on freebsd and linux gentoo servers ... any big issues with the upgrade process I should know about? | ||
| [2010/01/07 07:12:02] <FiXion> kjetilho: you wouldn't happen to have a puppet module for hpsmh ? ;) | ||
| [2010/01/07 07:12:14] <Volcane> andreid20: you read the release notes for 0.25.0, 1 and 2? | ||
| [2010/01/07 07:12:22] <andreid20> yup | ||
| [2010/01/07 07:13:05] <mattwynne> Volcane: I'm not sure about site.pp - at $confdir/manifests/site.pp there is a file which imports two other .pp files, but I can't see anything in the git tree which is explicity importing the modules | ||
| [2010/01/07 07:13:57] <Volcane> modules dont need importing | ||
| [2010/01/07 07:14:04] <Volcane> but your site.pp is importing classes/pp files etc | ||
| [2010/01/07 07:14:09] <Volcane> those needs to change per environment | ||
| [2010/01/07 07:14:16] <Volcane> so you'll hve a site.pp per environment | ||
| [2010/01/07 07:14:35] <mattwynne> what I think I'm going to do is have a confdir per environment | ||
| [2010/01/07 07:14:47] <mattwynne> because it's the contents of the confdir that's in the git repo | ||
| [2010/01/07 07:15:00] <Volcane> confdir isnt settable per evironment | ||
| [2010/01/07 07:15:09] <mattwynne> balls | ||
| [2010/01/07 07:15:10] <mattwynne> really? | ||
| [2010/01/07 07:15:36] <mattwynne> that's my plan scuppered then | ||
| [2010/01/07 07:15:38] <Volcane> that would imply multiple puppet.confs | ||
| [2010/01/07 07:16:03] <mattwynne> ? even though confdir is specified *in* puppet.conf? | ||
| [2010/01/07 07:16:28] <Volcane> try it | ||
| [2010/01/07 07:21:02] * mattwynne tries it | ||
| [2010/01/07 07:30:55] * robinbowes has upgraded to 0.25.2 | ||
| [2010/01/07 07:31:04] <robinbowes> All seems good, except: | ||
| [2010/01/07 07:31:06] <robinbowes> warning: Value of 'preferred_serialization_format' (pson) is invalid for report, using default (b64_zlib_yaml) | ||
| [2010/01/07 07:31:40] <Volcane> you can ignore that | ||
| [2010/01/07 07:32:01] <robinbowes> Can I silence it ? | ||
| [2010/01/07 07:32:07] <robinbowes> It's a bit noisy | ||
| [2010/01/07 07:32:25] <Volcane> u see it on every run ? | ||
| [2010/01/07 07:32:48] <robinbowes> yes | ||
| [2010/01/07 07:32:56] <Volcane> how do you run puppetd? | ||
| [2010/01/07 07:33:31] <robinbowes> I run it supervised with --no-daemonize | ||
| [2010/01/07 07:33:59] <Volcane> ah i see it too when i set preferred_serialization_format = pson | ||
| [2010/01/07 07:34:02] <Volcane> thats crappy | ||
| [2010/01/07 07:34:11] <Volcane> well u can shut it up with preferred_serialization_format=marshal | ||
| [2010/01/07 07:34:38] <robinbowes> puppet.conf? | ||
| [2010/01/07 07:34:50] <Volcane> yes | ||
| [2010/01/07 07:35:09] <robinbowes> Ah, OK, so the default is pson | ||
| [2010/01/07 07:35:24] <robinbowes> And you're using marhsal so you don't see the warning? | ||
| [2010/01/07 07:36:02] * slevin finished the puppet-update module and is running the update to 0.25.2 right now :) | ||
| [2010/01/07 07:36:16] <Volcane> yeah | ||
| [2010/01/07 07:36:24] <slevin> let puppet update itself :D | ||
| [2010/01/07 07:36:24] <robinbowes> k | ||
| [2010/01/07 07:36:29] <robinbowes> puppet-update module ? | ||
| [2010/01/07 07:36:36] <Volcane> left over from when it didnt have pson just json | ||
| [2010/01/07 07:36:37] * robinbowes has a puppet module | ||
| [2010/01/07 07:36:49] <Volcane> and json required gems to be installed | ||
| [2010/01/07 07:36:49] * FiXion has a puppet module as well | ||
| [2010/01/07 07:36:55] <FiXion> controls versions, config etc. | ||
| [2010/01/07 07:36:57] <slevin> yes i wrote a little module to update puppet | ||
| [2010/01/07 07:37:10] <robinbowes> Er, aren't you missing the point there? | ||
| [2010/01/07 07:37:13] <FiXion> slevin: don't you have a puppet module? to control config etc. | ||
| [2010/01/07 07:37:28] <slevin> of course i have, why? | ||
| [2010/01/07 07:37:43] <FiXion> you don't already enforce version in that module? | ||
| [2010/01/07 07:37:48] <robinbowes> Write a puppet module. Use "ensure => $version". Bump $version when you want to upgrade | ||
| [2010/01/07 07:37:50] <slevin> i install from source | ||
| [2010/01/07 07:38:00] <robinbowes> :rollseyes: | ||
| [2010/01/07 07:38:03] <Volcane> def mising the point :P | ||
| [2010/01/07 07:38:03] <FiXion> so you compile on each host? | ||
| [2010/01/07 07:38:12] <slevin> yes | ||
| [2010/01/07 07:38:30] <robinbowes> A clear example that care in the community is not working. :) | ||
| [2010/01/07 07:38:32] <FiXion> so no one host is the same as the others | ||
| [2010/01/07 07:38:47] <FiXion> yikes | ||
| [2010/01/07 07:39:06] <robinbowes> slevin: what platform? | ||
| [2010/01/07 07:39:06] <FiXion> even when I used Gentoo - I had a seperate build environment and rolled the same out on all hosts | ||
| [2010/01/07 07:39:24] <robinbowes> Even Gentoo does binary builds | ||
| [2010/01/07 07:39:39] <slevin> yes i can either build a rpm or s.th. but where's the problem just installing it on every host | ||
| [2010/01/07 07:39:57] <slevin> the ruby script does not much more than copying some files at the right place | ||
| [2010/01/07 07:40:02] <robinbowes> So, you have to install build tools and dependencies on *all* hosts | ||
| [2010/01/07 07:40:16] <slevin> no | ||
| [2010/01/07 07:40:17] <Volcane> do you also remove files thats not in newer releases anymore? | ||
| [2010/01/07 07:40:24] <slevin> it's just a ruby script | ||
| [2010/01/07 07:40:25] <Volcane> or cater for changes in internal file structuresin puppet? | ||
| [2010/01/07 07:40:48] <robinbowes> packages exist for a good reason | ||
| [2010/01/07 07:41:24] <slevin> erm... do you have a 0.25.2 package for CentOS / RedHat? | ||
| [2010/01/07 07:41:34] <Volcane> there's a rpm spec file in the tarball | ||
| [2010/01/07 07:41:57] <Volcane> rpmbuild -ba -D 'dist .el5' -D 'rhel 5' -D '_without_augeas 1' puppet.spec | ||
| [2010/01/07 07:41:59] <Volcane> done | ||
| [2010/01/07 07:42:04] <slevin> damn | ||
| [2010/01/07 07:42:22] * robinbowes has a 0.25.2 package, yes | ||
| [2010/01/07 07:42:33] <slevin> have to deal more with rpm | ||
| [2010/01/07 07:42:50] <robinbowes> Also, they are usually produced by some Redhat dude shortly after the official release | ||
| [2010/01/07 07:43:06] <Volcane> (and he uses the spec file in the tarball) | ||
| [2010/01/07 07:43:34] <slevin> hmm... can't know anything, eh | ||
| [2010/01/07 07:44:37] <robinbowes> http://tmz.fedorapeople.org/repo/puppet/ | ||
| [2010/01/07 07:44:56] <robinbowes> There's even a yum repo there | ||
| [2010/01/07 07:45:06] <robinbowes> He's not done 0.25.2 yet though | ||
| [2010/01/07 07:45:14] <Volcane> installing from source on redhat machines should be a dismisable offense :P | ||
| [2010/01/07 07:45:25] <robinbowes> Heh, +1 | ||
| [2010/01/07 07:46:22] @ Quit: scylla: "Connection timed out" | ||
| [2010/01/07 07:46:35] <slevin> that was the quickest way for me to get the latest puppet packages on the machines but... now that i know about the spec file it was a bad idea | ||
| [2010/01/07 07:47:21] <robinbowes> Could there be sys admin vacancies in Hamburg shortly, I wonder? ;) | ||
| [2010/01/07 07:47:51] <slevin> ... | ||
| [2010/01/07 07:49:47] @ Quit: choc: Read error: 60 (Operation timed out) | ||
| [2010/01/07 07:51:54] <jamesturnbull> robinbowes: .2 is in Todd's repo now I blieve | ||
| [2010/01/07 07:52:30] <robinbowes> No, just rc3 | ||
| [2010/01/07 07:55:18] @ choc joined channel #puppet | ||
| [2010/01/07 07:58:50] @ GioGio joined channel #puppet | ||
| [2010/01/07 07:59:01] <GioGio> hi there | ||
| [2010/01/07 07:59:45] <GioGio> and happy new year ;) | ||
| [2010/01/07 07:59:57] <GioGio> like a friend say : live long and prosper | ||
| [2010/01/07 08:00:03] <GioGio> :) | ||
| [2010/01/07 08:02:47] <GioGio> is there anyone who know where i can find a rpm for fedora 11 for the latest stable version of puppet please ? | ||
| [2010/01/07 08:03:37] <Volcane> you can make it youself with the included rpm spec file | ||
| [2010/01/07 08:03:50] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/07 08:04:24] <jamesturnbull> robinbowes: oh you are quite right | ||
| [2010/01/07 08:05:08] @ sheldonh joined channel #puppet | ||
| [2010/01/07 08:06:32] <sheldonh> if class my::fridge includes my::milk, and then sets file defaults with File { mode => ... }, are those defaults applied to file resources declared in my::milk? trying to track down an unexpected service refresh :) | ||
| [2010/01/07 08:07:09] <Volcane> it shouldnt | ||
| [2010/01/07 08:07:23] <Volcane> supposed to be in scope like vars as i understood it | ||
| [2010/01/07 08:07:34] <sheldonh> that's what i thought. is the feature actually documented? :) | ||
| [2010/01/07 08:07:54] <Volcane> yes, bt not sure the scoping rules is clearly doc'd | ||
| [2010/01/07 08:08:05] <Volcane> is in wiki:LanguageTutorial | ||
| [2010/01/07 08:08:07] <gepetto> Volcane: wiki:LanguageTutorial is http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial | ||
| [2010/01/07 08:09:13] <sheldonh> i get "//php5/php5::per_user_config/apache::service/File[/etc/init.d/apache]) Scheduling refresh of Exec[generate-user-php-config], which makes me think generate-user-php-config is subscribed to /etc/init.d/apache, or the latter notifies the former. but i've set up no such association | ||
| [2010/01/07 08:09:29] <sheldonh> unless File { ... } doesn't work like i thought it did :) | ||
| [2010/01/07 08:11:04] <Volcane> i guess if foo sets it, and then includes bar then just like a variable bar would be affected | ||
| [2010/01/07 08:11:16] <ricky> robinbowes, jamesturnbull: There are 0.25.2 builds at http://koji.fedoraproject.org/koji/packageinfo?packageID=3529 for what it's worth | ||
| [2010/01/07 08:13:05] @ bjaspan joined channel #puppet | ||
| [2010/01/07 08:16:52] @ Quit: Eghie: Read error: 110 (Connection timed out) | ||
| [2010/01/07 08:17:36] @ Eghie joined channel #puppet | ||
| [2010/01/07 08:18:35] <sheldonh> Volcane, that would kinda make sense. but i include bar, then set defaults :) | ||
| [2010/01/07 08:19:21] <Volcane> yeah that shouldnt cause it i wouldnt think - but who know :) | ||
| [2010/01/07 08:19:30] <Volcane> make a little test.pp and test | ||
| [2010/01/07 08:20:17] @ danielbln1 joined channel #puppet | ||
| [2010/01/07 08:20:17] @ Quit: danielbln: Nick collision from services. | ||
| [2010/01/07 08:20:22] @ danielbln1 is now known as danielbln | ||
| [2010/01/07 08:20:56] @ Quit: rmiller4pi81: Read error: 110 (Connection timed out) | ||
| [2010/01/07 08:21:48] @ Bass10 joined channel #puppet | ||
| [2010/01/07 08:24:34] @ pacalm joined channel #puppet | ||
| [2010/01/07 08:26:05] @ Quit: bjaspan: | ||
| [2010/01/07 08:27:11] @ lilmatt joined channel #puppet | ||
| [2010/01/07 08:27:51] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 08:28:58] @ brothers joined channel #puppet | ||
| [2010/01/07 08:29:25] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 08:29:29] @ brothers joined channel #puppet | ||
| [2010/01/07 08:29:41] @ Quit: vzctl_: Read error: 110 (Connection timed out) | ||
| [2010/01/07 08:33:19] @ Quit: mattwynne: "Leaving..." | ||
| [2010/01/07 08:35:12] @ pacalm_ joined channel #puppet | ||
| [2010/01/07 08:37:06] * ohadlevy finally created some graph alternatives to the rrd's | ||
| [2010/01/07 08:38:35] @ kolla joined channel #puppet | ||
| [2010/01/07 08:39:45] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 08:39:50] @ brothers joined channel #puppet | ||
| [2010/01/07 08:42:17] @ Quit: pacalm: Read error: 110 (Connection timed out) | ||
| [2010/01/07 08:42:18] @ pacalm_ is now known as pacalm | ||
| [2010/01/07 08:42:27] <robinbowes> Hey ohadlevy | ||
| [2010/01/07 08:42:38] <ohadlevy> robinbowes: hey | ||
| [2010/01/07 08:42:47] <robinbowes> I was looking for you earlier this week | ||
| [2010/01/07 08:42:51] <robinbowes> Can't remember what for though. | ||
| [2010/01/07 08:42:58] <ohadlevy> :) | ||
| [2010/01/07 08:43:03] <robinbowes> Can't have been important! | ||
| [2010/01/07 08:44:48] @ lero joined channel #puppet | ||
| [2010/01/07 08:45:12] <ohadlevy> robinbowes: probably foreman related? | ||
| [2010/01/07 08:45:17] <robinbowes> Probably | ||
| [2010/01/07 08:45:29] <robinbowes> I probably had a question about an upgrade | ||
| [2010/01/07 08:46:01] <robinbowes> It might have been something like... is the import process idempotent? | ||
| [2010/01/07 08:46:21] <ohadlevy> robinbowes: yep | ||
| [2010/01/07 08:46:37] @ bodepd joined channel #puppet | ||
| [2010/01/07 08:48:22] @ jab_doa joined channel #puppet | ||
| [2010/01/07 08:49:01] @ cwebber joined channel #puppet | ||
| [2010/01/07 08:49:23] <robinbowes> I think I found that out myself through trial and error | ||
| [2010/01/07 08:49:37] <robinbowes> "empirically" as we used to say... | ||
| [2010/01/07 08:49:54] @ vzctl_ joined channel #puppet | ||
| [2010/01/07 08:53:09] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 08:53:26] * ohadlevy wonders which other graphs can be useful... | ||
| [2010/01/07 08:54:02] @ brothers joined channel #puppet | ||
| [2010/01/07 08:59:02] @ bjaspan joined channel #puppet | ||
| [2010/01/07 08:59:56] @ fzzzt joined channel #puppet | ||
| [2010/01/07 09:04:29] <sheldonh> Volcane, test results aren't looking good for POLA :) | ||
| [2010/01/07 09:04:45] <FiXion> slevin: ohh - and I recommend having your own package mirror | ||
| [2010/01/07 09:05:09] <FiXion> when a new version is released - they remove the old versions - and with puppet especially - you want to run the same version on all clients :) | ||
| [2010/01/07 09:05:22] <FiXion> which is kinda hard if they removed it and you haven't upgraded all to the new version yet :) | ||
| [2010/01/07 09:06:54] <Volcane> sheldonh: file bugs :) | ||
| [2010/01/07 09:07:37] <sheldonh> Volcane, will do, once i've proven it exists in 0.25. i'm pretty sure nobody cares about 0.24 any more :) | ||
| [2010/01/07 09:08:21] <joe-mac> i still use .24.8 | ||
| [2010/01/07 09:08:25] <joe-mac> it's rock solid | ||
| [2010/01/07 09:08:31] <joe-mac> for me at least | ||
| [2010/01/07 09:09:40] <agaffney> same here | ||
| [2010/01/07 09:09:51] <agaffney> I might switch to 0.25 once it's in EPEL | ||
| [2010/01/07 09:10:00] <agaffney> and proven to work on EL4 ;) | ||
| [2010/01/07 09:10:11] <agaffney> I'm mostly EL5, but I've got a couple of older EL4 boxes around | ||
| [2010/01/07 09:10:33] @ jmarki joined channel #puppet | ||
| [2010/01/07 09:10:36] <joe-mac> i use ooboonto 8.04 and OpenBSD | ||
| [2010/01/07 09:10:53] <joe-mac> the portion that is configured by puppet on openbsd is trivial though | ||
| [2010/01/07 09:16:12] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 09:17:17] @ brothers joined channel #puppet | ||
| [2010/01/07 09:18:56] @ themurph joined channel #puppet | ||
| [2010/01/07 09:20:11] @ docelic joined channel #puppet | ||
| [2010/01/07 09:22:20] @ sektoid joined channel #puppet | ||
| [2010/01/07 09:24:50] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 09:25:18] @ brothers joined channel #puppet | ||
| [2010/01/07 09:25:40] <joe-mac> what's the best way to downcase a fact? | ||
| [2010/01/07 09:25:49] <joe-mac> is it even possible? | ||
| [2010/01/07 09:26:25] @ Quit: flakrat: "Leaving" | ||
| [2010/01/07 09:27:27] @ kaptk2 joined channel #puppet | ||
| [2010/01/07 09:28:50] <fzzzt> I have 0.25 on RHEL4 and it works fine | ||
| [2010/01/07 09:29:57] <fluxdude> joe-mac: do you really need to downcase it? | ||
| [2010/01/07 09:30:00] <fzzzt> How can I have multiple definitions get put into the same file? | ||
| [2010/01/07 09:30:18] <fluxdude> when doing comparisons, I think the comparisons in say case statements are case insensitive... | ||
| [2010/01/07 09:30:41] <fluxdude> fzzzt: clarify? | ||
| [2010/01/07 09:30:52] <robinbowes> joe-mac: you could probably do it using inline template | ||
| [2010/01/07 09:30:59] <fluxdude> you mean to specify multiple definitions in a single manifest file...? yes I think you can do that | ||
| [2010/01/07 09:31:38] <fzzzt> like, thing { parameter=>'hi' } thing { parameter=>'foo' } -> /etc/things.txt which contains the data, like 1: hi, 2: foo (though numerical order doesn't matter) | ||
| [2010/01/07 09:31:41] <joe-mac> i want to realize Nagios_hostgroup["${kernel}-servers"] | ||
| [2010/01/07 09:31:46] <joe-mac> no big deal | ||
| [2010/01/07 09:31:49] <fluxdude> robinbowes: ah yes, ruby inside template... or you can make a custom function and drop that in which can give you a function inside puppet code to do it | ||
| [2010/01/07 09:31:50] <joe-mac> worked around it | ||
| [2010/01/07 09:32:03] <joe-mac> yea i was thinking about inline_template but it seemed ugly | ||
| [2010/01/07 09:32:29] <ashp> idle question prompted by an argument at work: do you tend to split modules into 7-8 .pp's or do you tend to just shovel it all into one big init.pp? (this is for everyone) | ||
| [2010/01/07 09:32:34] <fluxdude> I did the custom function thing to extend puppet recently and it's really not that difficult | ||
| [2010/01/07 09:32:43] <fzzzt> basically take the parameters in multiple definitions and output them into one file | ||
| [2010/01/07 09:32:45] <ashp> my coworker was asking about when to split them and the best I could suggest was 'when it's big' | ||
| [2010/01/07 09:32:49] <fluxdude> ashp: I like 1 init.pp unless it's really big | ||
| [2010/01/07 09:32:58] * robinbowes uses separate .pps | ||
| [2010/01/07 09:33:13] <robinbowes> Generally, init.pp includes all the others | ||
| [2010/01/07 09:33:15] <fzzzt> ashp: I use one file until it's more than a few pages | ||
| [2010/01/07 09:33:18] <kjetilho> ashp: I like 1 init.pp unless the class structure needs to be exposed ;-) | ||
| [2010/01/07 09:33:28] <kjetilho> and it often needs to be | ||
| [2010/01/07 09:33:35] <fluxdude> I only use separate pps when they are really different things, but if it's one or two classes... then I'd just stick it in init.pp | ||
| [2010/01/07 09:34:00] <kjetilho> ie. for direct access from the outside (either include or inherit) | ||
| [2010/01/07 09:34:16] <joe-mac> that's largely a matter of taste imo | ||
| [2010/01/07 09:34:17] <ashp> I used to split them into loads of files but over time I've come to just use one giant init.pp unless I require some defines elsewhere etc | ||
| [2010/01/07 09:34:26] <ashp> and my coworker was asking which were right :) | ||
| [2010/01/07 09:34:33] <joe-mac> i put definitions in init.pp and stuff all sub classes should have, then put subclasses in files like $subclass_name.pp | ||
| [2010/01/07 09:34:56] <robinbowes> I use dirs for sub-classes | ||
| [2010/01/07 09:34:59] <ashp> I think for my defines for things like ldap::servertype I'm going to add include ldap::$name at the top of the define | ||
| [2010/01/07 09:35:11] <ashp> and then use seperate .pp's for each $name that I might have to capture specific logic for only that type | ||
| [2010/01/07 09:35:32] @ Quit: rmiller4pi8: "Leaving." | ||
| [2010/01/07 09:35:32] <ashp> as i have about 6 'types' of ldap server and it's turned into a nest of if $name {} stuff in the main define | ||
| [2010/01/07 09:35:40] @ flakrat joined channel #puppet | ||
| [2010/01/07 09:35:49] <ashp> then i'm going to consolidate all my older modules into one giant init.pp so it's not spread all over | ||
| [2010/01/07 09:36:01] <fzzzt> I do that with my titles | ||
| [2010/01/07 09:36:03] <ashp> well, one init.pp per module I mean, not ONE giant one :) | ||
| [2010/01/07 09:36:17] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/07 09:36:29] @ Quit: docelic: "http://www.spinlocksolutions.com/" | ||
| [2010/01/07 09:37:31] @ Quit: jcape: Read error: 60 (Operation timed out) | ||
| [2010/01/07 09:38:19] <robinbowes> ashp: I do something similar | ||
| [2010/01/07 09:38:28] <tmz> jamesturnbull, robinbowes: FWIW, 0.25.2 packages are being pushed to Fedora and EPEL updates-testing repos. It's in the Fedora repo now, but not yet in the EPEL repo. | ||
| [2010/01/07 09:38:40] <robinbowes> OK, cool | ||
| [2010/01/07 09:38:43] <tmz> For EL-5, here's the build: http://koji.fedoraproject.org/koji/buildinfo?buildID=149576 | ||
| [2010/01/07 09:38:52] <robinbowes> It will be 0.25.2-1 ? | ||
| [2010/01/07 09:39:30] <tmz> Yes (well, 1.1, due to a small issue building in the rawhide branch) | ||
| [2010/01/07 09:39:59] <ashp> robinbowes: and it works alright? I've been struggling with the best way to handle things like apache/ldap as we just end up with so many various 'types' which are basically the same old with a few extra bits tacked on, it's messy and annoying | ||
| [2010/01/07 09:40:07] <tmz> There are EL-4 packages as well, but they appear to be hit by #3013 | ||
| [2010/01/07 09:40:08] <gepetto> tmz: #3013 is http://projects.reductivelabs.com/issues/show/3013 "Puppet - Bug #3013: util.rb:execute broken on Ruby <1.8.3 - ReductiveLabs.com" | ||
| [2010/01/07 09:41:23] <robinbowes> ashp: I have a common base class, and extend it | ||
| [2010/01/07 09:41:44] <robinbowes> And have a define that creates the right class | ||
| [2010/01/07 09:42:07] <ashp> robinbowes: Hmm, can you show me an example of the define/common class so I acn see how you mean? | ||
| [2010/01/07 09:42:49] <robinbowes> eg. mysqld::instance{ type => 'foo', role => 'master', ... } | ||
| [2010/01/07 09:44:55] @ jcape joined channel #puppet | ||
| [2010/01/07 09:48:33] @ notbrien joined channel #puppet | ||
| [2010/01/07 09:49:04] @ Quit: teratoma: "leaving" | ||
| [2010/01/07 09:51:34] <sheldonh> our experience of 0.24.8 hasn't been great. we're in love with puppet, but this version feels flakey. once a day, we pull a report on long lost hosts and have to log into them to restart wedged puppetds and start up dead puppetds | ||
| [2010/01/07 09:51:56] <sheldonh> small price to pay for what we have, but we're hopeful that 0.25 is more solid | ||
| [2010/01/07 09:52:06] <monachus> sheldonh: are you sure that's puppet's issue and not something in your config that's wedging them? | ||
| [2010/01/07 09:52:30] <kjetilho> well, Puppet should be self-healing :/ | ||
| [2010/01/07 09:52:37] <monachus> we used to have similar issues - we put puppet under daemontools now to keep it from dying | ||
| [2010/01/07 09:52:49] <sheldonh> monachus, not sure, no. but from the straces and log excerpts we've managed to glean, it's looked like the same thing that happens if you kill -9 the puppetmaster processes at just the wrong time in the conversation :) | ||
| [2010/01/07 09:53:24] <sheldonh> monachus, if 0.25 doesn't nail it for us, we'll go with something like that (although never daemontools itself) | ||
| [2010/01/07 09:53:39] <monachus> wedging was often caused by LDAP issues - not even by puppet, but more system-level, like when the underlying pam_ldap/nss_ldap system loses the LDAP server and then reconnects, puppet never returns from the dead. it doesn't happen that often anymore | ||
| [2010/01/07 09:53:48] <fzzzt> I have a script that checks for it, and checks for a file to see if it should be disabled or not (e.g. testing) | ||
| [2010/01/07 09:53:48] <sheldonh> monachus, no ldap here | ||
| [2010/01/07 09:54:01] <fzzzt> It also checks RAM/CPU usage because we've had it run amok. | ||
| [2010/01/07 09:54:14] <monachus> we have nagios monitoring the state.yaml file, and if it's > 60m old, it bounces puppet for us | ||
| [2010/01/07 09:54:17] <sheldonh> monachus, could be a problem between the puppetmasters and the database they store logs and reports to | ||
| [2010/01/07 09:54:37] @ Cyis joined channel #puppet | ||
| [2010/01/07 09:54:39] <sheldonh> monachus, we're a way off from installing monit :) | ||
| [2010/01/07 09:54:54] <fzzzt> I'm working on a script that scans the master's log and reports on hosts that haven't checked in lately | ||
| [2010/01/07 09:54:56] <Volcane> sheldonh: ditch the daemon and run from cron | ||
| [2010/01/07 09:55:22] <sheldonh> Volcane, a) then we can't use puppetrun, and b) yes, but that only solves death, not hanging | ||
| [2010/01/07 09:55:45] <Volcane> sheldonh: ah, well there's a thin little puppetrun replacement in ext | ||
| [2010/01/07 09:55:45] <sheldonh> Volcane, i know, i know, mcollective > puppetrun ;) | ||
| [2010/01/07 09:55:45] <robinbowes> I must look at puppetrun | ||
| [2010/01/07 09:56:02] <Volcane> sheldonh: for cases where u run from cron | ||
| [2010/01/07 09:56:03] @ Quit: andreid20: Remote closed the connection | ||
| [2010/01/07 09:56:13] <Cyis> Volcane, hey you ever tried to setup a define for a config file and want to be able to have it either use a default file (source => ...) or alternatively pass it a template? | ||
| [2010/01/07 09:56:22] <Volcane> Cyis: yes | ||
| [2010/01/07 09:56:46] <sheldonh> Volcane, and then just have monit kill too-long-lived puppetds. interesting idea | ||
| [2010/01/07 09:57:01] <Volcane> sheldonh: not had them hang much i have to say | ||
| [2010/01/07 09:57:20] <Cyis> I figured I could use 'content' to send to the define and simply set it to template("...") for the second case but didn't think that'd work with a source file | ||
| [2010/01/07 09:57:39] @ bobbyz joined channel #puppet | ||
| [2010/01/07 09:57:42] @ Quit: jmarki: Remote closed the connection | ||
| [2010/01/07 09:57:42] <sheldonh> Volcane, it's a very low proportion. fewer than 7 every day. with over 1,200 hosts :) | ||
| [2010/01/07 09:58:16] <sheldonh> Volcane, actually, probably fewer than 1 a day. the 2-6 a day we get are mostly dead, not hung | ||
| [2010/01/07 09:58:39] <sheldonh> Volcane, the problem with rare problems is that they become common at scale. but then i'm preaching to the choir in #puppet lol | ||
| [2010/01/07 09:59:19] <Volcane> sheldonh: heh yeah, i found weird locked and odd ones with connectivity issues but otherwise no hassles, crashes yes and unhandled exceptions - cron solves those | ||
| [2010/01/07 09:59:36] @ sdodson left channel #puppet () | ||
| [2010/01/07 09:59:58] <Cyis> Volcane, suggestions/example/link you can throw my way? | ||
| [2010/01/07 10:00:21] @ jmarki joined channel #puppet | ||
| [2010/01/07 10:00:47] <Volcane> Cyis: i have a source and content options on my define, they default to something. then inside the define i check which one to use | ||
| [2010/01/07 10:01:36] <Cyis> Volcane, oh... so have something like... define blah ($content = '', $source = '') { .. } then case both ? | ||
| [2010/01/07 10:02:08] <Volcane> yes, and use the 'setting defaults' feature to affect a file resource in the define | ||
| [2010/01/07 10:02:48] @ rmiller4pi81 joined channel #puppet | ||
| [2010/01/07 10:03:18] @ mattwynne joined channel #puppet | ||
| [2010/01/07 10:03:57] @ loupgaroublond joined channel #puppet | ||
| [2010/01/07 10:04:00] <loupgaroublond> hi all | ||
| [2010/01/07 10:04:28] <loupgaroublond> where can i find a complete bnf of puppet's language? | ||
| [2010/01/07 10:04:50] <Volcane> i mean its all u need to do in the patch | ||
| [2010/01/07 10:04:54] <Volcane> oops, wrong # | ||
| [2010/01/07 10:07:02] <joe-mac> what does bnf stand for? there is the languagedocument | ||
| [2010/01/07 10:07:16] <joe-mac> wiki:LanguageDocument or is it wiki:LanguageReference | ||
| [2010/01/07 10:07:49] <loupgaroublond> backus naur format, it's the grammer that explains the language | ||
| [2010/01/07 10:08:09] @ rasputnik joined channel #puppet | ||
| [2010/01/07 10:08:11] <loupgaroublond> there's the language tutorial, it has a partial BNF | ||
| [2010/01/07 10:09:43] @ xerxas joined channel #puppet | ||
| [2010/01/07 10:09:51] <jamesturnbull> loupgaroublond: that's the only BNF currently | ||
| [2010/01/07 10:12:20] @ Quit: jmarki: Remote closed the connection | ||
| [2010/01/07 10:14:37] <joe-mac> interesting, will read about it | ||
| [2010/01/07 10:14:48] <joe-mac> i'm not a developer so, i guesst hat explains why i had no clue what it is | ||
| [2010/01/07 10:15:43] <joe-mac> o i've seen this before actuially, didn't recall the name | ||
| [2010/01/07 10:15:48] <joe-mac> or didn't recall the concept from the name | ||
| [2010/01/07 10:16:07] @ jcape1 joined channel #puppet | ||
| [2010/01/07 10:16:19] <loupgaroublond> ah, could be that too | ||
| [2010/01/07 10:16:55] <loupgaroublond> anywho, i'm having some issues with a variable not being set at the appropriate time, it begins with an underscore, i was wondering if it needs to be quoted with braces or something | ||
| [2010/01/07 10:17:06] <jamesturnbull> loupgaroublond: masterzen wrote that BNF if you have questions | ||
| [2010/01/07 10:17:59] @ Quit: rmiller4pi8: Read error: 110 (Connection timed out) | ||
| [2010/01/07 10:18:26] @ Quit: bpgoldsb: Read error: 60 (Operation timed out) | ||
| [2010/01/07 10:18:33] <masterzen> loupgaroublond: check the source (lib/puppet/parser/grammar.ra), there is a racc grammar (which is more or less a kind of BNF), it describes the whole puppet language. | ||
| [2010/01/07 10:18:33] @ jmarki joined channel #puppet | ||
| [2010/01/07 10:18:48] <Cyis> I hate when boring work tasks get in the way of working on my puppet configs :) | ||
| [2010/01/07 10:19:13] <loupgaroublond> masterzen, thanks :) | ||
| [2010/01/07 10:20:02] <joe-mac> same here Cyis which is why i sometimes take a stand against work | ||
| [2010/01/07 10:20:10] <joe-mac> and just do puppet (like i am doing today) | ||
| [2010/01/07 10:20:26] <joe-mac> heh... jk what i am doing now is for a nagios unification project | ||
| [2010/01/07 10:20:39] <joe-mac> really, it's part of infrastructure unification | ||
| [2010/01/07 10:20:56] <Cyis> I'm basically working to get our existing infrastructure under puppet control | ||
| [2010/01/07 10:21:01] @ Telmo joined channel #puppet | ||
| [2010/01/07 10:21:06] <Cyis> approaching it one part at a time | ||
| [2010/01/07 10:21:15] <joe-mac> yea i was doing that for months, here, now i have to do it for the west coast | ||
| [2010/01/07 10:24:46] @ giskard_ joined channel #puppet | ||
| [2010/01/07 10:25:20] @ bug joined channel #puppet | ||
| [2010/01/07 10:26:30] <joe-mac> rofl, i just got a dependency cycle that seems to go on for half my screen | ||
| [2010/01/07 10:26:33] @ Quit: giskard: Nick collision from services. | ||
| [2010/01/07 10:26:36] @ giskard_ is now known as giskard | ||
| [2010/01/07 10:28:00] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 10:29:33] @ andreid20 joined channel #puppet | ||
| [2010/01/07 10:29:53] <andreid20> anybody know how to specify a package name for gentoo in the new puuppet 25.2 | ||
| [2010/01/07 10:29:55] <andreid20> ? | ||
| [2010/01/07 10:30:24] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/07 10:30:42] <andreid20> I keep getting this error: Failed to retrieve current state of resource: No package found with the specified name [<unspecified category>/net-snmp] | ||
| [2010/01/07 10:31:04] <andreid20> and it worked with 24.8 | ||
| [2010/01/07 10:32:25] <masterzen> joe-mac: I think nigelk wrote a dependency-cycle demangler if you need it, maybe you can ping him... | ||
| [2010/01/07 10:33:09] <joe-mac> masterzen: i think i see it right now, about to make a commit that should greatly simplify it or even fix it | ||
| [2010/01/07 10:33:22] <joe-mac> getting nagios to work automatically based on puppet is kind of difficult | ||
| [2010/01/07 10:33:38] <masterzen> joe-mac: that's why I'm not using nagios :-) | ||
| [2010/01/07 10:34:02] <joe-mac> i've invested way too much intellectual capital in nagios to not use it :-P | ||
| [2010/01/07 10:34:08] <masterzen> joe-mac: lol | ||
| [2010/01/07 10:34:12] <joe-mac> it's probably taking up something like half my brainspace right now | ||
| [2010/01/07 10:34:20] <joe-mac> i'll forget how to breathe soon, but be able to write nagios definitions | ||
| [2010/01/07 10:34:22] @ rasputnik left channel #puppet ("Leaving") | ||
| [2010/01/07 10:34:40] @ Quit: bjaspan: | ||
| [2010/01/07 10:34:50] @ Quit: crdant: "Leaving." | ||
| [2010/01/07 10:34:52] <masterzen> joe-mac: do not forget to perform a brain backup before reformatting this part... in case you need to add some nagios stuff in a couple of month :-D | ||
| [2010/01/07 10:34:54] @ gnosek joined channel #puppet | ||
| [2010/01/07 10:36:17] <gnosek> hi all, I need a cluebat. one server (just one) cannot see its assigned classes ("no classes to store") | ||
| [2010/01/07 10:37:25] <gnosek> even though my node classifier tool (cobbler-ext-nodes) says everything's fine and the right class is seen in .../yaml/nodes/<fqdn>.yaml | ||
| [2010/01/07 10:39:56] @ Quit: bodepd: | ||
| [2010/01/07 10:41:22] @ gaveen joined channel #puppet | ||
| [2010/01/07 10:45:33] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 10:45:58] @ brothers joined channel #puppet | ||
| [2010/01/07 10:46:29] @ bjaspan joined channel #puppet | ||
| [2010/01/07 10:46:53] @ Quit: jcape1: Read error: 60 (Operation timed out) | ||
| [2010/01/07 10:47:59] @ Quit: rmiller4pi81: Read error: 110 (Connection timed out) | ||
| [2010/01/07 10:50:34] <Cyis> http://pastie.org/770215 that look sane? | ||
| [2010/01/07 10:51:42] <joe-mac> no, what's this File["config-${name}"] { content => $content } | ||
| [2010/01/07 10:51:52] <joe-mac> looks liek you are overriding, but this isn't a class | ||
| [2010/01/07 10:52:22] <Cyis> joe: I just edited... noticed that | ||
| [2010/01/07 10:52:59] <Cyis> yeah I'm trying to make it conditional for source/content | ||
| [2010/01/07 10:53:13] <joe-mac> not worth the work imo | ||
| [2010/01/07 10:53:47] <Cyis> got a better way to go about it? | ||
| [2010/01/07 10:56:09] <Cyis> I guess I could do the entire file resource within conditionals but then I'm pretty much repeating everything which isn't very DRY | ||
| [2010/01/07 10:57:16] <joe-mac> you can reduce verbosity, if that's your goal, by setting defaults in site.pp | ||
| [2010/01/07 10:58:05] <Cyis> this is inside a module I'm working on | ||
| [2010/01/07 10:58:57] @ Quit: tim|macbook: | ||
| [2010/01/07 10:59:08] <cwebber> if i change dnscertnames for the puppetmaster will I have to rebuild all of the cert databases on the clients | ||
| [2010/01/07 10:59:42] @ Quit: GioGio: "Page closed" | ||
| [2010/01/07 11:00:59] @ Quit: slevin: | ||
| [2010/01/07 11:01:21] @ littleidea joined channel #puppet | ||
| [2010/01/07 11:01:50] @ jcape joined channel #puppet | ||
| [2010/01/07 11:04:56] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/07 11:05:14] @ qwebirc8227 joined channel #puppet | ||
| [2010/01/07 11:10:03] @ labrown joined channel #puppet | ||
| [2010/01/07 11:13:51] @ Quit: gnosek: "bye" | ||
| [2010/01/07 11:17:57] @ mikerowehl joined channel #puppet | ||
| [2010/01/07 11:28:50] @ crdant joined channel #puppet | ||
| [2010/01/07 11:29:52] @ Quit: gebi: Read error: 60 (Operation timed out) | ||
| [2010/01/07 11:32:26] @ Quit: sheldonh: "Ex-Chat" | ||
| [2010/01/07 11:33:08] <lilmatt> can I set global variables in nodes.pp, for example "server role", that I can then access in recipes? | ||
| [2010/01/07 11:33:29] @ windowsrefund joined channel #puppet | ||
| [2010/01/07 11:33:31] <windowsrefund> hello | ||
| [2010/01/07 11:34:20] <windowsrefund> I'm getting a "403 Forbidden" when attempting to run puppetd against my master that has a passenger backend | ||
| [2010/01/07 11:34:26] <lilmatt> I want to set this, similar to a fact, but I was trying to avoid having to create a new facter script for each server type | ||
| [2010/01/07 11:34:36] <windowsrefund> does that have anything to do with a auth.conf or something? | ||
| [2010/01/07 11:34:56] <duritong> lilmatt: yes you can | ||
| [2010/01/07 11:35:04] <duritong> windowsrefund: might be | ||
| [2010/01/07 11:35:14] <lilmatt> windowsrefund: does the passenger user have filesystem permission to access the puppet files? | ||
| [2010/01/07 11:35:30] @ rmiller4pi81 joined channel #puppet | ||
| [2010/01/07 11:35:56] <lilmatt> duritong: is there a special syntax or anything? I'm setting $server_role in each node in nodes.pp, but it's empty when I go to access it in a module recipe | ||
| [2010/01/07 11:36:28] <windowsrefund> lilmatt: the passenger user is the user that owns config.ru, right? | ||
| [2010/01/07 11:36:29] @ rmiller4pi82 joined channel #puppet | ||
| [2010/01/07 11:36:43] <lilmatt> It's the user passenger is running as. | ||
| [2010/01/07 11:36:45] <z00dax> any views on how much data a variable can hold ? | ||
| [2010/01/07 11:36:51] <windowsrefund> ok | ||
| [2010/01/07 11:36:53] <duritong> lilmatt: are you using node inheritnace? | ||
| [2010/01/07 11:36:53] * windowsrefund checks | ||
| [2010/01/07 11:36:59] <lilmatt> duritong: yes | ||
| [2010/01/07 11:37:06] <duritong> lilmatt: then this is your problem | ||
| [2010/01/07 11:37:31] <duritong> you can't define a variable in a sub-node which should be present to a recipe which is included in a super node | ||
| [2010/01/07 11:38:08] <duritong> the solution is to either switch to an external node tool, which can handle that much better | ||
| [2010/01/07 11:38:22] <duritong> or you have to structure your node-hierarchy different | ||
| [2010/01/07 11:38:33] <duritong> I use the following way which works quite well for me: | ||
| [2010/01/07 11:38:33] <lilmatt> external node tool? | ||
| [2010/01/07 11:38:59] <duritong> lilmatt: wiki:ExternalNodes | ||
| [2010/01/07 11:38:59] <gepetto> duritong: lilmatt: wiki:ExternalNodes is http://reductivelabs.com/trac/puppet/wiki/ExternalNodes | ||
| [2010/01/07 11:39:04] <windowsrefund> lilmatt: I can help with that, I have variables defined for my nodes that I can use in manifests and I use inheritance | ||
| [2010/01/07 11:39:28] @ Quit: littleidea: | ||
| [2010/01/07 11:39:50] <windowsrefund> lilmatt: passenger is running as the puppet user and I just did a chown -R on my manifest directory... same 403 | ||
| [2010/01/07 11:39:58] @ rmiller4pi83 joined channel #puppet | ||
| [2010/01/07 11:40:06] <duritong> lilmatt: the best thing is to factor class hierarchies out of node inheritance into a class hierarchy and use node hierarchy only for variables | ||
| [2010/01/07 11:40:46] <lilmatt> duritong: that confused me | ||
| [2010/01/07 11:41:04] <lilmatt> so determine my classes based on variables set in node inheritance? | ||
| [2010/01/07 11:41:10] <duritong> nope | ||
| [2010/01/07 11:41:20] <duritong> wait I'm looking for the thread I wrote once | ||
| [2010/01/07 11:42:59] @ Quit: Ramonster: Client Quit | ||
| [2010/01/07 11:43:12] @ Quit: mqr: | ||
| [2010/01/07 11:43:40] <duritong> lilmatt: you find some answers here: http://groups.google.com/group/puppet-users/browse_thread/thread/49891a1cfb99f1c1/d373a3fc372f27e5?lnk=gst&q=config+node+inheritance+class+peter+meier#d373a3fc372f27e5 | ||
| [2010/01/07 11:44:03] <duritong> but I have once described my setup which works quite well | ||
| [2010/01/07 11:44:09] <duritong> but can't find the thread anymore | ||
| [2010/01/07 11:44:18] <duritong> but windowsrefund might also have a valuable solution | ||
| [2010/01/07 11:44:29] @ alfism joined channel #puppet | ||
| [2010/01/07 11:47:30] @ Quit: mattwynne: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 11:47:51] <windowsrefund> lilmatt: http://pastebin.com/m38a3199 | ||
| [2010/01/07 11:47:55] <lilmatt> tnx | ||
| [2010/01/07 11:47:58] <windowsrefund> that's how I do it | ||
| [2010/01/07 11:48:10] @ mattwynne joined channel #puppet | ||
| [2010/01/07 11:48:16] <duritong> windowsrefund: that's similar to mine | ||
| [2010/01/07 11:48:19] <duritong> :) | ||
| [2010/01/07 11:48:22] <windowsrefund> cool | ||
| [2010/01/07 11:49:05] <windowsrefund> I really still don't understand it at all | ||
| [2010/01/07 11:49:35] <windowsrefund> anyway, back to this 403 | ||
| [2010/01/07 11:50:32] <duritong> windowsrefund: you don't understand the node things at all? | ||
| [2010/01/07 11:50:37] <duritong> or passenger? | ||
| [2010/01/07 11:51:23] @ jcape1 joined channel #puppet | ||
| [2010/01/07 11:51:37] <windowsrefund> the node thing | ||
| [2010/01/07 11:51:40] <windowsrefund> or passenger | ||
| [2010/01/07 11:51:41] <windowsrefund> hahaha | ||
| [2010/01/07 11:51:47] <duritong> ;) | ||
| [2010/01/07 11:51:50] <windowsrefund> but nodes work | ||
| [2010/01/07 11:52:07] <windowsrefund> and I suspect this inheritance stuff will change anyway over time | ||
| [2010/01/07 11:52:32] <duritong> windowsrefund: about nodes: the problem is that with node inheritance includes are evaluated _before_ the variable in the sub-node is taken | ||
| [2010/01/07 11:52:38] <duritong> windowsrefund: yeah it will | ||
| [2010/01/07 11:52:49] @ Quit: rmiller4pi8: Read error: 110 (Connection timed out) | ||
| [2010/01/07 11:53:08] @ yarihm joined channel #puppet | ||
| [2010/01/07 11:54:05] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 11:55:26] <joe-mac> the confusion is when people do things with class inheritance and think it works the same with nodes | ||
| [2010/01/07 11:55:53] @ Quit: rmiller4pi82: Read error: 110 (Connection timed out) | ||
| [2010/01/07 11:56:05] @ WALoeIII joined channel #puppet | ||
| [2010/01/07 11:56:28] <duritong> joe-mac: ack | ||
| [2010/01/07 11:57:51] @ Quit: rmiller4pi81: Read error: 110 (Connection timed out) | ||
| [2010/01/07 11:59:00] @ Quit: skiold: | ||
| [2010/01/07 12:00:18] @ bodepd joined channel #puppet | ||
| [2010/01/07 12:01:48] @ Quit: Robbie_: Remote closed the connection | ||
| [2010/01/07 12:02:21] <Cyis> if I have: file { "/path/to/some/file": ..., alias => "some-file" } ... should I not be able to reference this as File["some-file"] later? | ||
| [2010/01/07 12:08:34] @ Quit: bug: | ||
| [2010/01/07 12:11:55] @ Quit: bjaspan: Read error: 110 (Connection timed out) | ||
| [2010/01/07 12:20:44] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 12:20:50] * windowsrefund is confused about auth.conf and/or namespaceauth.conf | ||
| [2010/01/07 12:21:19] <windowsrefund> I have neither and am running 0.25.1 | ||
| [2010/01/07 12:21:37] @ mattwynne joined channel #puppet | ||
| [2010/01/07 12:22:42] <bodepd> Cyis: yes | ||
| [2010/01/07 12:23:07] <duritong> windowsrefund: I would at least update to 0.25.2 | ||
| [2010/01/07 12:23:08] @ Quit: pacalm: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 12:23:12] <Cyis> think I got it working now... looks like I had some silent failures due to some syntax or something | ||
| [2010/01/07 12:23:16] @ Quit: mattwynne: Client Quit | ||
| [2010/01/07 12:23:42] <duritong> simply because there have been also a lot of fixes for passenger | ||
| [2010/01/07 12:23:49] <bodepd> windowsrefund: did you resolve your .25.2 issues with passenger/mongrel? | ||
| [2010/01/07 12:24:55] <windowsrefund> bodepd: passenger appears to be working for me now although I'm still on 0.25.1 | ||
| [2010/01/07 12:25:15] <bodepd> bodepd:I thought your questions were about a .25.2 upgrade? | ||
| [2010/01/07 12:25:20] <windowsrefund> maybe I'll upgrade to 0.25.2 | ||
| [2010/01/07 12:25:28] <bodepd> yeah, there are so many bug fixes | ||
| [2010/01/07 12:25:32] <windowsrefund> nope, haven't even thought about 0.25.2 yet | ||
| [2010/01/07 12:26:00] <windowsrefund> ok, looks like a basic namespaceauth.conf fixed my issue | ||
| [2010/01/07 12:26:13] <windowsrefund> is auth.conf still used? | ||
| [2010/01/07 12:28:03] <bodepd> check the 0.25.0 release notes | ||
| [2010/01/07 12:28:09] <bodepd> http://reductivelabs.com/trac/puppet/wiki/ReleaseNotes#0.25.0 | ||
| [2010/01/07 12:28:53] <bodepd> auth.conf will replace namespaceauth.conf in .26.0 | ||
| [2010/01/07 12:29:02] <windowsrefund> ok | ||
| [2010/01/07 12:29:04] <bodepd> its for access to the RESTful interface | ||
| [2010/01/07 12:29:25] @ Quit: poison: Remote closed the connection | ||
| [2010/01/07 12:37:40] @ Quit: jaredrhine: Read error: 110 (Connection timed out) | ||
| [2010/01/07 12:39:26] @ Quit: bgupta: Read error: 60 (Operation timed out) | ||
| [2010/01/07 12:48:03] @ lak joined channel #puppet | ||
| [2010/01/07 12:48:11] @ bgupta joined channel #puppet | ||
| [2010/01/07 12:51:12] @ Quit: marcoecc: Remote closed the connection | ||
| [2010/01/07 12:53:02] @ Quit: xerxas: "Leaving." | ||
| [2010/01/07 12:53:39] @ marcoecc joined channel #puppet | ||
| [2010/01/07 12:57:33] @ cwebber_ joined channel #puppet | ||
| [2010/01/07 12:59:35] @ Quit: cwebber: Read error: 60 (Operation timed out) | ||
| [2010/01/07 13:00:37] @ cwebber joined channel #puppet | ||
| [2010/01/07 13:01:31] @ Quit: cwebber_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 13:01:55] @ Boltsky joined channel #puppet | ||
| [2010/01/07 13:03:04] @ mqr joined channel #puppet | ||
| [2010/01/07 13:06:39] @ Quit: bgupta: Connection timed out | ||
| [2010/01/07 13:07:48] @ Quit: jcape1: Read error: 60 (Operation timed out) | ||
| [2010/01/07 13:08:22] @ mqr_ joined channel #puppet | ||
| [2010/01/07 13:08:23] @ Quit: mqr: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 13:09:11] @ toi joined channel #puppet | ||
| [2010/01/07 13:11:31] @ bjaspan joined channel #puppet | ||
| [2010/01/07 13:13:40] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 13:14:42] @ bgupta joined channel #puppet | ||
| [2010/01/07 13:15:24] @ bug joined channel #puppet | ||
| [2010/01/07 13:16:23] @ jaredrhine joined channel #puppet | ||
| [2010/01/07 13:16:47] @ Quit: jmarki: Remote closed the connection | ||
| [2010/01/07 13:18:26] @ Quit: marcoecc: Remote closed the connection | ||
| [2010/01/07 13:19:26] @ QMan joined channel #puppet | ||
| [2010/01/07 13:22:38] @ jmarki joined channel #puppet | ||
| [2010/01/07 13:22:39] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/07 13:23:10] @ Quit: mqr_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 13:23:11] @ giskard joined channel #puppet | ||
| [2010/01/07 13:23:20] @ ambroff joined channel #puppet | ||
| [2010/01/07 13:23:28] @ jcape joined channel #puppet | ||
| [2010/01/07 13:23:58] @ mqr joined channel #puppet | ||
| [2010/01/07 13:24:09] @ Quit: lak: | ||
| [2010/01/07 13:29:46] <Cope> evening | ||
| [2010/01/07 13:30:38] <Cope> I have one line in the default lighttpd config to change; it seems overkill to put the wole config file in puppet; is this an ideal use cse for augeas? Just to enable mod_proxy? | ||
| [2010/01/07 13:31:27] @ Quit: mqr: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 13:31:54] @ mqr joined channel #puppet | ||
| [2010/01/07 13:32:31] @ Quit: jmarki: Remote closed the connection | ||
| [2010/01/07 13:33:57] @ rasputnik joined channel #puppet | ||
| [2010/01/07 13:34:30] @ Quit: rasputnik: Client Quit | ||
| [2010/01/07 13:40:19] <bodepd> thats the use case for augeas. like you mentioned, its starts to break down if you want to manage the entire file. Another options is templates, or the file fragment pattern (but these deal with managing the entire file) | ||
| [2010/01/07 13:40:41] @ Quit: crdant: "Leaving." | ||
| [2010/01/07 13:41:37] @ rasputnik joined channel #puppet | ||
| [2010/01/07 13:41:55] <zahna> does the template function need to have it's path update like the source attribute in the File[] resource? | ||
| [2010/01/07 13:41:58] @ Quit: giskard: Read error: 110 (Connection timed out) | ||
| [2010/01/07 13:42:10] <zahna> when upgrading to 0.25.1? | ||
| [2010/01/07 13:42:25] @ Quit: cwebber: | ||
| [2010/01/07 13:42:45] <Cope> hrm; augtool looks to me like it hasn't indexed any files at all | ||
| [2010/01/07 13:42:54] <Cope> ls /files/ shows nothing | ||
| [2010/01/07 13:43:12] <bodepd> zahna: nope, just source | ||
| [2010/01/07 13:43:23] @ mqr_ joined channel #puppet | ||
| [2010/01/07 13:43:24] <zahna> bodepd: ok, thanks | ||
| [2010/01/07 13:43:38] @ xerxas joined channel #puppet | ||
| [2010/01/07 13:43:57] @ cwebber joined channel #puppet | ||
| [2010/01/07 13:44:30] @ Quit: mqr: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 13:45:03] @ Quit: rasputnik: Client Quit | ||
| [2010/01/07 13:45:26] <Cyis> well boss is letting us get home and work via remote since we have bad weather due in... ttyl | ||
| [2010/01/07 13:45:45] @ Cyis left channel #puppet ("Ex-Chat") | ||
| [2010/01/07 13:46:28] @ Quit: Eghie: Client Quit | ||
| [2010/01/07 13:47:11] @ Quit: xerxas: Client Quit | ||
| [2010/01/07 13:49:28] <Cope> ok this looks pretty non-trivial | ||
| [2010/01/07 13:50:51] @ Quit: MattyM: "ta ta" | ||
| [2010/01/07 13:53:08] @ giskard joined channel #puppet | ||
| [2010/01/07 13:53:29] @ Quit: brothers: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 13:53:34] @ Quit: fluxdude: Read error: 110 (Connection timed out) | ||
| [2010/01/07 13:53:53] @ brothers joined channel #puppet | ||
| [2010/01/07 13:54:26] <bodepd> I would say it depends for augeas, if the lenses exists its trivial... | ||
| [2010/01/07 13:54:50] <bodepd> I think its pretty easy to re-use the existing lenses for .ini files. otherwise it starts to get complicated | ||
| [2010/01/07 13:56:25] @ taggart joined channel #puppet | ||
| [2010/01/07 13:57:07] <taggart> Volcane: hey, I mostly have it working, but I am having problems with the name of the munin plugin | ||
| [2010/01/07 13:57:45] <taggart> Volcane: I want to do something like "smart_$name:" but that doesn't work, do I need to do an inline template or something? | ||
| [2010/01/07 13:58:25] @ mattwynne joined channel #puppet | ||
| [2010/01/07 13:59:04] @ crdant joined channel #puppet | ||
| [2010/01/07 13:59:04] <Volcane> need to see the code to be able to really answer | ||
| [2010/01/07 13:59:39] @ littleidea joined channel #puppet | ||
| [2010/01/07 13:59:41] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 14:00:18] @ jes5 joined channel #puppet | ||
| [2010/01/07 14:00:22] @ lak joined channel #puppet | ||
| [2010/01/07 14:01:26] <jaimeg> Hello Everyone... Quick Question... I wanting to present this product to the business but i need to know what kind of is available on this server. | ||
| [2010/01/07 14:01:51] @ Quit: crdant: Client Quit | ||
| [2010/01/07 14:01:54] <joe-mac> you're the missing the object in your prepositional phrase | ||
| [2010/01/07 14:02:02] <joe-mac> or something like that lol | ||
| [2010/01/07 14:02:18] <joe-mac> you want to know what version of puppet is on a server jaimeg? | ||
| [2010/01/07 14:02:43] <jaimeg> no.. I want to know how will we recover if the puppet master were to take a dump. | ||
| [2010/01/07 14:03:08] <jaimeg> Does the software allow replication? | ||
| [2010/01/07 14:03:24] @ Quit: bjaspan: Read error: 60 (Operation timed out) | ||
| [2010/01/07 14:03:24] <Volcane> the source code that it deploys is just files | ||
| [2010/01/07 14:03:28] <Volcane> put it in svn/git/whatever | ||
| [2010/01/07 14:03:30] <Volcane> and back up | ||
| [2010/01/07 14:04:01] <jaimeg> So there is no fault tolerance? | ||
| [2010/01/07 14:04:02] <loupgaroublond> jaimeg, in the event that the puppet master dies, your systems won't keel over, because once puppet runs on each system, it stays configured that way | ||
| [2010/01/07 14:04:09] <Volcane> jaimeg: http://reductivelabs.com/trac/puppet/wiki/Recipes/HighAvailability | ||
| [2010/01/07 14:04:14] <loupgaroublond> there are various ways you can cluster your puppet master too | ||
| [2010/01/07 14:04:20] <jaimeg> ty Volcane | ||
| [2010/01/07 14:04:30] <taggart> Volcane: found it, needed to be ' "smart_${name}": ' | ||
| [2010/01/07 14:04:32] @ Eghie joined channel #puppet | ||
| [2010/01/07 14:05:26] @ mqr joined channel #puppet | ||
| [2010/01/07 14:07:16] <qwebirc8227> hi there | ||
| [2010/01/07 14:07:43] <qwebirc8227> I've been looking around and can't find a way to display nodes in groups in the dashboard | ||
| [2010/01/07 14:07:46] @ Quit: alban2: Read error: 113 (No route to host) | ||
| [2010/01/07 14:07:57] @ alban2 joined channel #puppet | ||
| [2010/01/07 14:08:16] <qwebirc8227> is it something to be set on dashboard or on puppet side ? | ||
| [2010/01/07 14:08:16] @ Quit: themurph: Read error: 60 (Operation timed out) | ||
| [2010/01/07 14:10:36] @ Quit: mqr_: Read error: 110 (Connection timed out) | ||
| [2010/01/07 14:18:05] @ crdant joined channel #puppet | ||
| [2010/01/07 14:23:18] @ crdant1 joined channel #puppet | ||
| [2010/01/07 14:26:57] @ tecto joined channel #puppet | ||
| [2010/01/07 14:28:10] @ poison joined channel #puppet | ||
| [2010/01/07 14:31:58] <taggart> Volcane (or whoever): now I'm adding a (non-unique) munin plugin inside a define, but that blows up because it's attempting to define more than once | ||
| [2010/01/07 14:32:29] <taggart> is there some way I can add it only if it's not already set? | ||
| [2010/01/07 14:34:23] @ loupgaroublond left channel #puppet () | ||
| [2010/01/07 14:35:25] @ mecca joined channel #puppet | ||
| [2010/01/07 14:36:21] <whack> refactor so you don't define it multiple times | ||
| [2010/01/07 14:36:29] <taggart> I'm doing "munin::plugin { hddtemp_smartctl: ensure => present; }" inside the define | ||
| [2010/01/07 14:36:47] <whack> pastebin your munin::plugin define? | ||
| [2010/01/07 14:37:13] <whack> wait, you're specyfing that resource twice? | ||
| [2010/01/07 14:37:40] <taggart> well it's in the define and then I pass an array to the define | ||
| [2010/01/07 14:38:16] <taggart> I've thought about moving that particular plugin outside the define, and then inside setting a variable or something | ||
| [2010/01/07 14:38:34] @ Quit: crdant: Read error: 110 (Connection timed out) | ||
| [2010/01/07 14:38:44] <taggart> basically if I encounter at least one device that supports SMART then I want to turn that plugin on | ||
| [2010/01/07 14:40:18] <joe-mac> why do you have it inside a define if it's not unique? | ||
| [2010/01/07 14:40:36] <Volcane> make a class munin::smartctl | ||
| [2010/01/07 14:40:42] <Volcane> then just include it once you've determined you need it | ||
| [2010/01/07 14:40:46] <Volcane> u can include the same thing many times | ||
| [2010/01/07 14:41:27] <taggart> joe-mac: the define is checking if a device supports smart and then turning it on | ||
| [2010/01/07 14:41:41] <taggart> Volcane: ah that's one way of doing it | ||
| [2010/01/07 14:41:46] @ jfreeman joined channel #puppet | ||
| [2010/01/07 14:42:06] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 14:42:08] <joe-mac> that's what a class is for like volcane said | ||
| [2010/01/07 14:42:46] @ gebi joined channel #puppet | ||
| [2010/01/07 14:43:33] @ jcape joined channel #puppet | ||
| [2010/01/07 14:47:48] @ Quit: ambroff: Read error: 110 (Connection timed out) | ||
| [2010/01/07 14:48:24] @ Quit: bug: | ||
| [2010/01/07 14:48:42] @ bjaspan joined channel #puppet | ||
| [2010/01/07 14:49:14] @ Quit: lak: | ||
| [2010/01/07 14:49:32] @ Quit: brothers: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 14:49:39] @ lak joined channel #puppet | ||
| [2010/01/07 14:49:47] @ brothers joined channel #puppet | ||
| [2010/01/07 14:49:49] @ Quit: bjaspan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 14:50:21] @ bjaspan joined channel #puppet | ||
| [2010/01/07 14:56:52] @ bug joined channel #puppet | ||
| [2010/01/07 14:58:26] <taggart> Volcane, joe-mac: that works nicely thanks! | ||
| [2010/01/07 14:58:38] @ Djelibeybi joined channel #puppet | ||
| [2010/01/07 14:58:48] <joe-mac> np | ||
| [2010/01/07 14:58:52] @ Quit: poison: Remote closed the connection | ||
| [2010/01/07 14:58:56] @ mattwynne joined channel #puppet | ||
| [2010/01/07 15:00:50] @ LinuxCode joined channel #puppet | ||
| [2010/01/07 15:03:47] @ bpgoldsb joined channel #puppet | ||
| [2010/01/07 15:04:41] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 15:04:57] @ mattwynne joined channel #puppet | ||
| [2010/01/07 15:14:52] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 15:16:11] @ rasputnik joined channel #puppet | ||
| [2010/01/07 15:18:46] @ jcape joined channel #puppet | ||
| [2010/01/07 15:19:12] @ ambroff joined channel #puppet | ||
| [2010/01/07 15:19:51] <jaimeg> Hey guys. I was just reading a section of the puppet website. it stated that puppet was using webrick and that it could only handle 10 to 20 clients? | ||
| [2010/01/07 15:19:55] <jaimeg> is that acurate? | ||
| [2010/01/07 15:19:58] <barn> yes | ||
| [2010/01/07 15:20:03] <Volcane> by default | ||
| [2010/01/07 15:20:15] @ rasputnik left channel #puppet ("Leaving") | ||
| [2010/01/07 15:20:16] <barn> use mongrels + apache/nginx or passenger + apache/nginx for more | ||
| [2010/01/07 15:20:37] <jaimeg> what is webrick used for? | ||
| [2010/01/07 15:21:31] <barn> it's ruby's simple webstack | ||
| [2010/01/07 15:21:35] <Volcane> its the default werbserver in most rails apps | ||
| [2010/01/07 15:21:37] <Volcane> suitable for testing etc | ||
| [2010/01/07 15:22:03] <Volcane> and when u set it up with passenger or mongrel, u are replacing webrick with those | ||
| [2010/01/07 15:22:54] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/07 15:24:02] <LinuxCode> evening Volcane | ||
| [2010/01/07 15:24:15] <LinuxCode> Volcane, I ended up replacing my two tidy with execs | ||
| [2010/01/07 15:24:18] <LinuxCode> got fed up | ||
| [2010/01/07 15:24:21] <LinuxCode> hehe | ||
| [2010/01/07 15:24:36] @ Quit: QMan: Connection reset by peer | ||
| [2010/01/07 15:24:45] <Volcane> LinuxCode: hehe | ||
| [2010/01/07 15:25:03] @ QMan joined channel #puppet | ||
| [2010/01/07 15:26:28] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 15:27:09] <fzzzt> jaimeg: The number of nodes webrick will support depends on how complex they are. We're running 35 with no issues, but only two are what I would call complex. Most are pretty simple. | ||
| [2010/01/07 15:27:32] <fzzzt> I will probably migrate over anyway so I don't have to deal with when they are too complex, though. :) | ||
| [2010/01/07 15:29:07] @ poison joined channel #puppet | ||
| [2010/01/07 15:30:02] <fzzzt> I do get eval_generate errors a lot, though I have no idea what causes those, and it has always happened as far as I can remember... | ||
| [2010/01/07 15:31:56] @ ezmobius joined channel #puppet | ||
| [2010/01/07 15:32:37] <windowsrefund> does anyone see why my code in my puppet class won't run on a node that uses puppet::slave? http://pastebin.com/m6704b417 | ||
| [2010/01/07 15:33:27] <LinuxCode> Volcane, with tidy, can it be it only likes one tidy ? | ||
| [2010/01/07 15:33:49] <Volcane> never tried it | ||
| [2010/01/07 15:33:51] <LinuxCode> in a definition, or that maybe file and tidies cant have the same titles ? | ||
| [2010/01/07 15:34:13] <LinuxCode> because when I used the same paths, it gave me duplicate errors | ||
| [2010/01/07 15:34:35] <LinuxCode> replaced with execs, which were just changed into execs from those tidies, and that worked | ||
| [2010/01/07 15:35:21] @ themurph joined channel #puppet | ||
| [2010/01/07 15:35:38] @ Quit: littleidea: | ||
| [2010/01/07 15:37:18] @ rmiller4pi81 joined channel #puppet | ||
| [2010/01/07 15:43:59] @ ibolmo_ joined channel #puppet | ||
| [2010/01/07 15:44:42] <windowsrefund> I'm using 'include puppet' from within a define named puppet::server, yet the code in modules/puppet/manifests/init.pp does not run | ||
| [2010/01/07 15:45:03] @ jusfreeman joined channel #puppet | ||
| [2010/01/07 15:45:17] <windowsrefund> yet, it runs if I take that code and move it to puppet::packages (modules/puppet/manifests/packages.pp) | ||
| [2010/01/07 15:45:31] <fzzzt> I think puppet::server should require puppet happens first... | ||
| [2010/01/07 15:45:46] <ibolmo_> anyone have a samba recipe ? | ||
| [2010/01/07 15:45:51] <fzzzt> I read that somewhere | ||
| [2010/01/07 15:46:11] @ Quit: poison: Read error: 113 (No route to host) | ||
| [2010/01/07 15:46:30] <windowsrefund> fzzzt: well, regardless of order... the stuff in init.pp should run | ||
| [2010/01/07 15:46:43] <fzzzt> yep | ||
| [2010/01/07 15:46:51] <windowsrefund> I've got other working modules designed like this also | ||
| [2010/01/07 15:46:54] <windowsrefund> just so odd | ||
| [2010/01/07 15:50:07] <joe-mac> including inside defines is kind of wonky | ||
| [2010/01/07 15:50:50] <joe-mac> i was going to make one ibolmo_ since the samba fiole has an include functionality, and just push out a base config with an include for a hostshares.cfg or something | ||
| [2010/01/07 15:50:58] @ Quit: rmiller4pi83: Read error: 110 (Connection timed out) | ||
| [2010/01/07 15:51:18] <ibolmo_> would be great :D | ||
| [2010/01/07 15:51:38] <joe-mac> haven't gotten around to it | ||
| [2010/01/07 15:51:44] <joe-mac> probably because i hate samba | ||
| [2010/01/07 15:52:29] @ crashdummyMCH joined channel #puppet | ||
| [2010/01/07 15:53:50] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 15:53:51] <ibolmo_> haha | ||
| [2010/01/07 15:54:12] <ibolmo_> i'm getting started so any lessons learned would be great to know about ahead of time | ||
| [2010/01/07 15:54:27] <joe-mac> with samba? or puppet? | ||
| [2010/01/07 15:55:07] <ibolmo_> hehe both | ||
| [2010/01/07 15:55:14] <ibolmo_> but i understand the purpose of puppet | ||
| [2010/01/07 15:55:50] <ibolmo_> samba, ldap, and a bunch of other techs i'm still understanding where they all fit in a network | ||
| [2010/01/07 15:55:55] @ Quit: rmiller4pi8: Read error: 110 (Connection timed out) | ||
| [2010/01/07 15:56:22] <ibolmo_> my business is starting to grow and i'm having to have more IT infra | ||
| [2010/01/07 15:57:27] <joe-mac> don't eat yellow snow, everything else, just wing it | ||
| [2010/01/07 15:58:17] <ibolmo_> haha | ||
| [2010/01/07 15:58:19] <ibolmo_> exactly | ||
| [2010/01/07 15:58:22] <ibolmo_> i'm winging it | ||
| [2010/01/07 15:58:53] @ Quit: nexx|: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:00:24] @ Quit: themurph: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 16:00:30] @ nexx| joined channel #puppet | ||
| [2010/01/07 16:00:38] @ themurph joined channel #puppet | ||
| [2010/01/07 16:01:12] @ Quit: jfreeman: Read error: 110 (Connection timed out) | ||
| [2010/01/07 16:02:10] @ crdant joined channel #puppet | ||
| [2010/01/07 16:02:42] @ Quit: crdant1: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:04:16] <ashp> oh god our fucking developers | ||
| [2010/01/07 16:04:20] <ashp> i wish i had explosives | ||
| [2010/01/07 16:04:41] <ibolmo_> what'd they do? | ||
| [2010/01/07 16:05:14] <ashp> we just deployed this huge project built on sun's IDM | ||
| [2010/01/07 16:05:18] <ashp> we used v8.0 | ||
| [2010/01/07 16:05:27] <ashp> turns out there's _13_ big patches to that and nobody ever even checked for them | ||
| [2010/01/07 16:05:33] <windowsrefund> 0.25.2 appears to be fubar on opensolaris | ||
| [2010/01/07 16:05:34] <ashp> so we're 13 patches behind right out of the door | ||
| [2010/01/07 16:05:49] <ashp> they also developed an entire chunk of stuff in the IDM without ONCE looking at the input feed file that we REALLY get | ||
| [2010/01/07 16:05:58] <ashp> and instead developed it with a bunch of assumptions that data that doesn't exist will be present | ||
| [2010/01/07 16:06:24] <ibolmo_> haha ouch | ||
| [2010/01/07 16:07:24] @ crdant1 joined channel #puppet | ||
| [2010/01/07 16:07:43] @ Quit: crdant: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:09:30] @ Quit: ibolmo_: "Page closed" | ||
| [2010/01/07 16:10:19] @ fzzzt left channel #puppet () | ||
| [2010/01/07 16:11:31] @ Quit: labrown: "Leaving" | ||
| [2010/01/07 16:17:25] @ Quit: Demosthenes: "leaving" | ||
| [2010/01/07 16:17:39] @ paxos joined channel #puppet | ||
| [2010/01/07 16:17:51] @ Quit: bjaspan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:17:51] @ bjaspan_ joined channel #puppet | ||
| [2010/01/07 16:17:58] @ bjaspan_ is now known as bjaspan | ||
| [2010/01/07 16:18:42] @ qwebirc59094 joined channel #puppet | ||
| [2010/01/07 16:18:51] <qwebirc59094> Hello | ||
| [2010/01/07 16:19:14] @ Quit: bodepd: | ||
| [2010/01/07 16:19:23] <qwebirc59094> I am brand new to this project. I just stumbled across it, because I am looking for a way to create a new user on 400+ servers. | ||
| [2010/01/07 16:20:09] <joe-mac> and other stuff too i hope? | ||
| [2010/01/07 16:20:14] <joe-mac> be a waste to use puppet for only that | ||
| [2010/01/07 16:20:43] @ tessier is now known as turgid | ||
| [2010/01/07 16:21:18] @ Quit: jusfreeman: Read error: 110 (Connection timed out) | ||
| [2010/01/07 16:21:24] @ turgid is now known as tessier | ||
| [2010/01/07 16:21:26] <qwebirc59094> quit | ||
| [2010/01/07 16:21:29] <qwebirc59094> logout | ||
| [2010/01/07 16:21:30] <qwebirc59094> exit | ||
| [2010/01/07 16:21:31] <qwebirc59094> stop | ||
| [2010/01/07 16:21:33] <qwebirc59094> !xit | ||
| [2010/01/07 16:21:39] <qwebirc59094> logoff | ||
| [2010/01/07 16:21:42] @ qwebirc59094 left channel #puppet () | ||
| [2010/01/07 16:21:47] <taggart> haha | ||
| [2010/01/07 16:22:46] @ Quit: Bass10: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:23:00] @ Quit: Eghie: Remote closed the connection | ||
| [2010/01/07 16:24:11] <eric0> :wq! | ||
| [2010/01/07 16:24:15] @ Bass10 joined channel #puppet | ||
| [2010/01/07 16:26:07] @ crdant joined channel #puppet | ||
| [2010/01/07 16:26:44] <paxos> congratulations on 0.25.2 release.. even see that my two puppetrun bugs were fixed! | ||
| [2010/01/07 16:28:57] <paxos> I think all the bugs that I've encountered while using 0.25.1 have been fixed in 0.25.2.. awesome! | ||
| [2010/01/07 16:33:09] @ JM joined channel #puppet | ||
| [2010/01/07 16:33:56] @ Quit: Bass10: Read error: 60 (Operation timed out) | ||
| [2010/01/07 16:36:55] @ Quit: crdant1: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:37:51] @ crdant1 joined channel #puppet | ||
| [2010/01/07 16:38:56] @ bodepd joined channel #puppet | ||
| [2010/01/07 16:39:16] <joe-mac> grrrrrrr i think something is not getting updated in my stored config db and causing me to go on a chase | ||
| [2010/01/07 16:39:41] @ Quit: crdant: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 16:44:09] <joe-mac> how do i look at exported resources in the db? | ||
| [2010/01/07 16:45:19] <joe-mac> i su ck at sql | ||
| [2010/01/07 16:48:28] <Juzzy> whats stored in sql? | ||
| [2010/01/07 16:50:32] <joe-mac> my configuration | ||
| [2010/01/07 16:50:56] <joe-mac> it looks like the target attribute on an exported nagios_host resource is not getting updated when i change the manifest | ||
| [2010/01/07 16:51:05] <joe-mac> of course it could be PEBKAC... | ||
| [2010/01/07 16:53:54] <danielbln> anyone else have clients unexpectedly die when the master is out of reach? | ||
| [2010/01/07 16:54:10] <danielbln> my master crashed and now I have had 200 puppet client processes just vanish | ||
| [2010/01/07 16:54:28] * danielbln shows a look of disapproval | ||
| [2010/01/07 16:55:36] @ jcape joined channel #puppet | ||
| [2010/01/07 16:55:42] <joe-mac> yea that was weird, i had to clear stuff frmo the db | ||
| [2010/01/07 16:56:44] @ Quit: toi: Read error: 113 (No route to host) | ||
| [2010/01/07 16:56:46] @ Quit: tecto: | ||
| [2010/01/07 16:56:56] <danielbln> guess I'll have to do the 'ole cronjob-to-check-if-service-is-running trick | ||
| [2010/01/07 16:57:12] <danielbln> let's see if this happens with 0.25.2 | ||
| [2010/01/07 16:57:48] <Volcane> there's a ticket - probably closed but there are may cases that can case it | ||
| [2010/01/07 16:58:26] <danielbln> what do you do to catch that behaviour? | ||
| [2010/01/07 16:58:26] <tessier> danielbln: Speaking of which...How do most people configure puppet to ensure that puppet itself is running? I noticed that cfengine had a cron to start cfengine and cfengine ensured that cron was running and the cronjob was in place. | ||
| [2010/01/07 16:58:42] <tessier> Do people configure puppet in the same way? | ||
| [2010/01/07 16:58:57] <danielbln> I have a Nagios script check the state.yaml and have some monitoring monkey restart the service | ||
| [2010/01/07 16:59:19] <danielbln> but that's not ideal when all clients in the infrastructure die | ||
| [2010/01/07 16:59:20] <Volcane> danielbln: you comment on the tiket with the output from --debug --trace and let RL fix it :) | ||
| [2010/01/07 16:59:27] <danielbln> will do | ||
| [2010/01/07 17:00:21] <Djelibeybi> tessier, danielbln: I have a cronjob that checks that puppetd is running | ||
| [2010/01/07 17:00:33] * Volcane just runs it from cron | ||
| [2010/01/07 17:00:58] @ Quit: crdant1: "Leaving." | ||
| [2010/01/07 17:00:59] <Djelibeybi> Volcane: yeah, I'm considering that for my next round of upgrades. Switching from puppetd to puppet --one-time in cron | ||
| [2010/01/07 17:01:10] @ eventi joined channel #puppet | ||
| [2010/01/07 17:01:24] <bodepd> windowsrefund: can you be more specific about what your issue with .25.2 on opensolaris is? | ||
| [2010/01/07 17:01:27] <Volcane> yup, with a random sleep | ||
| [2010/01/07 17:01:30] <danielbln> mh, I never considered using one-time, that's a nice way of doing it | ||
| [2010/01/07 17:01:35] <Volcane> much++ happier since doing that | ||
| [2010/01/07 17:02:24] @ Quit: lilmatt: Client Quit | ||
| [2010/01/07 17:03:16] <eventi> hey - suddenly got a message from puppetd: "Configuration could not be instantiated: Parameter path failed: File paths must be fully qualified" | ||
| [2010/01/07 17:03:22] <eventi> won't tell me where it failed | ||
| [2010/01/07 17:03:35] <bodepd> eventi: paths must start with leading '/' | ||
| [2010/01/07 17:03:37] <whack> Volcane: I use puppetd --test + randomsleep in daemontools | ||
| [2010/01/07 17:03:49] <whack> fwiw, much happier than with puppetd in daemon mode and splay | ||
| [2010/01/07 17:04:06] <taggart> yes here too | ||
| [2010/01/07 17:05:04] <danielbln> whack: what do you do against client failure? | ||
| [2010/01/07 17:05:15] <Volcane> $(($RANDOM % 900)) && /usr/sbin/puppetd --onetime | ||
| [2010/01/07 17:05:21] <Volcane> erm, put sleep infront | ||
| [2010/01/07 17:05:44] <danielbln> don't forget to tell cron to use bash if you want to use it as a direct crontab line though | ||
| [2010/01/07 17:05:55] <danielbln> sh is oblivious to the random variable :/ | ||
| [2010/01/07 17:05:56] * Volcane has it in a script and cron the script | ||
| [2010/01/07 17:06:04] <danielbln> yeah, that'll do | ||
| [2010/01/07 17:07:02] <danielbln> had about 500 cron mails the other day, all complaining about what to do with the random variable | ||
| [2010/01/07 17:08:12] <joe-mac> i throw cron mails in the dumpster | ||
| [2010/01/07 17:08:15] <whack> danielbln: the randomsleep averages about 30 minutes | ||
| [2010/01/07 17:08:20] <whack> danielbln: daemontools restarts it when the sleep finishes | ||
| [2010/01/07 17:08:34] <danielbln> ic | ||
| [2010/01/07 17:08:45] @ somnoliento joined channel #puppet | ||
| [2010/01/07 17:08:45] <eventi> bodepd: any way to find out where in the manifest it fails? | ||
| [2010/01/07 17:08:54] <whack> so I do puppet first, then sleep, so I can instantly run puppet on demand with 'svc -t /service/puppet' since it doesn't sleep at the start | ||
| [2010/01/07 17:09:05] <bodepd> it should list the line number, can you pastie the failure message? | ||
| [2010/01/07 17:09:59] <eventi> http://pastie.org/770791 | ||
| [2010/01/07 17:10:04] <eventi> this one doesnt | ||
| [2010/01/07 17:10:37] <eventi> event with --debug | ||
| [2010/01/07 17:11:34] <bodepd> this is strange, this doesnt seem to be coming form you manifests. | ||
| [2010/01/07 17:11:57] <bodepd> you can propagate the line number in the source with --trace | ||
| [2010/01/07 17:12:09] <bodepd> what version is the client/server? | ||
| [2010/01/07 17:13:01] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 17:14:33] <whack> eventi: does the file you're looking for (puppet://....) exist on your fileserver? | ||
| [2010/01/07 17:15:24] <eventi> whack: that would give a different error | ||
| [2010/01/07 17:16:13] <eventi> bodepd: --trace doesn't help - it says what line in puppetd fails, but not what manifest | ||
| [2010/01/07 17:16:50] @ Quit: lero: Remote closed the connection | ||
| [2010/01/07 17:17:26] <bodepd> try with --pluginsync false | ||
| [2010/01/07 17:17:35] @ jcape joined channel #puppet | ||
| [2010/01/07 17:17:55] <bodepd> something seems to be screwed up with your plugins. | ||
| [2010/01/07 17:18:34] <bodepd> this woudnt be related to a manifest, this might be a directory org issue. Maybe a version issue, what version is this on? what are the client/server versions? | ||
| [2010/01/07 17:18:56] <eventi> sorry - it worked an hour ago just stopped - one sec | ||
| [2010/01/07 17:19:14] <eventi> client 0.24.8 | ||
| [2010/01/07 17:19:35] <eventi> server 0.24.8 | ||
| [2010/01/07 17:24:52] <eventi> disabling pluginsync failed - but I reverted my configs, so I think it is on the client side | ||
| [2010/01/07 17:26:42] @ tomas_ joined channel #puppet | ||
| [2010/01/07 17:29:41] <crashdummyMCH> hi, I'm new to puppet and am trying to find out if puppet works with opensuse zypper and can't find any reliable information. Anyone have any experience with opensuse? | ||
| [2010/01/07 17:32:15] <danielbln> Volcane: do you ever miss the puppetrun functionality? | ||
| [2010/01/07 17:32:16] @ Quit: bug: | ||
| [2010/01/07 17:32:26] <danielbln> since you use cron and all | ||
| [2010/01/07 17:32:39] <Volcane> danielbln: nope, http://code.google.com/p/mcollective-plugins/wiki/AgentPuppetd | ||
| [2010/01/07 17:33:14] <Volcane> miles better than puppetrun | ||
| [2010/01/07 17:33:32] <Volcane> mc-puppet --with-class /someclass/ --with-fact somefact=somevalue runonce | ||
| [2010/01/07 17:33:49] <danielbln> gah, yet another layer | ||
| [2010/01/07 17:34:33] @ Quit: rsquared: "Leaving" | ||
| [2010/01/07 17:34:51] @ rsquared joined channel #puppet | ||
| [2010/01/07 17:34:59] <Volcane> but also in the ext directory i think is a puppetrun replacement for cron environments | ||
| [2010/01/07 17:36:57] @ Quit: kolla: Remote closed the connection | ||
| [2010/01/07 17:38:05] @ Quit: themurph: Read error: 60 (Operation timed out) | ||
| [2010/01/07 17:40:43] @ jcape1 joined channel #puppet | ||
| [2010/01/07 17:41:22] @ lilmatt joined channel #puppet | ||
| [2010/01/07 17:43:01] @ Quit: windowsrefund: "leaving" | ||
| [2010/01/07 17:47:32] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/07 17:49:25] @ littleidea joined channel #puppet | ||
| [2010/01/07 17:51:14] @ docelic joined channel #puppet | ||
| [2010/01/07 17:51:52] <joe-mac> it it just me or do the nagios types sometimes fuck up and write random unreadable characters to the config files | ||
| [2010/01/07 17:51:58] <danielbln> ha! got the mofo | ||
| [2010/01/07 17:52:05] <danielbln> usr/lib/ruby/1.8/timeout.rb:60:in `rbuf_fill': execution expired (Timeout::Error) | ||
| [2010/01/07 17:54:48] <joe-mac> anybody seen a bug with corrupt data being written by naginator? | ||
| [2010/01/07 17:54:49] <joe-mac> http://www.pastie.org/770861 | ||
| [2010/01/07 17:55:06] <joe-mac> that really, sucks | ||
| [2010/01/07 17:55:53] <joe-mac> i know it's puppet doing this, if i remove the file and it gets re created by naginator, the corrupt data still gets written | ||
| [2010/01/07 17:57:36] @ joe-mac left channel #puppet () | ||
| [2010/01/07 17:58:03] @ Quit: crashdummyMCH: Read error: 110 (Connection timed out) | ||
| [2010/01/07 18:00:57] @ Quit: cwebber: | ||
| [2010/01/07 18:15:07] @ kolla joined channel #puppet | ||
| [2010/01/07 18:18:46] @ lero joined channel #puppet | ||
| [2010/01/07 18:25:41] @ pting joined channel #puppet | ||
| [2010/01/07 18:29:01] <jrojas> has anyone seen 0.25.2 clients just exit without saying a word? | ||
| [2010/01/07 18:31:13] <danielbln> not yet, but with 0.25.1 | ||
| [2010/01/07 18:31:19] <danielbln> I'm testing 0.25.2 in a minute | ||
| [2010/01/07 18:31:35] <danielbln> was/is your master unavailable by any chance? | ||
| [2010/01/07 18:31:48] <jrojas> no | ||
| [2010/01/07 18:32:00] <jrojas> its on my dev farm so its lightly used | ||
| [2010/01/07 18:32:15] <jrojas> im stracing the puppet daemon now, not sure where it it disappearing | ||
| [2010/01/07 18:32:45] <jrojas> it writes the finished catalog run then goes away | ||
| [2010/01/07 18:33:52] <jrojas> very strange. | ||
| [2010/01/07 18:34:20] @ jmarki joined channel #puppet | ||
| [2010/01/07 18:34:49] <LinuxCode> jrojas, out of curiosity, what are you using to strace ? | ||
| [2010/01/07 18:35:04] <jrojas> strace -p <pidof puppet daemon> | ||
| [2010/01/07 18:35:10] <LinuxCode> k | ||
| [2010/01/07 18:35:42] <jrojas> i see a trace output but im not getting it in the logs.. | ||
| [2010/01/07 18:35:49] <danielbln> mh, I think 0.25.2 breaks init-script status detection | ||
| [2010/01/07 18:36:28] <danielbln> the client performs process table detection although the service is configured with "hasstatus => true" | ||
| [2010/01/07 18:36:53] <danielbln> anyone else notice that problem? debian here, nfs-common service | ||
| [2010/01/07 18:39:36] @ Quit: _elmata_: Read error: 60 (Operation timed out) | ||
| [2010/01/07 18:40:00] @ _elmata_ joined channel #puppet | ||
| [2010/01/07 18:40:01] @ jmarki left channel #puppet ("Leaving") | ||
| [2010/01/07 18:40:47] <jrojas> Failed to generate additional resources using 'eval_generate': Invalid argument | ||
| [2010/01/07 18:40:50] <jrojas> hmm is that just one of those informational messages? | ||
| [2010/01/07 18:41:52] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/07 18:44:11] @ mattwynne joined channel #puppet | ||
| [2010/01/07 18:45:22] @ Quit: docelic: "http://www.spinlocksolutions.com/" | ||
| [2010/01/07 18:49:11] @ Quit: bobbyz: Read error: 60 (Operation timed out) | ||
| [2010/01/07 18:51:06] <lero> i'm running 0.25.2 w/o problems here | ||
| [2010/01/07 18:51:11] <lero> both server/clients | ||
| [2010/01/07 18:53:05] <danielbln> is there a way to crank up debug levels? | ||
| [2010/01/07 18:53:12] <jrojas> --debug --trace | ||
| [2010/01/07 18:53:12] <danielbln> or is --debug max. verbosity? | ||
| [2010/01/07 18:53:16] <danielbln> even more? | ||
| [2010/01/07 18:53:23] <paxos> --debug gives you a lot | ||
| [2010/01/07 18:53:28] <jrojas> im seeing a Errno::EINVAL but im not able to capture it in the log files | ||
| [2010/01/07 18:53:42] <danielbln> I know, but I need just a little more information out of it | ||
| [2010/01/07 18:54:06] <danielbln> oh well, let's have a look at the code then | ||
| [2010/01/07 18:54:47] <jrojas> danielbln: what type of info are you looking for that its not spitting out | ||
| [2010/01/07 18:55:02] <danielbln> debug: Service[nfs-common](provider=debian): Executing '/etc/init.d/nfs-common status' | ||
| [2010/01/07 18:55:11] <danielbln> I'd like to know, what exit status puppet received | ||
| [2010/01/07 18:55:25] <danielbln> since it's not reacting to the status as is should | ||
| [2010/01/07 18:55:28] <jrojas> hm | ||
| [2010/01/07 18:55:37] <danielbln> *as it | ||
| [2010/01/07 18:55:42] <ricky> mqr: Well, speak of the devil ^ ;-) | ||
| [2010/01/07 18:58:04] @ Quit: notbrien: | ||
| [2010/01/07 18:58:11] <ricky> (the EINVAL, that is) | ||
| [2010/01/07 18:58:33] <jrojas> ricky: yeah looking at the tickets now.... | ||
| [2010/01/07 18:58:51] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 18:59:15] <danielbln> jrojas: found the exit code.. oddly enough, it's 256 oO | ||
| [2010/01/07 18:59:28] * ricky notices that there's a read_nonblock in util/selinux.rb which we also don't handle EINVAL for. | ||
| [2010/01/07 19:00:15] <jrojas> Starting puppet: /usr/lib/ruby/site_ruby/1.8/puppet/util/suidmanager.rb:29:in `asuser': Invalid argument (Errno::EINVAL | ||
| [2010/01/07 19:00:22] <jrojas> ignore the starting puppet | ||
| [2010/01/07 19:01:13] @ Quit: gebi: Read error: 60 (Operation timed out) | ||
| [2010/01/07 19:01:19] @ qwebirc3398 joined channel #puppet | ||
| [2010/01/07 19:01:53] @ Quit: qwebirc3398: Client Quit | ||
| [2010/01/07 19:03:30] <jrojas> lin 29 is old_groups = self.groups | ||
| [2010/01/07 19:05:53] <tessier> Does puppet make a backup copy of files it replaces? | ||
| [2010/01/07 19:07:05] <tessier> Crap. I think I overwrote something important. :( | ||
| [2010/01/07 19:07:06] <barn> it can do | ||
| [2010/01/07 19:07:42] <tessier> Does it by default or is that something I have to configure? | ||
| [2010/01/07 19:08:06] <barn> you have to configure it I'm afraid | ||
| [2010/01/07 19:09:23] @ Quit: pting: Read error: 110 (Connection timed out) | ||
| [2010/01/07 19:10:18] @ mattwynne joined channel #puppet | ||
| [2010/01/07 19:10:27] @ LancerDeuce joined channel #puppet | ||
| [2010/01/07 19:11:16] @ Cyis joined channel #puppet | ||
| [2010/01/07 19:11:20] @ pting joined channel #puppet | ||
| [2010/01/07 19:13:03] <ricky> jrojas: Hmm, how many supplementary groups is whatever you're running puppet as in? | ||
| [2010/01/07 19:13:19] <jrojas> hmma log. | ||
| [2010/01/07 19:13:22] <jrojas> a lot | ||
| [2010/01/07 19:13:32] <ricky> Or does a short ruby script with just "Process.groups" also give the EINTR? | ||
| [2010/01/07 19:13:44] <ricky> If anything this looks like a bug in ruby - it's hardcoded to 32 :-( | ||
| [2010/01/07 19:14:09] <jrojas> that would suck | ||
| [2010/01/07 19:14:13] @ xerxas joined channel #puppet | ||
| [2010/01/07 19:14:14] <jrojas> its all from ldapo | ||
| [2010/01/07 19:14:15] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/07 19:15:03] <jrojas> things is, dont have this issue at all with 24.8 | ||
| [2010/01/07 19:15:10] <jrojas> wow i am horrible at typing today | ||
| [2010/01/07 19:16:02] @ Quit: littleidea: | ||
| [2010/01/07 19:16:04] <ricky> 24.8 apparently doesn't use Process.groups | ||
| [2010/01/07 19:16:08] <ricky> That really does suck :-( | ||
| [2010/01/07 19:16:10] <jrojas> hmm | ||
| [2010/01/07 19:16:13] <jrojas> well shit | ||
| [2010/01/07 19:16:24] <ricky> There's even a variable describing the limit: Process.maxgroups | ||
| [2010/01/07 19:16:30] <ricky> (In ruby, that is) | ||
| [2010/01/07 19:16:52] <jrojas> hmm | ||
| [2010/01/07 19:17:14] <ricky> I'd definitely file a ticket to see if the limitation can be worked around in puppet | ||
| [2010/01/07 19:17:16] @ Quit: WALoeIII: Read error: 113 (No route to host) | ||
| [2010/01/07 19:18:30] <ricky> The usage was introduced in ticket #1806 for what it's worth | ||
| [2010/01/07 19:18:31] <gepetto> ricky: #1806 is http://projects.reductivelabs.com/issues/show/1806 "Puppet - Bug #1806: supplementary groups are not reset - ReductiveLabs.com" | ||
| [2010/01/07 19:18:32] <jrojas> yeah, for some reason root is in 57 groups... uhg i hate infrastructures that are old | ||
| [2010/01/07 19:20:49] @ Quit: jab_doa: "Verlassend" | ||
| [2010/01/07 19:25:12] <Cyis> Hmm... if you have a File["somefile"] resource already and you wanted to create a symlink for it... could you use the File[] resource itself to reference it or am I just thinking crazy? | ||
| [2010/01/07 19:26:58] <taggart> Cyis: you mean the same File resource or a new one for the symlink? | ||
| [2010/01/07 19:27:30] <jrojas> hmm | ||
| [2010/01/07 19:27:32] <jrojas> i think i fixed it.. | ||
| [2010/01/07 19:27:36] <jrojas> or at least found awork-around | ||
| [2010/01/07 19:28:30] <Cyis> taggart, ie- I have File["/path/to/fileA"] and I want to have /path/to/fileB be a symlink to it... I could do file { "/path/to/fileB": ensure => "/path/to/fileA" } but could I do file { "/path/to/fileB": ensure => File["/path/to/fileA"] } | ||
| [2010/01/07 19:29:19] <Cyis> I know it doesn't save anything when you use full path... but if you were using an alias (ie- file { "/path/to/fileA": alias => "fileA", ... } then use File["fileA"]) | ||
| [2010/01/07 19:30:05] <taggart> hmm I don't know, did you try it? | ||
| [2010/01/07 19:30:11] <Cyis> or is my brain just running on overdrive with too much blood in the caffine stream | ||
| [2010/01/07 19:30:43] <Cyis> hadn't tried yet... was just working on one of my defines and thought about it... | ||
| [2010/01/07 19:30:48] @ Quit: kaptk2: "Leaving." | ||
| [2010/01/07 19:31:14] <Volcane> you cant | ||
| [2010/01/07 19:31:35] <Cyis> Volcane, ok... overdrive it is :) | ||
| [2010/01/07 19:31:37] @ Quit: QMan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 19:32:11] <Cyis> was trying to be clever with creating multiple symlinks to a script that behaves differently based on the basename it is called with | ||
| [2010/01/07 19:32:22] @ QMan joined channel #puppet | ||
| [2010/01/07 19:35:53] @ Quit: jaredrhine: Read error: 60 (Operation timed out) | ||
| [2010/01/07 19:37:15] <mqr> ricky (and any other members of the EINTR interest group): I'm trying to figure out two things regarding EINTR from IO#sysread: | ||
| [2010/01/07 19:37:36] <mqr> 1) what circumstances lead to it | ||
| [2010/01/07 19:37:45] <mqr> 2) A way to force these circumstances | ||
| [2010/01/07 19:39:40] <ricky> You'd need to send a signal to the process while the read syscall is running (a perfectly timed kill pid in the middle of when the read is running) | ||
| [2010/01/07 19:40:32] <ricky> I have on idea how to actually get that perfect timing, but in theory, it can happen | ||
| [2010/01/07 19:41:15] <ricky> man 7 signal should explain it in more detali | ||
| [2010/01/07 19:42:43] <mqr> I've tried setting up a source process that extends the window, flood killing, etc., with no luck. | ||
| [2010/01/07 19:43:09] * ricky sees if he can setup a situatoin | ||
| [2010/01/07 19:43:28] <mqr> with sysread the window looks like it should be pretty wide to begin with; | ||
| [2010/01/07 19:43:52] <mqr> as long as the source doesn't close of fill the buffer, right? | ||
| [2010/01/07 19:44:01] <mqr> s/of/or/ | ||
| [2010/01/07 19:44:10] <jrojas> adding Process.maxgroups=128 to configurer.rb worked for me in keepting it from dying after one run | ||
| [2010/01/07 19:44:13] <ricky> Yeah. | ||
| [2010/01/07 19:44:23] <jrojas> granted there is a better way to fix i | ||
| [2010/01/07 19:44:24] <jrojas> t | ||
| [2010/01/07 19:44:29] <ricky> jrojas: Hah, nice - I had no idea you could configure that | ||
| [2010/01/07 19:45:12] <jrojas> in class << self | ||
| [2010/01/07 19:45:18] <jrojas> it works without issue for now.. | ||
| [2010/01/07 19:45:21] <mqr> As for the one in SELinux, that is coming from /proc/mounts, and shouln't even die on you. | ||
| [2010/01/07 19:46:50] <mqr> Another thing that I've not gotten my head around: what does the Process.groups have to do with EINTR? | ||
| [2010/01/07 19:47:57] <ricky> getgroups apparently returns EINTR in a situation that has absolutely nothing to do with interrupted syscalls | ||
| [2010/01/07 19:48:17] <mqr> Ah. | ||
| [2010/01/07 19:50:41] <Cyis> head has been in code too long... need to step back and take a fresh look at things :) | ||
| [2010/01/07 19:50:58] <duritong> anybody an idea how I can delete a user from a group with augeas? | ||
| [2010/01/07 19:51:39] <lero> what's the difference running puppetmaster standalone and with mongrel? | ||
| [2010/01/07 19:52:42] @ Quit: brothers: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 19:52:51] <ricky> mqr: OK, here's a somewhat convoluted case, but it doesn't look to far from the signal handler that puppetd installs: http://dpaste.com/142329/ | ||
| [2010/01/07 19:52:55] @ brothers joined channel #puppet | ||
| [2010/01/07 19:52:56] <ricky> Run that and kill -USR1 the pid it prints | ||
| [2010/01/07 19:53:36] <mqr> ricky: I'm not seeing where any of the getgroups-related stuff generates an EINTR in the ruby source, nor is it listed in the man pages I've looked at. What OS, do you know? | ||
| [2010/01/07 19:54:02] @ Quit: paxos: | ||
| [2010/01/07 19:54:13] <mqr> Thanks, I will try that now. | ||
| [2010/01/07 19:54:39] <ricky> I was looking at proc_getgroups in process.c in the ruby source code - it calls the getgroups C function, which can return EINTR | ||
| [2010/01/07 19:55:00] @ bodepd left channel #puppet () | ||
| [2010/01/07 19:57:06] * ricky goes off for a bit, thanks for looking atthis | ||
| [2010/01/07 19:57:22] <mqr> ricky... | ||
| [2010/01/07 19:58:56] @ Quit: tomas_: "Leaving..." | ||
| [2010/01/07 20:02:34] @ Quit: LancerDeuce: "Leaving" | ||
| [2010/01/07 20:05:07] @ Quit: alfism: "http://opensolaris.com/" | ||
| [2010/01/07 20:07:44] @ Quit: ezmobius: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 20:07:56] <duritong> lero: standalone runs with webrick and is single threaded | ||
| [2010/01/07 20:08:16] @ pjsz joined channel #puppet | ||
| [2010/01/07 20:08:42] <duritong> lero: if you got a certain amount of hosts contacting your master you'll have to switch over to a multi-threaded solution, like mongrel or passenger. the latter seems to be prefered nowadays | ||
| [2010/01/07 20:10:14] <lero> hmm | ||
| [2010/01/07 20:10:18] <lero> we are using mongrel right now | ||
| [2010/01/07 20:10:23] <lero> to server about ~100 machines | ||
| [2010/01/07 20:10:34] @ Quit: explody: Remote closed the connection | ||
| [2010/01/07 20:11:06] <lero> it's better with passenger? | ||
| [2010/01/07 20:11:30] <jrojas> lero: ive got passenger doing about 100+ requests a second on 24.8 for 1k+ hosts | ||
| [2010/01/07 20:12:10] <lero> what kind of machine | ||
| [2010/01/07 20:12:14] <lero> the server side | ||
| [2010/01/07 20:12:29] <jrojas> its a xen vm, 4cpus and a bunch of memory | ||
| [2010/01/07 20:12:34] <lero> hmm nie | ||
| [2010/01/07 20:12:35] <lero> nice | ||
| [2010/01/07 20:12:55] <lero> i'm running on 4 x xeon 3.2ghz and 4gb of ram | ||
| [2010/01/07 20:13:05] <lero> memory usage right now = 544MB | ||
| [2010/01/07 20:13:20] @ rdi joined channel #puppet | ||
| [2010/01/07 20:13:42] <jrojas> 6GB of ram for me, after a day of running i run out of ram right around 5am | ||
| [2010/01/07 20:13:52] <jrojas> from mem leaks and such | ||
| [2010/01/07 20:13:55] <lero> hehe | ||
| [2010/01/07 20:14:10] <jrojas> avg 5min load avg: 27.70 | ||
| [2010/01/07 20:14:16] <lero> i'll see how this 0.25.2 runs.. i rebooted the machine yesterday to install a new kernel | ||
| [2010/01/07 20:14:17] @ Quit: danielbln: "Leaving." | ||
| [2010/01/07 20:14:33] <lero> using ruby 1.8.6-383 from fc12/13, recompiled | ||
| [2010/01/07 20:14:38] <lero> on rhel5.4 | ||
| [2010/01/07 20:14:54] <lero> jrojas: hmm | ||
| [2010/01/07 20:15:17] <lero> my load is nice.. 0.06 | ||
| [2010/01/07 20:16:42] <jrojas> im also serving a lot of files | ||
| [2010/01/07 20:17:02] <lero> my config is simple :) | ||
| [2010/01/07 20:17:04] @ Quit: QMan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 20:17:07] <lero> we use cfengine too | ||
| [2010/01/07 20:17:12] <lero> for other areas | ||
| [2010/01/07 20:17:17] <jrojas> lero: yeah my manifests are far from simple. | ||
| [2010/01/07 20:17:26] <lero> cf2... they are moving to cf3 | ||
| [2010/01/07 20:17:26] @ QMan joined channel #puppet | ||
| [2010/01/07 20:17:33] <lero> but cf3 is a bitch to config | ||
| [2010/01/07 20:17:41] <lero> puppet is nice, thought it's slow :/ | ||
| [2010/01/07 20:17:55] <lero> they serve about 3k machines from one p4 cf2 server | ||
| [2010/01/07 20:17:57] <lero> hehe | ||
| [2010/01/07 20:18:22] <jrojas> it takes about 30 secs for a manifest to run on one of my servers. | ||
| [2010/01/07 20:20:36] <lero> nice :D | ||
| [2010/01/07 20:22:25] @ Quit: pjsz: | ||
| [2010/01/07 20:22:39] <rdi> Hi, all. I've been looking into using puppet and I've got a question that I can't seem to answer from the docs. Was hoping someone here could help out. | ||
| [2010/01/07 20:24:57] @ pjsz joined channel #puppet | ||
| [2010/01/07 20:25:31] @ Quit: mqr: | ||
| [2010/01/07 20:25:57] <rdi> If I have a service that depends on a list of files (via subscribe) will it be restarted when *any* file in the list changes, or when *every* file in the last changes? | ||
| [2010/01/07 20:26:06] <z00dax> any | ||
| [2010/01/07 20:26:22] <rdi> Thans, z00dax. | ||
| [2010/01/07 20:26:32] <z00dax> but this should be easy to test, dont take my word for it | ||
| [2010/01/07 20:26:41] <z00dax> specially since i have no idea how you did the setup to start with | ||
| [2010/01/07 20:26:46] <z00dax> or what your manifests look like | ||
| [2010/01/07 20:27:09] <rdi> Haven't written any yet - trying to get a sense of how that particular angle plays out | ||
| [2010/01/07 20:28:10] <rdi> I have a service that has a number of data files that it reads at startup, and I want to have it restart when the files are updated - but only after the updates are completed. Not sure if that's more than puppet can handle. | ||
| [2010/01/07 20:29:16] @ Quit: pjsz: Client Quit | ||
| [2010/01/07 20:31:26] <eric0> you might try notify from each of the files to the service rather than subscribe. | ||
| [2010/01/07 20:31:42] <z00dax> yeah, was thinking of that | ||
| [2010/01/07 20:31:55] <z00dax> it will make for, atleast, easier to understand manfiests | ||
| [2010/01/07 20:32:39] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/07 20:33:21] @ Quit: jcape1: Read error: 110 (Connection timed out) | ||
| [2010/01/07 20:33:46] <rdi> Thanks, eric0. Seems like that might still cause a restart after each file updates, but I haven't looked at a notify example yet. | ||
| [2010/01/07 20:37:34] <rdi> h'm. Just looked at the metaparameters, and maybe "before" would do what I want. I'll dig into that some more. Thanks. | ||
| [2010/01/07 20:38:37] @ Quit: taggart: Read error: 110 (Connection timed out) | ||
| [2010/01/07 20:38:45] @ bug joined channel #puppet | ||
| [2010/01/07 20:39:27] @ Quit: gaveen: Read error: 110 (Connection timed out) | ||
| [2010/01/07 20:39:54] @ taggart joined channel #puppet | ||
| [2010/01/07 20:40:47] @ Quit: bjaspan: | ||
| [2010/01/07 20:40:55] @ gaveen joined channel #puppet | ||
| [2010/01/07 20:41:27] @ Quit: bug: Client Quit | ||
| [2010/01/07 20:42:09] @ rdi left channel #puppet () | ||
| [2010/01/07 20:42:15] @ Quit: QMan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 20:43:09] @ QMan joined channel #puppet | ||
| [2010/01/07 20:44:40] @ jaredrhine joined channel #puppet | ||
| [2010/01/07 20:49:40] @ Quit: rmiller4pi81: Read error: 110 (Connection timed out) | ||
| [2010/01/07 20:52:19] @ Quit: mikerowehl: | ||
| [2010/01/07 20:52:32] @ nakano_ is now known as nakano | ||
| [2010/01/07 20:56:04] @ Quit: jes5: "Leaving." | ||
| [2010/01/07 20:56:14] @ mqr joined channel #puppet | ||
| [2010/01/07 20:56:52] @ Filbert_ joined channel #puppet | ||
| [2010/01/07 20:56:53] @ Quit: Filbert: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 20:57:40] @ Quit: somnoliento: "Ex-Chat" | ||
| [2010/01/07 20:59:43] <ricky> mqr: I'm back now, was there something wrong? | ||
| [2010/01/07 21:00:00] @ Quit: lak: | ||
| [2010/01/07 21:04:07] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/07 21:05:48] @ Quit: jaredrhine: Read error: 110 (Connection timed out) | ||
| [2010/01/07 21:06:38] <mqr> Only with my brain. | ||
| [2010/01/07 21:08:43] <mqr> ricky: I'm going to work up a modification to your patch for #3013, and try to get a 0.25.3 out in the next day or so. | ||
| [2010/01/07 21:08:44] <gepetto> mqr: ricky: #3013 is http://projects.reductivelabs.com/issues/show/3013 "Puppet - Bug #3013: util.rb:execute broken on Ruby <1.8.3 - ReductiveLabs.com" | ||
| [2010/01/07 21:09:04] @ explody joined channel #puppet | ||
| [2010/01/07 21:09:13] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 21:10:18] <ricky> Sounds good | ||
| [2010/01/07 21:10:46] <explody> is it possible to extend default facts from facter? I don't mean adding new ones, but for example adding support for missing OSes in existing facts (e.g. HP-UX doesn't get netmask and it doesn't properly read ip addresses) | ||
| [2010/01/07 21:11:29] <explody> aside from monkey patching the original source, I mean | ||
| [2010/01/07 21:12:33] @ bjaspan joined channel #puppet | ||
| [2010/01/07 21:13:56] <taggart> explody: why not patch the original source and submit your patches to the maintainers? I'm sure they would like them | ||
| [2010/01/07 21:14:14] <explody> that's fine too | ||
| [2010/01/07 21:14:27] <taggart> explody: the puppet repo I work on additional facts | ||
| [2010/01/07 21:14:51] @ QMan joined channel #puppet | ||
| [2010/01/07 21:15:02] <explody> I was just thinking if there was a modular way to extend it, I could add support for things and probably redistribute without messing with the upstream | ||
| [2010/01/07 21:15:07] <taggart> explody: I don't know about modifying default facts in that manner, but as a hack you could just make your own with another name (until you get your fixes in upstream) | ||
| [2010/01/07 21:15:25] <explody> yeah I've done that for several other facts | ||
| [2010/01/07 21:19:55] @ mqr_ joined channel #puppet | ||
| [2010/01/07 21:20:38] @ Quit: mqr: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 21:20:46] <explody> oh wow, there is no explicit support for hp-ux at all in facter. heh, no wonder the fact list is a bit small :P | ||
| [2010/01/07 21:22:57] @ Quit: mqr_: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 21:23:06] @ mqr joined channel #puppet | ||
| [2010/01/07 21:27:14] @ scylla joined channel #puppet | ||
| [2010/01/07 21:27:53] <taggart> explody: yeah that's because nobody uses it ;P | ||
| [2010/01/07 21:28:27] <explody> god I wish | ||
| [2010/01/07 21:29:07] <explody> I have I think 3 PA-RISC and 4 IA64s | ||
| [2010/01/07 21:31:25] @ Quit: ashp: Remote closed the connection | ||
| [2010/01/07 21:32:09] <explody> for those 7 boxes, I have to put in as much maintenance effort as I do for 200+ linux and another 50 Sol9/10 | ||
| [2010/01/07 21:32:59] <jb> i have about 50 aix boxes :/ | ||
| [2010/01/07 21:33:18] <explody> 9 here | ||
| [2010/01/07 21:33:28] <jb> i despise IBM. | ||
| [2010/01/07 21:33:40] <explody> yeah they drive me nuts | ||
| [2010/01/07 21:34:19] <jb> everyting is ass backwards | ||
| [2010/01/07 21:34:22] <jb> and their website is a monster | ||
| [2010/01/07 21:34:50] <explody> I wouldn't mind as much if we had like 20+ of each, then it would feel like we have an equitably mixed network, but have like 5 of some random platform make management suck. Lots of overhead for few machines | ||
| [2010/01/07 21:35:10] <jb> yeah that sucks.. | ||
| [2010/01/07 21:35:13] <jb> all linux/aix here | ||
| [2010/01/07 21:35:17] <jb> well, and windows :) | ||
| [2010/01/07 21:36:24] <explody> we're about as heterogeneous as it gets: RHEL3/4/5,SLES9/10,Ubuntu,Debian,AIX5.2/5.3/6,HP-UX 11.11/11.23/11.31 and Sol 8/9/10 | ||
| [2010/01/07 21:36:40] <explody> and windows blah/blah/blah/I don't do windows | ||
| [2010/01/07 21:37:33] <jb> we are all EL5 and AIX 5.3/6.1 on the nix side.. | ||
| [2010/01/07 21:37:50] <jb> i sqaused SLES9 bs | ||
| [2010/01/07 21:37:54] <jb> squashed | ||
| [2010/01/07 21:38:20] <explody> I'm about to | ||
| [2010/01/07 21:38:42] <explody> I wish I could squash everything that uses RPM :P | ||
| [2010/01/07 21:39:07] <jb> :) | ||
| [2010/01/07 21:39:11] <jb> yum :P | ||
| [2010/01/07 21:40:25] <explody> it certainly helps, but RH and SLES both still get on my nerves | ||
| [2010/01/07 21:41:47] @ jaredrhine joined channel #puppet | ||
| [2010/01/07 21:42:07] @ mqr_ joined channel #puppet | ||
| [2010/01/07 21:42:09] @ Quit: mqr: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 21:42:34] <nevyn> I like debian use it all the time my laptop runs debian | ||
| [2010/01/07 21:42:47] <nevyn> everything at work is suse/redhat and I can see why | ||
| [2010/01/07 21:43:21] <nevyn> fundamentally the debian approach to maintainence breaks apps. | ||
| [2010/01/07 21:43:21] @ ashp joined channel #puppet | ||
| [2010/01/07 21:43:36] <ohadlevy> jb: do you use puppet with aix? | ||
| [2010/01/07 21:44:02] <nevyn> explody: otoh.. depenancy on a file is a stupid feature. | ||
| [2010/01/07 21:44:34] <explody> nevyn: That's their approach though, the packaging system doesn't suck. RPM has been awful since the day it was born | ||
| [2010/01/07 21:45:03] <explody> TBH, they're both pretty broken, but apt/dpkg bugs me a lot less | ||
| [2010/01/07 21:45:08] <nevyn> explody: well deb is an incremental improvement on rpm. they're functionally equiv | ||
| [2010/01/07 21:45:24] @ littleidea joined channel #puppet | ||
| [2010/01/07 21:45:45] <nevyn> deb has a richer set of features (suggests, requires etc) | ||
| [2010/01/07 21:46:09] <nevyn> and less braindamaged features (the package system cannot depend on a file only on packages) | ||
| [2010/01/07 21:46:39] <nevyn> but given yum and a rhn subscription things pretty much work | ||
| [2010/01/07 21:46:56] <taggart> explody: ia64 and hppa run Debian :) | ||
| [2010/01/07 21:46:56] <jb> ohadlevy: i'm going to start, yeah | ||
| [2010/01/07 21:47:08] <jb> ohadlevy: the service/package provider patches are available.. | ||
| [2010/01/07 21:47:12] <ohadlevy> jb: I was wondering if foreman works when those too ;) | ||
| [2010/01/07 21:47:21] <explody> nevyn: agreed | ||
| [2010/01/07 21:47:25] <nevyn> ia64 is an awesome platform if you care about raw computrons | ||
| [2010/01/07 21:47:31] <jb> ohadlevy: andrew, the guy who has the foreman ticket about AIX has done a lot of work | ||
| [2010/01/07 21:47:34] <nevyn> and not about power | ||
| [2010/01/07 21:47:41] <kjetilho> Debian wouldn't be so bad if it hadn't been for Debian Policy | ||
| [2010/01/07 21:47:48] <ohadlevy> yeah, but he had some issues with the facts importers | ||
| [2010/01/07 21:47:56] <jb> yeah i'll know soon enough | ||
| [2010/01/07 21:47:57] <kjetilho> unfortunately Debian seems to attract nutcases | ||
| [2010/01/07 21:48:04] <explody> taggart: don't tease me, I'll cry | ||
| [2010/01/07 21:48:04] <jb> ohadlevy: i did have some questions/feature suggestions to bug you about | ||
| [2010/01/07 21:48:06] <nevyn> unfortunatly for most workloads today computrons/watt is the critical metric in decision making | ||
| [2010/01/07 21:48:09] <taggart> kjetilho: what's wrong with policy? | ||
| [2010/01/07 21:48:33] <nevyn> and amd64/ia32e have pretty much everything licked there.. | ||
| [2010/01/07 21:48:46] <taggart> explody: also for a while Debian/hppa had support for running hpux/hppa binaries | ||
| [2010/01/07 21:48:48] <nevyn> Power might be in with a chance. | ||
| [2010/01/07 21:48:48] <kjetilho> taggart: e.g., hostname is not FQDN | ||
| [2010/01/07 21:49:33] <kjetilho> taggart: and no real counterpart for RHEL's chkconfig | ||
| [2010/01/07 21:49:35] <explody> hostname isn't supposed to be fqdn, it's the hostname | ||
| [2010/01/07 21:49:37] <taggart> explody: there was a guy running some hpux license server on an old hppa 712 workstation running debian :) | ||
| [2010/01/07 21:49:46] <kjetilho> explody: that way lies madness | ||
| [2010/01/07 21:49:54] <explody> taggart: awesome | ||
| [2010/01/07 21:49:56] <taggart> kjetilho: update-rc.d? | ||
| [2010/01/07 21:50:17] <explody> kjetilho: uh no, that way lies the difference between a hostname and a domainname | ||
| [2010/01/07 21:50:17] <taggart> explody: I am guessing you have some similar legacy reason for needing to support it | ||
| [2010/01/07 21:50:28] <kjetilho> taggart: nope, won't do the trick. since daemons are linked up by default, you need to find a magic config file which will remove the disabling | ||
| [2010/01/07 21:50:59] <explody> taggart: very much so, we have quite a few HP-UX customers. Most are moving away from PA-RISC to IA64, which doesn't help me at all | ||
| [2010/01/07 21:51:01] <kjetilho> explody: what's a domainname? it's certainly not the stuff after the first period | ||
| [2010/01/07 21:51:37] <kjetilho> explody: that's the thing about Debian. it works for small installations, like perhaps a couple hundred servers | ||
| [2010/01/07 21:51:47] <kjetilho> for larger installations, the policies just break down | ||
| [2010/01/07 21:51:53] <taggart> kjetilho: oh I see what you are saying. yes this is a fundamental philosopy difference: in RedHat everything is installed by default and disabled by default, in Debian nothing is installed by default and if you install it then it's enabled by default | ||
| [2010/01/07 21:52:04] <explody> taggart: yeah, I hate that | ||
| [2010/01/07 21:52:24] @ GGD joined channel #puppet | ||
| [2010/01/07 21:52:25] <taggart> kjetilho: I'd still like to see something like what you're talking about so the sysadmin could set their own policy | ||
| [2010/01/07 21:52:35] <kjetilho> taggart: no, it's just halfway enabled. you often need to find the correct file in /etc/default/foo and comment out the DISABLE statement | ||
| [2010/01/07 21:52:41] <explody> kjetilho: so you're concerned about naming collisions in separate domains? i.e. host1.a.domain.com and host1.b.domain.com? | ||
| [2010/01/07 21:52:52] <kjetilho> explody: yes, happens all the time | ||
| [2010/01/07 21:53:22] <explody> um, yeah if you don't use unique hostnames. I totally understand what you're talking about but it's not that hard to avoid | ||
| [2010/01/07 21:53:26] <taggart> kjetilho: well it depends on the package, I would say it's "enabled in a safe state" which for some services is disabled :P | ||
| [2010/01/07 21:53:36] <kjetilho> if you use unique hostnames, why use subdomains? | ||
| [2010/01/07 21:54:07] <kjetilho> taggart: yeah. but it means you need special procedures for each service to enable them | ||
| [2010/01/07 21:54:07] <explody> any number of reasons, logical divisions, geographic or network segments | ||
| [2010/01/07 21:54:15] <kjetilho> taggart: in RHEL it's simple. use chkconfig | ||
| [2010/01/07 21:54:47] <taggart> I worked in a company who's infrastructure predated domainnames and a lot of the legacy hostnames had all sorts of info in them that wound up being redundant with the domainname | ||
| [2010/01/07 21:54:49] <kjetilho> explody: sure, but if those divisions need to cooperate to avoid collisions, what good is it really to separate them? | ||
| [2010/01/07 21:55:40] <taggart> kjetilho: yeah, it would be a nice thing to standardize | ||
| [2010/01/07 21:55:42] * kjetilho was convinced FQDN was the only sane thing after managing Ultirx NFS servers in 1991 | ||
| [2010/01/07 21:56:02] <kjetilho> *Ultrix | ||
| [2010/01/07 21:57:18] <explody> kjetilho: I see your point, but again it's just so easy to avoid. i.e. use non-unique names, and use the FQDN...there's no reason that `hostname` needs to return the FQDN though, because any way you cut it the FQDN is not a machine's hostname | ||
| [2010/01/07 21:57:33] <kjetilho> yes it is | ||
| [2010/01/07 21:57:50] <kjetilho> my name is not Kjetil. it is Kjetil Torgrim Homme | ||
| [2010/01/07 21:58:05] <kjetilho> why should computers be different? | ||
| [2010/01/07 21:59:03] <kjetilho> it's just painful to make the nodename non-FQDN. just establishing the FQDN from a short name is error prone | ||
| [2010/01/07 21:59:24] <kjetilho> it will most often require a lookup on a remote host (the DNS server) | ||
| [2010/01/07 21:59:41] <kjetilho> a basic fact like the hosts name should *not* require external dependencies | ||
| [2010/01/07 22:00:07] <explody> so you don't maintain your hosts file either? | ||
| [2010/01/07 22:00:27] <kjetilho> no, the hosts file only contains localhost | ||
| [2010/01/07 22:01:24] <explody> whatever floats yer boat | ||
| [2010/01/07 22:01:34] <kjetilho> yeah, and it sure ain't Debian | ||
| [2010/01/07 22:01:58] <taggart> hehe | ||
| [2010/01/07 22:02:06] <explody> apparently not :P | ||
| [2010/01/07 22:02:06] @ Quit: brothers: | ||
| [2010/01/07 22:02:25] <taggart> non-Debian is more painful | ||
| [2010/01/07 22:02:34] <kjetilho> but I think it's partly due to my early exposure to SunOS. us oldtimers seem to prefer RHEL over Debian/Ubuntu | ||
| [2010/01/07 22:02:36] <explody> btw, you might read the manpage for "hostname" on an RH5 box | ||
| [2010/01/07 22:03:09] <jb> ohadlevy: is there any thing in place to notify when failure report comes in for a node? | ||
| [2010/01/07 22:03:12] <taggart> kjetilho: can't teach an old dog new tricks? :) | ||
| [2010/01/07 22:03:20] <kjetilho> explody: what about it? | ||
| [2010/01/07 22:03:28] <jb> ohadlevy: i think it would be useful in foreman to be notified if a node has a failure.. by email and/or by SNMP traps maybe? | ||
| [2010/01/07 22:03:43] <kjetilho> taggart: SMF, ZFS, Zones, Comstar -- I'd think we can learn new tricks | ||
| [2010/01/07 22:03:55] <ohadlevy> jb: currently there is a a rake tasks that summirze all puppet activity and send it via an email | ||
| [2010/01/07 22:04:15] <ohadlevy> and you can restrict it to an env or a fact base filtering (e.g. that you get only the hosts you are interested in) | ||
| [2010/01/07 22:04:33] <ohadlevy> but its not triggered via a report... hmm.. that could be easy to add | ||
| [2010/01/07 22:04:39] <jb> ok thats nice | ||
| [2010/01/07 22:04:48] <jb> yeah, I think that would be useful | ||
| [2010/01/07 22:04:57] <ohadlevy> jb: do you use the git version? | ||
| [2010/01/07 22:04:59] <jb> puppet failure -> generate SNMP trap to monitoring system -> alert admins | ||
| [2010/01/07 22:05:00] <jb> yep | ||
| [2010/01/07 22:05:10] <explody> kjetilho: I'm just pointing out that although redhat likes to put the FQDN in the hostname, it's not like they've rewritten what hostname/domainname/fqdn *is*, it's the same manpage as most any other linux | ||
| [2010/01/07 22:05:11] @ Demosthenes joined channel #puppet | ||
| [2010/01/07 22:05:29] <ohadlevy> try typing : rake -D reports:summarize | ||
| [2010/01/07 22:05:31] <explody> ergo, you can put the fqdn in hostname on a debian box too if you want | ||
| [2010/01/07 22:05:42] <kjetilho> explody: sure. you're free to use either short or FQDN on RHEL | ||
| [2010/01/07 22:05:49] <jb> ohadlevy: ahh | ||
| [2010/01/07 22:05:56] <kjetilho> explody: but on Debian, *only* short hostname is supported, by policy | ||
| [2010/01/07 22:06:21] <explody> what does supported mean? | ||
| [2010/01/07 22:06:46] <explody> oh oh I see | ||
| [2010/01/07 22:06:54] <explody> you mean literally | ||
| [2010/01/07 22:07:14] <jb> ohadlevy: ok, yeah, the triggered per report might be a nice feature, if it's not to difficult | ||
| [2010/01/07 22:07:16] <kjetilho> well, I guess no one is actually getting support for Debian. so the point is moot | ||
| [2010/01/07 22:07:30] <ohadlevy> jb: add it, I'll do it somewhere in the next few days | ||
| [2010/01/07 22:07:36] <jb> ohadlevy: sounds good. | ||
| [2010/01/07 22:07:40] <explody> I've never needed it ;) | ||
| [2010/01/07 22:07:45] <kjetilho> hah. | ||
| [2010/01/07 22:07:52] <jb> ohadlevy: also, automatic deletion of reports would be nice.. | ||
| [2010/01/07 22:08:04] <jb> ohadlevy: i have like 13,000 in two days :) | ||
| [2010/01/07 22:08:08] <explody> have on RHEL and SLES, though that was due to commercial 3rd party stuff, not the OS | ||
| [2010/01/07 22:08:11] @ Quit: Boltsky: | ||
| [2010/01/07 22:08:13] @ Quit: littleidea: | ||
| [2010/01/07 22:08:30] <ohadlevy> jb: you really dont read the docs dont you? ;) | ||
| [2010/01/07 22:08:39] @ jusfreeman joined channel #puppet | ||
| [2010/01/07 22:08:39] <jb> ohadlevy: i do, i swear :) | ||
| [2010/01/07 22:08:42] <ohadlevy> rake -D reports:expire | ||
| [2010/01/07 22:08:58] @ tomas_ joined channel #puppet | ||
| [2010/01/07 22:09:07] <jb> ohadlevy: thanks! | ||
| [2010/01/07 22:09:22] <jb> so what graphs are you working on? | ||
| [2010/01/07 22:09:24] <ohadlevy> its even in the puppet foreman modules | ||
| [2010/01/07 22:09:36] <explody> enough OT for tonight, I'm out cheers | ||
| [2010/01/07 22:09:44] <ohadlevy> I've added hosts graphs - e.g. something like: | ||
| [2010/01/07 22:10:22] <ohadlevy> http://chart.apis.google.com/chart?chbh=25&chf=bg,s,E6DFCF&chxt=x,y&chs=800x200&cht=lc&chd=s:LKKKLKJJLLKQLQMNKLMKLLLMaNPOLNKKLLMKKOLKKKKKLKKK,nnmoopnnqpo4xzu6rrpoqvro6r9xt5qpqoqpnrpnppnqppno&chtt=Runtime&chxl=0:%7C07+Jan+03:15%7C07+Jan+15:00%7C08+Jan+02:45%7C1:%7C0%7C43.56&chdl=Config+Retrival%7CTotal&chg=10,50,1,5&chm=o,0066FF,1,-1,6%7Co,80C65A,0,-1,6&chco=0077CC,7700CC | ||
| [2010/01/07 22:10:23] <taggart> kjetilho: I also started on Ultrix (1992) but moved on to Slackware (1994) and Redhat (1996) and then Debian (1997) and never looked back. So not all old timers are still on RH :) | ||
| [2010/01/07 22:10:37] <ohadlevy> and http://chart.apis.google.com/chart?chbh=5,0,3&chf=bg,s,E6DFCF&chxt=x,y&chs=800x200&cht=lc&chd=s:AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9,AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9A,AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA,AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA,AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA&chtt=Resources&chxl=0:%7C07+Jan+03:20%7C07+Jan+15:04%7C08+Jan+02:49%7C1:%7C0%7C1&chdl=applied%7C | ||
| [2010/01/07 22:10:44] <jb> ohadlevy: oh nice, like dashboard :) | ||
| [2010/01/07 22:11:10] <kjetilho> taggart: I guess you are the exception to the rule, then ;-) | ||
| [2010/01/07 22:11:14] <ohadlevy> jb: well, too many people were complaining about the rrd graphs, so it was too easy just to add it, as its all in the db anyway | ||
| [2010/01/07 22:11:23] <explody> oh yeah, RH in '95 here, Debian since maybe '98 | ||
| [2010/01/07 22:11:40] <taggart> kjetilho: hehe | ||
| [2010/01/07 22:11:40] <explody> <= not new school either | ||
| [2010/01/07 22:11:47] <jb> ohadlevy: good idea | ||
| [2010/01/07 22:12:20] <taggart> explody: I'll admit to doing hpux from 96-98, but only because I was developing it :P | ||
| [2010/01/07 22:12:37] <kjetilho> hehe | ||
| [2010/01/07 22:12:46] <explody> haha | ||
| [2010/01/07 22:13:25] <kjetilho> I did HP-UX sysadmin in '93. not my most pleasant job experience | ||
| [2010/01/07 22:13:27] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:13:33] <ohadlevy> jb: I'll probably commit some of this stuff later today, will be happy for feedback | ||
| [2010/01/07 22:13:36] <kjetilho> it's a lot better these days | ||
| [2010/01/07 22:13:48] @ QMan joined channel #puppet | ||
| [2010/01/07 22:13:50] @ brothers joined channel #puppet | ||
| [2010/01/07 22:13:53] <ohadlevy> jb: the only problem unlike the rrd's, is that we dont store the reports for ever | ||
| [2010/01/07 22:14:04] @ Quit: mqr_: Read error: 110 (Connection timed out) | ||
| [2010/01/07 22:14:29] <jb> right | ||
| [2010/01/07 22:14:30] <jb> hm | ||
| [2010/01/07 22:14:36] @ littleidea joined channel #puppet | ||
| [2010/01/07 22:14:41] @ Quit: littleidea: Client Quit | ||
| [2010/01/07 22:14:43] <taggart> jb: I also would like error reports | ||
| [2010/01/07 22:15:03] <taggart> jb: right now I only have something that tells me when hosts last checked in | ||
| [2010/01/07 22:15:32] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:15:49] @ littleidea joined channel #puppet | ||
| [2010/01/07 22:16:55] @ Quit: JM: Client Quit | ||
| [2010/01/07 22:17:04] <ohadlevy> taggart: you could always use foreman :) | ||
| [2010/01/07 22:17:09] @ Bass10 joined channel #puppet | ||
| [2010/01/07 22:17:34] <taggart> ohadlevy: is that the rake thing you mention? | ||
| [2010/01/07 22:18:00] <ohadlevy> taggart: the rake tasks are part of it - http://theforeman.org | ||
| [2010/01/07 22:18:38] @ Quit: Bass10: SendQ exceeded | ||
| [2010/01/07 22:18:44] @ QMan joined channel #puppet | ||
| [2010/01/07 22:19:01] <ohadlevy> taggart: you can you use it only for puppet reportings if you like | ||
| [2010/01/07 22:19:05] <taggart> ohadlevy: it should have been named overseer and had silver glasses and a shotgun... | ||
| [2010/01/07 22:19:14] @ Bass10 joined channel #puppet | ||
| [2010/01/07 22:19:22] <ohadlevy> :) | ||
| [2010/01/07 22:19:26] <taggart> oooooo systems inventory | ||
| [2010/01/07 22:19:40] <taggart> I've been wanting that and haven't found anything good to do it | ||
| [2010/01/07 22:19:57] <taggart> I was contemplating writing my own | ||
| [2010/01/07 22:20:08] @ Boltsky joined channel #puppet | ||
| [2010/01/07 22:20:20] <taggart> I don't know if facts would be enough, I was going to do some parsing of hwinfo/dmidecode | ||
| [2010/01/07 22:20:40] <taggart> but I guess facts could be written for all that | ||
| [2010/01/07 22:20:59] <ohadlevy> taggart: facter does parse dmidecode | ||
| [2010/01/07 22:21:59] @ Quit: Bass10: Client Quit | ||
| [2010/01/07 22:22:26] <taggart> ohadlevy: oh cool | ||
| [2010/01/07 22:22:57] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:23:50] @ QMan joined channel #puppet | ||
| [2010/01/07 22:26:50] <ohadlevy> jb: are you using master branch or edge? | ||
| [2010/01/07 22:27:01] <jb> taggart: i can vouch for foreman :) | ||
| [2010/01/07 22:27:12] <jb> hm | ||
| [2010/01/07 22:27:24] @ Quit: Boltsky: Read error: 60 (Operation timed out) | ||
| [2010/01/07 22:27:58] <jb> good question | ||
| [2010/01/07 22:28:18] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:28:19] <ohadlevy> probably master ;) | ||
| [2010/01/07 22:28:21] <ohadlevy> git branch will tell you | ||
| [2010/01/07 22:28:41] <jb> 0.1-3 | ||
| [2010/01/07 22:28:44] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:28:45] <jb> master | ||
| [2010/01/07 22:28:56] @ QMan joined channel #puppet | ||
| [2010/01/07 22:29:06] <ohadlevy> if you want, the new stuff usually goes to edge first | ||
| [2010/01/07 22:29:45] <jb> how do I go about gettin edge? | ||
| [2010/01/07 22:30:02] <ohadlevy> git checkout -t origin/edge -b edge | ||
| [2010/01/07 22:30:12] <ohadlevy> this will create a local edge branch | ||
| [2010/01/07 22:30:20] <ohadlevy> you might need to do git pull first | ||
| [2010/01/07 22:30:33] <jb> ok, I'll check it out | ||
| [2010/01/07 22:30:58] @ Bass10 joined channel #puppet | ||
| [2010/01/07 22:31:35] @ brothers joined channel #puppet | ||
| [2010/01/07 22:39:36] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:40:28] @ Quit: gaveen: "Leaving" | ||
| [2010/01/07 22:45:15] @ lak joined channel #puppet | ||
| [2010/01/07 22:45:31] @ GGD_ joined channel #puppet | ||
| [2010/01/07 22:46:43] @ Quit: lak: Client Quit | ||
| [2010/01/07 22:48:11] @ brothers joined channel #puppet | ||
| [2010/01/07 22:50:15] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:50:56] @ brothers joined channel #puppet | ||
| [2010/01/07 22:51:04] @ mqr joined channel #puppet | ||
| [2010/01/07 22:52:41] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 22:54:09] @ nevyn_ joined channel #puppet | ||
| [2010/01/07 22:58:41] @ QMan joined channel #puppet | ||
| [2010/01/07 23:01:51] @ crdant joined channel #puppet | ||
| [2010/01/07 23:01:52] @ Quit: QMan: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 23:02:40] @ Quit: GGD: Read error: 110 (Connection timed out) | ||
| [2010/01/07 23:04:04] @ QMan joined channel #puppet | ||
| [2010/01/07 23:05:52] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 23:05:54] @ Quit: nevyn: Read error: 101 (Network is unreachable) | ||
| [2010/01/07 23:07:46] @ lak joined channel #puppet | ||
| [2010/01/07 23:09:46] @ brothers joined channel #puppet | ||
| [2010/01/07 23:12:46] @ jaredrhine1 joined channel #puppet | ||
| [2010/01/07 23:15:36] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 23:16:37] @ brothers joined channel #puppet | ||
| [2010/01/07 23:17:19] @ Quit: jaredrhine: Read error: 110 (Connection timed out) | ||
| [2010/01/07 23:18:01] @ Quit: Djelibeybi: "Leaving" | ||
| [2010/01/07 23:31:59] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 23:32:56] @ brothers joined channel #puppet | ||
| [2010/01/07 23:34:55] @ Quit: QMan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/07 23:35:52] @ QMan joined channel #puppet | ||
| [2010/01/07 23:36:13] @ Quit: lak: | ||
| [2010/01/07 23:37:43] @ Boltsky joined channel #puppet | ||
| [2010/01/07 23:39:55] @ Quit: lero: "Leaving..." | ||
| [2010/01/07 23:42:24] @ nevyn_ is now known as nevyn | ||
| [2010/01/07 23:43:35] @ bug joined channel #puppet | ||
| [2010/01/07 23:51:30] @ Quit: brothers: Read error: 104 (Connection reset by peer) | ||
| [2010/01/07 23:52:31] @ brothers joined channel #puppet | ||
| [2010/01/07 23:55:10] @ jes5 joined channel #puppet | ||
| [2010/01/07 23:55:38] @ jcape joined channel #puppet | ||
| [2010/01/07 23:55:55] @ Quit: lazzurs: Remote closed the connection | ||
| [2010/01/07 23:56:00] @ lazzurs joined channel #puppet | ||
| [2010/01/07 23:56:33] @ Quit: astinus: Remote closed the connection | ||
| [2010/01/07 23:56:41] @ astinus joined channel #puppet |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!