| [2011/12/02 00:09:51] @ Log started by gepetto | ||
| [2011/12/02 00:09:51] @ Quit: Spads: Ping timeout: 252 seconds | ||
| [2011/12/02 00:10:04] @ Quit: kolla_: Remote host closed the connection | ||
| [2011/12/02 00:10:49] <joe-mac> can someone explain to me the significance of the 'instances' method in a type? shouldn't it be that if an objkect exists in instances, it is considered to already be present? | ||
| [2011/12/02 00:13:18] <whack> joe-mac: 'instances' I think is to query for all existing instances of that type on the system for stuff like 'puppet resource' and maybe some other stuff | ||
| [2011/12/02 00:14:49] @ Ramonster joined channel #puppet | ||
| [2011/12/02 00:15:32] <joe-mac> indeed whack | ||
| [2011/12/02 00:15:41] <joe-mac> I have a problem with a type though- it keeps re writing on every run | ||
| [2011/12/02 00:17:26] @ verwilst joined channel #puppet | ||
| [2011/12/02 00:18:11] <whack> something you wrote? or a standard one? | ||
| [2011/12/02 00:18:44] @ Quit: lak: Quit: Leaving. | ||
| [2011/12/02 00:19:11] @ lak joined channel #puppet | ||
| [2011/12/02 00:19:12] @ Quit: lak: Changing host | ||
| [2011/12/02 00:19:12] @ lak joined channel #puppet | ||
| [2011/12/02 00:19:45] <joe-mac> it's the sshkey type but modified | ||
| [2011/12/02 00:19:54] <joe-mac> to support more than one keytype per host and to support new ecdsa keys | ||
| [2011/12/02 00:20:05] <joe-mac> it's working except seems to be re writing every dam run | ||
| [2011/12/02 00:21:08] <whack> ahh, I forget the exact details lemme see my type code | ||
| [2011/12/02 00:21:55] <whack> the exists? method might be poorly implemented, maybe? | ||
| [2011/12/02 00:22:12] <whack> if you 'ensure => present' does it say 'ensuring present' on all runs? or does it silentlyrewrite the file every time? | ||
| [2011/12/02 00:22:24] <joe-mac> it's the opposite of silent | ||
| [2011/12/02 00:23:43] <joe-mac> it's like you said ensure created on every run | ||
| [2011/12/02 00:24:32] <whack> probably 'exists?' in your type isn't doing it right then? | ||
| [2011/12/02 00:24:42] <whack> My puppet type internals is a bit rusty though | ||
| [2011/12/02 00:24:48] <joe-mac> I mean it's just an addition to the existing one, let me check | ||
| [2011/12/02 00:25:22] @ raphdg joined channel #puppet | ||
| [2011/12/02 00:25:54] @ Kazuhiro joined channel #puppet | ||
| [2011/12/02 00:25:59] @ Quit: lak: Quit: Leaving. | ||
| [2011/12/02 00:26:46] <masterzen> joe-mac: is it a parsed file type? | ||
| [2011/12/02 00:27:00] @ fronlius joined channel #puppet | ||
| [2011/12/02 00:27:13] <joe-mac> yes | ||
| [2011/12/02 00:27:16] <joe-mac> masterzen, it is | ||
| [2011/12/02 00:27:32] <joe-mac> sweet- I never knew about the 'whit' type desc "The smallest possible resource type, for when you need a resource and naught else." | ||
| [2011/12/02 00:27:49] <joe-mac> that's pretty sweet, you could do something like an if else and just use the resource to notify soemthing | ||
| [2011/12/02 00:27:53] @ Quit: sdog1: Quit: Leaving. | ||
| [2011/12/02 00:28:15] <joe-mac> hm actually that might not work if there is nothing that can change | ||
| [2011/12/02 00:29:03] <masterzen> joe-mac: can you pastie us the code or diff against the original? | ||
| [2011/12/02 00:29:07] <joe-mac> masterzen: do you know what I am doing wrong off the top of your head, is it what whack said? | ||
| [2011/12/02 00:29:16] <joe-mac> I'll pastie the whole thing hold on | ||
| [2011/12/02 00:29:34] <masterzen> joe-mac: no, I don't have any clue yet... It might be that insync? broke or sth... | ||
| [2011/12/02 00:29:49] @ Quit: dissipate_: Ping timeout: 240 seconds | ||
| [2011/12/02 00:30:14] <masterzen> joe-mac: I believe the whit resource is removed from the catalog at some point during compilation, so you can't use that to order/event | ||
| [2011/12/02 00:30:57] <joe-mac> http://www.pastie.org/2953813 | ||
| [2011/12/02 00:31:09] <joe-mac> that's the type, do you need the provider too | ||
| [2011/12/02 00:31:32] @ Kazuhiro left channel #puppet () | ||
| [2011/12/02 00:32:33] <joe-mac> where does the default target come from? I can't figure that one out | ||
| [2011/12/02 00:34:36] @ octane- joined channel #puppet | ||
| [2011/12/02 00:38:07] <joe-mac> the provider has changes too | ||
| [2011/12/02 00:38:13] <joe-mac> I will paste it casue I think you need to see it too | ||
| [2011/12/02 00:38:17] <masterzen> joe-mac: what version is this coming from? I'm surprised about the self.instances in the type. this usually is in the provider | ||
| [2011/12/02 00:38:36] <joe-mac> I put it there masterzen- I can shift it to the provider if need be | ||
| [2011/12/02 00:38:45] <joe-mac> this is my fork of the regular type | ||
| [2011/12/02 00:38:48] <joe-mac> this isn't the PL one | ||
| [2011/12/02 00:38:54] @ ezryx joined channel #puppet | ||
| [2011/12/02 00:39:02] <joe-mac> I need two features it doesn't have so I am overriding it | ||
| [2011/12/02 00:39:42] <masterzen> joe-mac: I see | ||
| [2011/12/02 00:40:19] <masterzen> joe-mac: my understanding is that the self.instances is a provider thing. | ||
| [2011/12/02 00:40:43] <joe-mac> ok- this is the provider http://www.pastie.org/2953843 | ||
| [2011/12/02 00:40:48] <joe-mac> so I should move instances tehre | ||
| [2011/12/02 00:40:52] <joe-mac> let me do that and see if it magically fixes it | ||
| [2011/12/02 00:44:40] <joe-mac> masterzen: after I did that I get Could not evaluate: No ability to determine if sshkey exists | ||
| [2011/12/02 00:44:56] <joe-mac> maybe my instances method is borked | ||
| [2011/12/02 00:45:29] <joe-mac> I guess I could dump out instances to a file? not so sure what the best debug methodology is in this case | ||
| [2011/12/02 00:46:00] @ sdog joined channel #puppet | ||
| [2011/12/02 00:46:35] <joe-mac> yea I bet that method has a logic erorr in it or something | ||
| [2011/12/02 00:46:50] @ trickyf joined channel #puppet | ||
| [2011/12/02 00:47:16] @ Quit: sdog: Changing host | ||
| [2011/12/02 00:47:16] @ sdog joined channel #puppet | ||
| [2011/12/02 00:51:41] @ Quit: bmahe: Quit: Leaving | ||
| [2011/12/02 00:53:40] <joe-mac> scratch that, it's like that method never even gets called | ||
| [2011/12/02 00:54:27] @ Razique joined channel #puppet | ||
| [2011/12/02 00:55:00] <joe-mac> yea it's just not called, wtf | ||
| [2011/12/02 00:55:27] @ 64MABAUR8 is now known as Kubicek | ||
| [2011/12/02 00:56:45] <joe-mac> hm, self.instances is in the type for yumrepo | ||
| [2011/12/02 00:57:05] <Dominic> there is also a method called "prefetch" that relates to instances, not sure how | ||
| [2011/12/02 00:57:26] <Dominic> ah, the package provider calls instances from prefetch | ||
| [2011/12/02 00:57:33] <joe-mac> yea prefetch supposed to be called | ||
| [2011/12/02 00:57:37] <joe-mac> on all resources | ||
| [2011/12/02 00:57:56] <joe-mac> prefetch calls the instances method | ||
| [2011/12/02 00:58:06] <joe-mac> so maybe prefetch on parsedfiles isn't a "normal" case | ||
| [2011/12/02 00:59:50] @ Quit: sdog: Remote host closed the connection | ||
| [2011/12/02 01:00:22] @ mogitaff joined channel #puppet | ||
| [2011/12/02 01:01:19] <whack> dark magic. | ||
| [2011/12/02 01:07:27] <masterzen> joe-mac: parsedfile are different than any other providers/types | ||
| [2011/12/02 01:07:59] @ lkoranda joined channel #puppet | ||
| [2011/12/02 01:08:43] <masterzen> joe-mac: you shouldn't need a self.instances (which is usually called only with ralsh), because the parsedfile provider does it for you with the result of parsing the file | ||
| [2011/12/02 01:09:41] <masterzen> joe-mac: what were you trying to achieve with the self.instances? | ||
| [2011/12/02 01:10:02] <joe-mac> masterzen: well, I had to change the provider to the type name wasn't married to the host name | ||
| [2011/12/02 01:11:21] <masterzen> joe-mac: oh, because you can have more than one host but different key type, is that right? | ||
| [2011/12/02 01:11:27] <joe-mac> indeed | ||
| [2011/12/02 01:11:43] @ EL84 joined channel #puppet | ||
| [2011/12/02 01:12:05] <joe-mac> I said that all weird- "to the type" shouldn't be in that sentence should so "so that the" | ||
| [2011/12/02 01:12:07] <masterzen> joe-mac: you better change the provider to reflect this | ||
| [2011/12/02 01:12:18] <joe-mac> I did | ||
| [2011/12/02 01:12:31] <joe-mac> the provider now looks at the fields as host type key instead of name type key | ||
| [2011/12/02 01:12:36] <joe-mac> or do you mean some other change? | ||
| [2011/12/02 01:12:56] <joe-mac> basically, I need to tell puppet how to turn each line into an object so it doesn't keep re writing | ||
| [2011/12/02 01:14:59] <masterzen> joe-mac: it rewrites because your individual lines as read by the parsed file provider don't have a proper name. The provider should provide this name | ||
| [2011/12/02 01:15:13] <joe-mac> right | ||
| [2011/12/02 01:15:20] <joe-mac> so how does the provider 'provide' the nane | ||
| [2011/12/02 01:15:26] <masterzen> joe-mac: ie, you should read host, type, key but should set name, host, type and key in the resource | ||
| [2011/12/02 01:15:44] <joe-mac> right and that's what instances does | ||
| [2011/12/02 01:15:46] <masterzen> joe-mac: since your name is in fact "host-type", just add this in the post parse | ||
| [2011/12/02 01:15:59] <joe-mac> oh | ||
| [2011/12/02 01:16:05] <joe-mac> ok let me try that | ||
| [2011/12/02 01:16:07] <masterzen> joe-mac: yes, but instances is never called by puppet | ||
| [2011/12/02 01:16:21] <joe-mac> right, only ralsh you say | ||
| [2011/12/02 01:16:59] @ Spads joined channel #puppet | ||
| [2011/12/02 01:19:05] @ aleya joined channel #puppet | ||
| [2011/12/02 01:19:19] @ aleya left channel #puppet () | ||
| [2011/12/02 01:19:34] <joe-mac> moment of truth... | ||
| [2011/12/02 01:20:14] <joe-mac> nope still says created every run after I added to the post parse | ||
| [2011/12/02 01:20:17] @ apollo13 joined channel #puppet | ||
| [2011/12/02 01:20:36] @ Quit: alanevans: Ping timeout: 260 seconds | ||
| [2011/12/02 01:20:56] <joe-mac> hash[:name] = hash[:host] + "-" + hash[:type] | ||
| [2011/12/02 01:21:00] <joe-mac> that's what I added to the post | ||
| [2011/12/02 01:21:05] <joe-mac> looks fine to me | ||
| [2011/12/02 01:23:00] @ matti joined channel #puppet | ||
| [2011/12/02 01:23:00] @ Quit: matti: Changing host | ||
| [2011/12/02 01:23:00] @ matti joined channel #puppet | ||
| [2011/12/02 01:24:52] <masterzen> joe-mac: hmm, strange, I'm now out of clue | ||
| [2011/12/02 01:25:57] @ Holocaine joined channel #puppet | ||
| [2011/12/02 01:28:21] <joe-mac> :-( | ||
| [2011/12/02 01:29:05] <joe-mac> maybe that needs to go in pre gen? | ||
| [2011/12/02 01:29:11] <joe-mac> I've never messed with parsed file before | ||
| [2011/12/02 01:29:19] <joe-mac> really only package | ||
| [2011/12/02 01:30:16] @ MacDuke joined channel #puppet | ||
| [2011/12/02 01:31:13] @ Quit: cosimo: Quit: leaving | ||
| [2011/12/02 01:32:30] <joe-mac> alright, that did something | ||
| [2011/12/02 01:34:07] @ crs joined channel #puppet | ||
| [2011/12/02 01:34:41] @ Quit: raphdg: Remote host closed the connection | ||
| [2011/12/02 01:35:54] <crs> Hi there. I am getting duplicate definition error but in regards to the same very definition. Have created separate type(?) with define and have put it there. Am still getting this. Is it a bug? Version 2.7.6. | ||
| [2011/12/02 01:35:59] @ Kazuhiro joined channel #puppet | ||
| [2011/12/02 01:36:23] <kjetilho> crs: each resource created in your define must have a unique name | ||
| [2011/12/02 01:36:39] @ Kazuhiro left channel #puppet () | ||
| [2011/12/02 01:36:40] <kjetilho> so typically you need to include $name (from your define) as part of it | ||
| [2011/12/02 01:40:50] <crs> kjetilho: Absolutely true. The problem is that puppet is complaining that this resource is already defined _in the same line_. | ||
| [2011/12/02 01:41:13] @ Quit: MacDuke: Quit: Page closed | ||
| [2011/12/02 01:41:20] <kjetilho> yes | ||
| [2011/12/02 01:41:39] <kjetilho> Puppet isn't smart enough to report the line where the define is called instead | ||
| [2011/12/02 01:42:07] <crs> Oh, I see! So what am I missing? $name? Where exactly? | ||
| [2011/12/02 01:42:38] <kjetilho> whereever it is useful... file { "/tmp/$name": } | ||
| [2011/12/02 01:43:11] <kjetilho> if the resource *is* the same for all instances of your define, you can move it out and make it virtual with a @ | ||
| [2011/12/02 01:43:18] <kjetilho> and then use realize in your define | ||
| [2011/12/02 01:43:36] <joe-mac> FFS | ||
| [2011/12/02 01:43:37] <kjetilho> realize can be called as many times as you like | ||
| [2011/12/02 01:43:43] <joe-mac> who here understands parsedfile well | ||
| [2011/12/02 01:45:53] <crs> kjetilho: I didn't get to virtual resources yet. :|If there is a way not to do it - I would like to skip it right now. In my define I got one exec only and do not really need $name there at all. How could I get around it? | ||
| [2011/12/02 01:46:14] @ Quit: bgupta: Remote host closed the connection | ||
| [2011/12/02 01:46:21] @ bgupta joined channel #puppet | ||
| [2011/12/02 01:46:34] @ Quit: KaiserBeto: Quit: Leaving. | ||
| [2011/12/02 01:47:39] <kjetilho> crs: this kind of virtual resources isn't very magic. you don't need exported resources | ||
| [2011/12/02 01:47:56] <kjetilho> @user { 'apache': } | ||
| [2011/12/02 01:48:02] <kjetilho> realize User['apache'] | ||
| [2011/12/02 01:48:13] <kjetilho> that is equivalent to user { 'apache': } | ||
| [2011/12/02 01:48:53] @ sdog joined channel #puppet | ||
| [2011/12/02 01:50:47] @ jonhattan joined channel #puppet | ||
| [2011/12/02 01:51:23] <crs> kjetilho: Can I have realised (exec) with refreshonly? And then how could I realise it only if something else happens? | ||
| [2011/12/02 01:52:14] <kjetilho> if it is refreshonly, you don't need to worry about not realising it | ||
| [2011/12/02 01:52:32] <crs> Oh, I see. OK, let me test what I got now. | ||
| [2011/12/02 01:52:33] <kjetilho> (it's not really possible, anyway) | ||
| [2011/12/02 01:53:13] @ less_than_puppet joined channel #puppet | ||
| [2011/12/02 01:53:29] <kjetilho> what is realised or not is decided when the puppetmaster compiles the manifest. so what happens during the node applying it can't change the manifest | ||
| [2011/12/02 01:54:04] <less_than_puppet> hi, ralsh on one node is only giving me some types (ralsh --type) however on another node it is giving me a lot more types, why is this? | ||
| [2011/12/02 01:54:40] @ Quit: sdog: Changing host | ||
| [2011/12/02 01:54:40] @ sdog joined channel #puppet | ||
| [2011/12/02 01:56:25] <crs> kjetilho: Understood. I need to run this exec only if it is called by other exec which got unless set up... | ||
| [2011/12/02 01:56:59] <kjetilho> what are you trying to do? | ||
| [2011/12/02 01:57:18] <kjetilho> when you start chaining execs, it is usually time to look for a custom type written in Ruby | ||
| [2011/12/02 01:58:06] <crs> kjetilho: I am trying to get campfire notifier after git pull. | ||
| [2011/12/02 01:58:23] <kjetilho> are you using vcsrepo ? | ||
| [2011/12/02 01:58:55] <crs> Nope, just execs. | ||
| [2011/12/02 01:59:15] @ brtl joined channel #puppet | ||
| [2011/12/02 01:59:42] <kjetilho> you should definitely check it out :) | ||
| [2011/12/02 02:00:06] <kjetilho> https://github.com/puppetlabs/puppet-vcsrepo/ | ||
| [2011/12/02 02:00:23] <kjetilho> it's awesomesauce | ||
| [2011/12/02 02:00:30] <kjetilho> esp. compared to using execs | ||
| [2011/12/02 02:02:10] <crs> Will have a look at this. For now, would just love to add that one more extra exec to the chain. :) | ||
| [2011/12/02 02:03:51] @ raphdg joined channel #puppet | ||
| [2011/12/02 02:11:43] <less_than_puppet> Anyone know how to update ralsh so it uses more types? | ||
| [2011/12/02 02:12:31] @ rocketeer joined channel #puppet | ||
| [2011/12/02 02:15:43] @ Quit: less_than_puppet: Quit: Leaving | ||
| [2011/12/02 02:16:06] @ less_than_puppet joined channel #puppet | ||
| [2011/12/02 02:16:40] @ Quit: bzrk: Remote host closed the connection | ||
| [2011/12/02 02:17:08] @ Quit: choonming: Quit: Leaving | ||
| [2011/12/02 02:19:11] @ Quit: jonhattan: Read error: No route to host | ||
| [2011/12/02 02:19:15] <crs> kjetilho: Have spent too much time with puppet. I do realise now what my problem was. I had exec { "campfire-notify": ... }. Added $app to the name which then is being different for many 'callers' and it works just fine. Thanks for all you tips. That lead me to realization. Thank you muchly! | ||
| [2011/12/02 02:19:32] <kjetilho> yw! | ||
| [2011/12/02 02:20:52] @ jonhattan joined channel #puppet | ||
| [2011/12/02 02:24:46] @ Quit: dezwart: Quit: -ENOTTY | ||
| [2011/12/02 02:26:29] @ Quit: kelseyhightower: Ping timeout: 244 seconds | ||
| [2011/12/02 02:29:38] @ Quit: _flex: Remote host closed the connection | ||
| [2011/12/02 02:33:17] @ Guest88325 is now known as yann2 | ||
| [2011/12/02 02:35:34] @ duerF joined channel #puppet | ||
| [2011/12/02 02:40:04] @ thestash joined channel #puppet | ||
| [2011/12/02 02:41:06] @ Quit: rutger_: Ping timeout: 260 seconds | ||
| [2011/12/02 02:44:39] <vStone> How can I ensure a class is loaded before the package type is used anywhere | ||
| [2011/12/02 02:48:08] @ sdd joined channel #puppet | ||
| [2011/12/02 02:48:27] <sdd> can i cuztomize the puppet dashboard ? | ||
| [2011/12/02 02:48:47] <sdd> by which language do i need to do it ? | ||
| [2011/12/02 02:49:36] @ Quit: rubic: Quit: Leaving. | ||
| [2011/12/02 02:50:03] <neilh> vStone: Class[$myclass] -> Package <| |> IIRC | ||
| [2011/12/02 02:51:34] <vStone> purrfect | ||
| [2011/12/02 02:51:35] <vStone> thx | ||
| [2011/12/02 02:52:56] @ rutger_ joined channel #puppet | ||
| [2011/12/02 02:52:59] @ flatline joined channel #puppet | ||
| [2011/12/02 02:53:06] @ Quit: sdog: Quit: Leaving. | ||
| [2011/12/02 02:53:13] @ sdog joined channel #puppet | ||
| [2011/12/02 02:53:39] @ jonhattan left channel #puppet () | ||
| [2011/12/02 02:55:39] @ Quit: Enf: Quit: Leaving | ||
| [2011/12/02 02:56:38] @ Quit: less_than_puppet: Quit: Leaving | ||
| [2011/12/02 03:00:19] @ gaveen joined channel #puppet | ||
| [2011/12/02 03:00:19] @ Quit: gaveen: Changing host | ||
| [2011/12/02 03:00:20] @ gaveen joined channel #puppet | ||
| [2011/12/02 03:01:36] <thestash> not sure if someone can give me a hand on this bug report, my setup is the last entry: http://projects.puppetlabs.com/issues/3548 | ||
| [2011/12/02 03:04:51] @ Quit: sdd: Quit: Page closed | ||
| [2011/12/02 03:04:56] @ adimania joined channel #puppet | ||
| [2011/12/02 03:05:41] @ Quit: adimania: Client Quit | ||
| [2011/12/02 03:07:41] <robinbowes> Is there any reference to puppet templating, what functions are available and how to use them? | ||
| [2011/12/02 03:08:46] <_rc_> http://docs.puppetlabs.com/guides/templating.html | ||
| [2011/12/02 03:08:57] <robinbowes> Ta | ||
| [2011/12/02 03:09:04] <robinbowes> Am fighting hiera at the moment | ||
| [2011/12/02 03:09:27] <robinbowes> Not working with numbers in the yaml file, so have to wrap them in quotes, which means they become strings | ||
| [2011/12/02 03:09:54] <robinbowes> then I get: Failed to parse template role/core/sysctl.erb: can't convert String into Integer from the template | ||
| [2011/12/02 03:09:57] <robinbowes> <sigh> | ||
| [2011/12/02 03:10:24] <_rc_> well once you're in the template you're in ruby | ||
| [2011/12/02 03:11:09] <robinbowes> Yeah, but I'm not fluent in ruby :/ | ||
| [2011/12/02 03:12:52] <_rc_> christmas is coming, buy a book and a copy of irb | ||
| [2011/12/02 03:13:52] @ d3c joined channel #puppet | ||
| [2011/12/02 03:14:13] <robinbowes> I have a couple | ||
| [2011/12/02 03:14:14] <_rc_> Strings have a .to_i method http://www.ruby-doc.org/core-1.9.3/String.html#method-i-to_i | ||
| [2011/12/02 03:14:56] <robinbowes> Ah yes - thanks. Let me try that... | ||
| [2011/12/02 03:15:43] <robinbowes> Ah, not integers | ||
| [2011/12/02 03:16:11] <robinbowes> to_f | ||
| [2011/12/02 03:16:26] <d3c> I'm seeing a problem where running puppet via mcollective (puppetd app, runs puppetd --onetime) will not pick up changes on nodes. if I run puppet agent -o --ignorecache, it will. I added ignorecache = true to puppet.conf under [main]. any reason this wouldn't work? | ||
| [2011/12/02 03:17:42] @ ckauhaus joined channel #puppet | ||
| [2011/12/02 03:18:03] <robinbowes> d3c: I don't see one - try it and see! | ||
| [2011/12/02 03:18:32] <robinbowes> I generally use --test, which I believe includes several common options including --ignorecache | ||
| [2011/12/02 03:20:55] <d3c> robinbowes: the mcollective puppetd agent runs puppet once via puppetd --once. see https://github.com/puppetlabs/mcollective-plugins/blob/master/agent/puppetd/agent/puppetd.rb#L141 - this however doesn't seem to do the trick even though my config looks like this: http://pastebin.com/6tqXc0AQ | ||
| [2011/12/02 03:21:17] <d3c> robinbowes: that config is for nodes | ||
| [2011/12/02 03:21:45] <_rc_> when you say it doesn't do the trick, what does that mean? | ||
| [2011/12/02 03:22:35] <_rc_> is it taking two runs to make a manifest change manifest that kind of thing? | ||
| [2011/12/02 03:22:42] <d3c> _rc_: running via mco puppetd runonce -f --with-fact role=web (or whatever fact), it doesn't pick up changes on the node. if I run puppet manually on the node via puppet agent -o --ignorecache, it will. | ||
| [2011/12/02 03:23:15] <_rc_> what if you make the second run puppetd --once ? | ||
| [2011/12/02 03:23:27] <_rc_> er --onetime even | ||
| [2011/12/02 03:24:22] <d3c> _rc_: will try and see what it does then. sec | ||
| [2011/12/02 03:24:46] <_rc_> as it may be you have the caching bug on your master. what version do you run on the master? and is it under passenger? | ||
| [2011/12/02 03:25:13] @ Quit: KaZeR_W: Read error: Connection reset by peer | ||
| [2011/12/02 03:26:15] <d3c> _rc_: 2.6.6 and is running under passenger+httpd | ||
| [2011/12/02 03:26:39] <d3c> _rc_: nodes are all 2.6.6 as well | ||
| [2011/12/02 03:27:03] <_rc_> right, I think that might be your issue | ||
| [2011/12/02 03:27:04] @ bzrk joined channel #puppet | ||
| [2011/12/02 03:27:09] <d3c> _rc_: any known issues with caching? (seems like it isn't the first time you see this issue) | ||
| [2011/12/02 03:27:17] <_rc_> I just need to find my notes on what version fixed it | ||
| [2011/12/02 03:27:49] @ mlout joined channel #puppet | ||
| [2011/12/02 03:27:52] <_rc_> http://projects.puppetlabs.com/issues/5318 | ||
| [2011/12/02 03:27:58] <_rc_> fixed in 2.6.9 | ||
| [2011/12/02 03:28:36] <robinbowes> Is it normal to need to use .to_f to do calculations in templates, based on facts? | ||
| [2011/12/02 03:29:10] <_rc_> it's not normal to do a lot of maths in templates, but if you are, then yes | ||
| [2011/12/02 03:29:14] <robinbowes> For example, I've ended up with this (which works): vm.nr_hugepages = <%= (memorysize_raw.to_f * hugepage_ratio.to_f / hugepage_pagesize.to_i).to_i %> | ||
| [2011/12/02 03:29:26] <_rc_> as ruby treats strings and numbers as distinct things | ||
| [2011/12/02 03:29:38] <robinbowes> Yeah, and facts are all strings | ||
| [2011/12/02 03:30:02] <robinbowes> So I'm thinking "yes" I do need to explicitly cast the strings as numbers | ||
| [2011/12/02 03:30:49] <robinbowes> I'm doing the calculations in the template as I don't see how to do it in puppet | ||
| [2011/12/02 03:31:16] <d3c> _rc_: thanks, appreciate your help. will get puppet upgraded. going for 2.7.6 - no known caching issues there, right? ;) | ||
| [2011/12/02 03:31:26] <robinbowes> eg. to take the integer part of a calculation | ||
| [2011/12/02 03:31:40] <robinbowes> d3c: might be better to stay with in 2.6.x | ||
| [2011/12/02 03:31:42] <_rc_> d3c: I think that one got fixed in 2.7.2 | ||
| [2011/12/02 03:31:52] <_rc_> but 2.7 is a big upgrade | ||
| [2011/12/02 03:31:59] <robinbowes> 2.7.x is ... what he said! :) | ||
| [2011/12/02 03:33:03] <d3c> what do I have to look out for? I mean, I wrote all manifests within the last month, it's just that amazon has 2.6.6 in their repo | ||
| [2011/12/02 03:35:14] <neilh> what's the syntax in a template for getting a variable defined in a hash? I.e. in puppet, I have $hash= { key1 => 'value1' }, In the template what do I need inside the <%= %> ? | ||
| [2011/12/02 03:35:17] <d3c> robinbowes, _rc_: ^ :-) | ||
| [2011/12/02 03:36:29] <_rc_> d3c: say bump a few times too, it really makes answering intractable problems trivial | ||
| [2011/12/02 03:37:16] <_rc_> neilh: standard ruby syntax - hash["key1"]. But you didn't say what part of the hash you want to get at | ||
| [2011/12/02 03:38:08] <d3c> _rc_: lol, I just forgot to mention you. was just to be nice so your client notified you. wasn't to ask you to hurry up or anything. | ||
| [2011/12/02 03:38:21] <_rc_> d3c: a great deal changed http://projects.puppetlabs.com/projects/1/wiki/Release_Notes | ||
| [2011/12/02 03:39:13] <_rc_> if you just want to fix your issue, the smallest change is to go to 2.6.12 | ||
| [2011/12/02 03:39:52] <neilh> _rc_: I want the template to generate "value1". I'v tried the syntax you suggest (only difference from what I'd already tried was the quotes), yet I get an error: "Failed to parse template nbx_apache/vhost-pay.erb: can't convert String into Integer" | ||
| [2011/12/02 03:40:11] <robinbowes> neilh: that's a different problem! | ||
| [2011/12/02 03:40:21] <d3c> _rc_: will read the release notes. thanks once again | ||
| [2011/12/02 03:40:40] @ Quit: bolt12: Quit: Leaving. | ||
| [2011/12/02 03:40:49] <robinbowes> neilh: I've ended up with this: | ||
| [2011/12/02 03:40:53] <robinbowes> vm.nr_hugepages = <%= (memorysize_raw.to_f * hugepage_ratio.to_f / hugepage_pagesize.to_i).to_i %> | ||
| [2011/12/02 03:41:16] <_rc_> neilh: show us the actual template? | ||
| [2011/12/02 03:41:21] <robinbowes> Not using hashes, but shows how to explicitly cast varialbes to numbers | ||
| [2011/12/02 03:41:53] <_rc_> it may be that the error message is that you've actually got an array alled hash | ||
| [2011/12/02 03:42:06] <neilh> robinbowes: nothing is (or should be) an integer - it's all strings AFAIK | ||
| [2011/12/02 03:42:17] <robinbowes> See, you have key1 => 'value1', so hash['key1'] returns a string | ||
| [2011/12/02 03:42:24] <_rc_> your hash didn't make it into template space | ||
| [2011/12/02 03:42:28] <robinbowes> OK, so what are you trying to do with the valie in the template? | ||
| [2011/12/02 03:42:35] <neilh> _rc_: oh, ok - will change that to guarantee no name conflict. 2s | ||
| [2011/12/02 03:42:48] <robinbowes> As _rc_ asks, pastie your whole template | ||
| [2011/12/02 03:42:51] <_rc_> don't call a thing hash | ||
| [2011/12/02 03:42:57] <_rc_> or test | ||
| [2011/12/02 03:43:12] <_rc_> http://pastie.org/2954500 | ||
| [2011/12/02 03:43:18] <_rc_> hash is a function in ruby | ||
| [2011/12/02 03:44:02] * neilh now feels a muppet - there was a name clash - it now works. d'oh - thx guys | ||
| [2011/12/02 03:44:21] <neilh> yes, rookie mistake ;) | ||
| [2011/12/02 03:45:19] <_rc_> using a sigil for variables might be ugly, but at least you know you're talking about a variable | ||
| [2011/12/02 03:48:17] @ Quit: sascha_: Read error: Connection reset by peer | ||
| [2011/12/02 03:50:14] <thestash> miserably bumping my question one last time, would appreciate if someone can give me a hand on this bug report, my problem/setup is the last entry: http://projects.puppetlabs.com/issues/3548 | ||
| [2011/12/02 03:51:07] <_rc_> PL thinks it's fixed. Have you make super duper sure that the puppet user can read /etc/puppet/environments/testing/files/system/myfolder/somefilehere ? | ||
| [2011/12/02 03:51:24] <thestash> yes | ||
| [2011/12/02 03:51:47] <thestash> is there a way i can see where the fileserver is serving from ? | ||
| [2011/12/02 03:52:09] <_rc_> try running the puppetmaster under --debug | ||
| [2011/12/02 03:52:19] <_rc_> and if that doesn't help, strace it | ||
| [2011/12/02 03:53:34] <_rc_> I didn't even think the fileserver was environment aware, I thought there was one, and you used it for all environments | ||
| [2011/12/02 03:54:00] <d3c> any standard way to build a RPM from puppet source? I found the way to build a gem (rake puppetpackages). | ||
| [2011/12/02 03:54:32] <_rc_> d3c: it might already be done for you. http://yum.puppetlabs.com/ | ||
| [2011/12/02 03:54:53] <thestash> _rc_: i'm using modules on the next version yet i unfortunately have to maintain for some more weeks a version without it | ||
| [2011/12/02 03:55:53] <d3c> _rc_: ah, nice. didn't know of that repo. thanks! | ||
| [2011/12/02 03:57:21] @ Quit: mlout: Quit: Leaving. | ||
| [2011/12/02 04:00:46] <thestash> _rc_: 02/Dec/2011:11:58:51 +0000] "GET /testing/file_metadatas/files/system/myfolder/somefilehere?&links=manage&checksum_type=md5&&recurse=true HTTP/1.1" 404 86 "-" "-" | ||
| [2011/12/02 04:00:53] <thestash> see that on my apache logs | ||
| [2011/12/02 04:01:03] <thestash> 404 | ||
| [2011/12/02 04:01:15] <thestash> yet it seems to pass "testing" there | ||
| [2011/12/02 04:01:44] <d3c> btw, is there any way to force a yum clean all before puppet runs? I tried putting an exec with yum clean all in a stage before everything else but that doesn't seem to make puppet upgrade packages though. this is checked even before this first stage. | ||
| [2011/12/02 04:02:37] <thestash> d3c: we've had to do the same with an exec | ||
| [2011/12/02 04:03:08] <d3c> thestash: but did it work for you? I have to run the manifest twice. first run does yum clean all. second run actually finds the updates | ||
| [2011/12/02 04:03:28] <thestash> yeah it worked at the time | ||
| [2011/12/02 04:05:08] <thestash> we still use it | ||
| [2011/12/02 04:06:47] @ Quit: sdog: Changing host | ||
| [2011/12/02 04:06:47] @ sdog joined channel #puppet | ||
| [2011/12/02 04:06:48] <thestash> _rc_: [02/Dec/2011:12:01:18 +0000] "GET /production/file_metadata/files/system/myfolder/somefilehere?&links=manage&checksum_type=md5&&recurse=true HTTP/1.1" 404 66 "-" "-" | ||
| [2011/12/02 04:06:57] <d3c> thestash: I have an 'init' stage that runs before anything else. but still I have to run puppet twice for it to pick up updates. how did you do it? in the same way? | ||
| [2011/12/02 04:07:53] <thestash> we have about 5 pre-stages before main | ||
| [2011/12/02 04:08:23] <thestash> but atm we're using the clean all on the main itself | ||
| [2011/12/02 04:09:13] <thestash> bit complex our system since we have internal yum repos and stuff | ||
| [2011/12/02 04:10:20] <thestash> _rc_: is it me or is the problem this call? "/production/file_metadata/" as compared to "/testing/file_metadatas/" | ||
| [2011/12/02 04:10:24] <thestash> notice the S | ||
| [2011/12/02 04:10:51] <d3c> thestash: strange that it isn't working for me then. we have our own repo as well. you're just doing /usr/bin/yum clean all, right? | ||
| [2011/12/02 04:11:08] <thestash> # Make a yum clean all exec { "clean-repos": command => "yum clean all", } | ||
| [2011/12/02 04:11:55] <thestash> we put the paths "/usr/bin etc..."on a default | ||
| [2011/12/02 04:12:10] <d3c> thestash: and that's in the main stage? | ||
| [2011/12/02 04:13:32] <thestash> no sorry | ||
| [2011/12/02 04:13:37] <thestash> it's on a pre stage | ||
| [2011/12/02 04:13:48] <thestash> mentioned before it was on main but it isn't | ||
| [2011/12/02 04:13:52] @ mlout joined channel #puppet | ||
| [2011/12/02 04:14:19] @ jamezpolley left channel #puppet () | ||
| [2011/12/02 04:15:42] @ Quit: sdog: Read error: Connection reset by peer | ||
| [2011/12/02 04:15:53] @ sdog joined channel #puppet | ||
| [2011/12/02 04:16:54] @ fff joined channel #puppet | ||
| [2011/12/02 04:16:58] <fff> hi | ||
| [2011/12/02 04:17:28] <thestash> _rc_: ignore the "S" thing i said before ... was just chuck testa playing a trick | ||
| [2011/12/02 04:18:17] @ fff left channel #puppet () | ||
| [2011/12/02 04:19:34] @ fronlius_ joined channel #puppet | ||
| [2011/12/02 04:20:35] <d3c> thestash: and that exact pre-stage contains only yum? | ||
| [2011/12/02 04:20:37] @ Quit: viq: Ping timeout: 240 seconds | ||
| [2011/12/02 04:21:04] @ Quit: fronlius: Ping timeout: 248 seconds | ||
| [2011/12/02 04:21:04] @ fronlius_ is now known as fronlius | ||
| [2011/12/02 04:21:08] <vStone> problem: during the same puppet run, I install a package that creates a repo file, and then I try to modify it with yumrepo. The result = a empty repo file (with stuff I defined in yumrepo present) | ||
| [2011/12/02 04:21:09] @ ben_duyujie joined channel #puppet | ||
| [2011/12/02 04:22:34] @ dim0 joined channel #puppet | ||
| [2011/12/02 04:25:08] @ Quit: ezmobius|zoning|: Quit: peace out | ||
| [2011/12/02 04:25:19] <d3c> is <%= scope.lookupvar('foo') %> the way to go if I wanna use facter facts in templates? getting warnings after upgrading to 2.7.2. previously just did <%= foo %> | ||
| [2011/12/02 04:25:36] @ Quit: vincent_vdk: Quit: leaving | ||
| [2011/12/02 04:27:35] @ Quit: Holocaine: | ||
| [2011/12/02 04:27:43] @ Quit: sdog: Changing host | ||
| [2011/12/02 04:27:43] @ sdog joined channel #puppet | ||
| [2011/12/02 04:27:46] <kjetilho> d3c: no, @fact is the way to go | ||
| [2011/12/02 04:27:54] @ Quit: leifmadsen: Quit: This is how I roll; animal print pants out of control! | ||
| [2011/12/02 04:27:57] <kjetilho> facts are special ... | ||
| [2011/12/02 04:27:57] <robinbowes> Ew | ||
| [2011/12/02 04:27:58] @ leifmadsen joined channel #puppet | ||
| [2011/12/02 04:28:18] <kjetilho> it was lak himself which recommended it to me! | ||
| [2011/12/02 04:28:18] <robinbowes> So, in 2.7.x, you have to pre-fix all facts used in template vars with @ ? | ||
| [2011/12/02 04:28:25] <kjetilho> yes | ||
| [2011/12/02 04:28:28] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 04:28:41] <robinbowes> Major pain | ||
| [2011/12/02 04:28:58] <d3c> easier than scope.lookupvar though. read about that somewhere | ||
| [2011/12/02 04:29:14] <d3c> http://blog.penumbra.be/2011/08/dynamic-lookup-is-deprecated-when-using-facts-in-templates/ | ||
| [2011/12/02 04:29:20] <d3c> anyway, will be appending @'s then. thanks | ||
| [2011/12/02 04:29:29] <robinbowes> pre-pending | ||
| [2011/12/02 04:29:30] <kjetilho> yes, but with lookupvar you don't need to know where the variable comes from... | ||
| [2011/12/02 04:29:35] <kjetilho> so it's a tradeoff | ||
| [2011/12/02 04:30:08] <d3c> robinbowes: right. lol | ||
| [2011/12/02 04:30:16] @ Quit: vincent_vdk: Client Quit | ||
| [2011/12/02 04:30:31] <robinbowes> Why is that being done? | ||
| [2011/12/02 04:30:35] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 04:30:39] @ Quit: vincent_vdk: Changing host | ||
| [2011/12/02 04:30:39] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 04:30:42] <robinbowes> ie. turning off dynamic variable lookup? | ||
| [2011/12/02 04:30:51] @ RhysMorgan joined channel #puppet | ||
| [2011/12/02 04:30:59] @ Quit: pugnacity: Read error: No route to host | ||
| [2011/12/02 04:31:01] <robinbowes> That's one of the things I thought was really cool about puppet templating when I first came across it | ||
| [2011/12/02 04:31:08] <RhysMorgan> Hi, is there a way to set multiple puppet master addresses in pupp.conf | ||
| [2011/12/02 04:31:21] <RhysMorgan> i.e if the first one fails try the second etc | ||
| [2011/12/02 04:33:42] @ Quit: vincent_vdk: Client Quit | ||
| [2011/12/02 04:35:24] @ Quit: ben_duyujie: Ping timeout: 260 seconds | ||
| [2011/12/02 04:36:31] @ pugnacity joined channel #puppet | ||
| [2011/12/02 04:37:06] @ Quit: skm: Quit: Computer has gone to sleep. | ||
| [2011/12/02 04:37:10] @ ben_duyujie joined channel #puppet | ||
| [2011/12/02 04:37:11] @ viq joined channel #puppet | ||
| [2011/12/02 04:37:38] @ cosimo joined channel #puppet | ||
| [2011/12/02 04:40:06] @ Quit: duerF: Ping timeout: 260 seconds | ||
| [2011/12/02 04:42:06] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2011/12/02 04:42:14] @ tom_ joined channel #puppet | ||
| [2011/12/02 04:42:14] <tom_> hi | ||
| [2011/12/02 04:42:57] @ Javid__ joined channel #puppet | ||
| [2011/12/02 04:43:10] <Javid__> I need help with facter network | ||
| [2011/12/02 04:43:25] <tom_> Would that work in a .pp "if ($dbs != undef)" ? | ||
| [2011/12/02 04:43:34] <Javid__> facter network is not working on debian any solution | ||
| [2011/12/02 04:44:00] <tom_> I can't easily change the "undef" value to something else. | ||
| [2011/12/02 04:44:06] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 04:44:33] <fragfutter> Javid__: what is not working? | ||
| [2011/12/02 04:44:49] @ Quit: vincent_vdk: Client Quit | ||
| [2011/12/02 04:45:08] <Javid__> <Fragfutter> I didnt get de network | ||
| [2011/12/02 04:45:36] <Javid__> facter NetMask is working | ||
| [2011/12/02 04:46:38] <Javid__> <fragfutter > I am using facter version 1.5.7 | ||
| [2011/12/02 04:46:57] @ Quit: gaveen: Remote host closed the connection | ||
| [2011/12/02 04:47:48] @ skm joined channel #puppet | ||
| [2011/12/02 04:48:01] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 04:50:01] <thestash> d3c: yes | ||
| [2011/12/02 04:52:15] @ ben_duyujie left channel #puppet () | ||
| [2011/12/02 04:52:37] <leifmadsen> I'm reading the Pro Puppet book and getting started learning about MCollective. The book recommends RabbitMQ as the middleware messaging system -- is that still recommended? | ||
| [2011/12/02 04:52:49] @ vincent_1dk joined channel #puppet | ||
| [2011/12/02 04:52:49] <leifmadsen> just looking to test with best practices and understand how quickly things can change :) | ||
| [2011/12/02 04:53:05] @ Quit: vincent_1dk: Client Quit | ||
| [2011/12/02 04:53:31] @ Quit: pierre1: Ping timeout: 260 seconds | ||
| [2011/12/02 04:53:37] <Javid__> has anyonw problem with "facter network" | ||
| [2011/12/02 04:56:21] @ duerF joined channel #puppet | ||
| [2011/12/02 04:58:39] @ johnston joined channel #puppet | ||
| [2011/12/02 04:58:54] @ Quit: vincent_vdk: Quit: leaving | ||
| [2011/12/02 04:59:00] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 05:00:12] @ Quit: indigo: Quit: leaving | ||
| [2011/12/02 05:00:39] @ Quit: Javid__: Quit: Page closed | ||
| [2011/12/02 05:06:54] @ Quit: Mandus: Ping timeout: 244 seconds | ||
| [2011/12/02 05:07:46] @ flatline_ joined channel #puppet | ||
| [2011/12/02 05:07:46] @ Quit: flatline: Read error: Connection reset by peer | ||
| [2011/12/02 05:08:09] @ flatline__ joined channel #puppet | ||
| [2011/12/02 05:09:05] @ Quit: flatline_: Read error: Connection reset by peer | ||
| [2011/12/02 05:09:27] @ flatline_ joined channel #puppet | ||
| [2011/12/02 05:12:25] @ Quit: nuknad: Quit: nuknad | ||
| [2011/12/02 05:13:14] @ Quit: flatline__: Ping timeout: 255 seconds | ||
| [2011/12/02 05:17:36] @ Quit: blentz: Ping timeout: 248 seconds | ||
| [2011/12/02 05:20:03] @ Mandus joined channel #puppet | ||
| [2011/12/02 05:20:55] @ Quit: vincent_vdk: Quit: leaving | ||
| [2011/12/02 05:23:39] @ mgw joined channel #puppet | ||
| [2011/12/02 05:23:52] @ choonming joined channel #puppet | ||
| [2011/12/02 05:29:00] @ giskard joined channel #puppet | ||
| [2011/12/02 05:29:03] <giskard> hey | ||
| [2011/12/02 05:29:11] <giskard> how can i re-ship a file if the content is changed? | ||
| [2011/12/02 05:29:35] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 05:29:49] <giskard> with this http://paste.debian.net/147809/ | ||
| [2011/12/02 05:29:56] <giskard> it's copied only if the file is not there | ||
| [2011/12/02 05:35:45] @ Quit: Razique: Quit: Razique | ||
| [2011/12/02 05:35:51] @ Razique joined channel #puppet | ||
| [2011/12/02 05:37:08] @ Jenza joined channel #puppet | ||
| [2011/12/02 05:38:53] @ Quit: vincent_vdk: Quit: leaving | ||
| [2011/12/02 05:40:25] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 05:41:50] <R1ck> giskard: if the checksum is different, it should copy the file | ||
| [2011/12/02 05:41:52] @ blentz joined channel #puppet | ||
| [2011/12/02 05:43:03] @ octe joined channel #puppet | ||
| [2011/12/02 05:43:24] <octe> how do i deal with two seperate modules both requiring a common package to be installed? | ||
| [2011/12/02 05:43:31] <octe> get duplicate package definitions | ||
| [2011/12/02 05:45:25] @ KaZeR_W joined channel #puppet | ||
| [2011/12/02 05:45:56] @ ryancoleman joined channel #puppet | ||
| [2011/12/02 05:47:14] @ bhearsum left channel #puppet () | ||
| [2011/12/02 05:50:26] @ uphillian joined channel #puppet | ||
| [2011/12/02 05:52:51] <fragfutter> octe: define common class and include it in both | ||
| [2011/12/02 05:53:45] <octe> that only seems logical sometimes | ||
| [2011/12/02 05:53:55] <octe> kind of breaks the independency of a module | ||
| [2011/12/02 05:56:17] @ Quit: tom_: Quit: leaving | ||
| [2011/12/02 05:57:37] <d3c> if a config file changes and a dependant service does not have status, will puppet do nothing to it then or restart it? | ||
| [2011/12/02 05:57:56] <d3c> I'm seeing a problem with php-fpm that doesn't respond to config file changes. had to reload manually | ||
| [2011/12/02 06:01:09] <thestash> d3c: notify ? | ||
| [2011/12/02 06:01:35] <neilh> A regex used in an if respects captures that you can then refer back to with $1, $2, etc. Does this work within node definition? (http://www.pastie.org/2954982) | ||
| [2011/12/02 06:02:35] <d3c> thestash: I do notify it (a service class) | ||
| [2011/12/02 06:03:20] <d3c> thestash: I just found this: http://puppet-manifest-share.googlecode.com/svn-history/r16/trunk/anjuke-puppet/modules/php/manifests/php-fpm.pp | ||
| [2011/12/02 06:03:21] <gepetto> d3c: thestash: r16 is http://projects.puppetlabs.com/repositories/revision/puppet/16 "Revision 16b23112" | ||
| [2011/12/02 06:03:41] <d3c> thestash: seems like that guy added a exec resource to reload php-fpm | ||
| [2011/12/02 06:03:55] @ Quit: KaZeR_W: Read error: Connection reset by peer | ||
| [2011/12/02 06:04:21] <thestash> yeah looks like it | ||
| [2011/12/02 06:04:36] <thestash> subscribe => File["php-fpm.conf"] | ||
| [2011/12/02 06:04:42] <thestash> even though he has a subscribe | ||
| [2011/12/02 06:04:58] <Bardack> grrrr I m facing to a problem with my puppet :( | ||
| [2011/12/02 06:05:19] <Bardack> well with a module to be exact | ||
| [2011/12/02 06:05:28] <d3c> thestash: but I think that's the problem. php-fpm won't reload via puppet unless you do service php-fpm reload. you tried it? | ||
| [2011/12/02 06:05:38] @ KaZeR_W joined channel #puppet | ||
| [2011/12/02 06:05:43] <thestash> no, sorry | ||
| [2011/12/02 06:05:46] <Bardack> I ve a module : makeconfig.pp , containing an exec to a script (located in files/ of the module) | ||
| [2011/12/02 06:05:49] <thestash> i'm a puppet n00b too :P | ||
| [2011/12/02 06:06:17] <Bardack> this script will be executed on the client, and will call facter for a reason X | ||
| [2011/12/02 06:06:34] <Bardack> if I m doing it in the way: facter hostname | ||
| [2011/12/02 06:06:40] <d3c> heh alright | ||
| [2011/12/02 06:06:41] <Bardack> it returns empty | ||
| [2011/12/02 06:06:55] <Bardack> if I run /usr/bin/facter, I ve a unknown uname -s in kernel.rb ... | ||
| [2011/12/02 06:07:05] <Bardack> but , if I run manually the script on the client, everything is fine | ||
| [2011/12/02 06:07:13] @ astrostl joined channel #puppet | ||
| [2011/12/02 06:07:30] <Bardack> it seems that te problem comes from the fact that facter command is called inside an exec of a puppet module | ||
| [2011/12/02 06:07:32] <Bardack> or something | ||
| [2011/12/02 06:07:35] <Bardack> any idea ^ | ||
| [2011/12/02 06:07:37] <Bardack> ? | ||
| [2011/12/02 06:07:46] @ hal-j joined channel #puppet | ||
| [2011/12/02 06:09:36] <Bardack> notice: /Stage[main]/Jboss::Makeconfig/Exec[makefile.pl]/returns: /usr/lib/ruby/site_ruby/1.8/facter/kernel.rb:14: command not found: uname -s | ||
| [2011/12/02 06:09:40] <Bardack> something like that ^ | ||
| [2011/12/02 06:12:43] <_rc_> Bardack: PATH | ||
| [2011/12/02 06:13:12] <_rc_> your path at that time is excluding /bin which is where uname lives | ||
| [2011/12/02 06:13:45] <_rc_> http://pastie.org/2955024 is a reduction | ||
| [2011/12/02 06:14:01] <Bardack> mmm | ||
| [2011/12/02 06:14:19] <thestash> Exec { path => '/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin'} :) | ||
| [2011/12/02 06:14:21] @ brandon_dacrib joined channel #puppet | ||
| [2011/12/02 06:14:35] <Bardack> oh ok like that | ||
| [2011/12/02 06:14:43] <Bardack> in the exec | ||
| [2011/12/02 06:14:45] <Bardack> gonna check :) | ||
| [2011/12/02 06:15:04] <_rc_> http://docs.puppetlabs.com/references/stable/type.html#exec | ||
| [2011/12/02 06:15:11] @ tremble joined channel #puppet | ||
| [2011/12/02 06:16:05] <Bardack> mmmm well | ||
| [2011/12/02 06:16:12] <Bardack> no error anymore but empty value | ||
| [2011/12/02 06:16:19] <Bardack> while by running it manually it set the value | ||
| [2011/12/02 06:16:36] <_rc_> by running what? | ||
| [2011/12/02 06:16:59] @ joe-mac1 joined channel #puppet | ||
| [2011/12/02 06:17:08] <Bardack> my perl script | ||
| [2011/12/02 06:17:13] <Bardack> normally called by exec | ||
| [2011/12/02 06:17:53] @ kelseyhightower joined channel #puppet | ||
| [2011/12/02 06:17:55] <Bardack> youhou works | ||
| [2011/12/02 06:18:08] @ Quit: kelseyhightower: Client Quit | ||
| [2011/12/02 06:18:11] <_rc_> still need more input. what does it do. what were you expecting it to do. if you can't explain show us the code | ||
| [2011/12/02 06:18:12] <Bardack> USR/BIN wasnt enough | ||
| [2011/12/02 06:18:13] <Bardack> thx :) | ||
| [2011/12/02 06:18:23] @ kelseyhightower joined channel #puppet | ||
| [2011/12/02 06:18:47] <legoklods> On my puppetmaster server I have put a module in /etc/puppet/modules and on the puppetmaster it works (i.e. running `puppetd --test` spews some errors, because the module contains a custom fact that calls an external script that's not present - but that's fixable), but the module ain't copied to clients (i.e. their /etc/puppet/modules stays empty). - What have I broken? | ||
| [2011/12/02 06:19:01] @ kolla_ joined channel #puppet | ||
| [2011/12/02 06:19:03] <_rc_> Bardack: remember when you run a thing in your shell the environment is far richer | ||
| [2011/12/02 06:19:28] <_rc_> legoklods: did you tell the clients to pluginsync? | ||
| [2011/12/02 06:19:34] <ashp> lavaman: I see you :D | ||
| [2011/12/02 06:21:25] <lisa> with puppetlabs-firewall is there a way to flush the chains (iptables -F) before performing any defined firewall{} rules? | ||
| [2011/12/02 06:23:15] <ryancoleman> legoklods: even after you've set 'pluginsync = true' on the agent, they won't copy down the module. Puppet will just copy down the custom content (like custom facts) and place them in its library. Running agent with —test —verbose will present with you details on what's going on. | ||
| [2011/12/02 06:25:28] <legoklods> _rc_ & ryancoleman: I did indeed miss pluginsync=true on the agent. Doesn't --test imply --verbose? | ||
| [2011/12/02 06:25:43] @ Quit: joe-mac1: Ping timeout: 244 seconds | ||
| [2011/12/02 06:26:18] @ Quit: Stenbryggen: Quit: Leaving. | ||
| [2011/12/02 06:26:19] <ryancoleman> legoklods: it does indeed, so ignore my sleepy mind. | ||
| [2011/12/02 06:27:01] @ Quit: Vasik|work: Ping timeout: 240 seconds | ||
| [2011/12/02 06:27:12] <fim> is there any way to `evaluate` variables inside a template? Eg: from var="repo.domain.com/${architecture}" to get the architecture variable expanded | ||
| [2011/12/02 06:28:15] @ spiette joined channel #puppet | ||
| [2011/12/02 06:28:15] <_rc_> fim: yes, it's covered in the docs. http://docs.puppetlabs.com/guides/templating.html | ||
| [2011/12/02 06:28:16] <tremble> fim: 'repo.domain.com/<%= architecture %>' ? | ||
| [2011/12/02 06:28:22] @ marczak_ joined channel #puppet | ||
| [2011/12/02 06:28:33] @ lborda joined channel #puppet | ||
| [2011/12/02 06:29:17] <fim> _rc_: it is? I've been looking at this page. Let me check again | ||
| [2011/12/02 06:30:37] <fim> tremble: the problem is that the url is not set. I mean that I want to be able to define variables that utilize puppet vars. The $arch could be at the beginning ie "${arch}.domain.com/repo" | ||
| [2011/12/02 06:30:52] @ mcarpent joined channel #puppet | ||
| [2011/12/02 06:31:08] @ Quit: marczak: Ping timeout: 260 seconds | ||
| [2011/12/02 06:31:15] <tremble> fim: Try taking a look at inline_template then | ||
| [2011/12/02 06:32:04] <fim> I was under the impression that you can't use it from within a template | ||
| [2011/12/02 06:32:05] <robinbowes> fim: not wure what you're asking | ||
| [2011/12/02 06:32:10] <robinbowes> *sure | ||
| [2011/12/02 06:32:15] <_rc_> you want scope.lokkupvar | ||
| [2011/12/02 06:32:21] <_rc_> also in that document | ||
| [2011/12/02 06:32:49] <_rc_> and you can use the inline_template function in a template, it's just perverse | ||
| [2011/12/02 06:33:01] <ryancoleman> fim: If you want to use facter variables, those are available in templates as just the plain word. ex. operatingsystem | ||
| [2011/12/02 06:33:20] @ riha left channel #puppet () | ||
| [2011/12/02 06:33:31] <ryancoleman> though I think I'm missing what you're asking.. so nevermind. | ||
| [2011/12/02 06:33:33] <_rc_> it'd be scope.function_inline_template([" super embedded template "]) | ||
| [2011/12/02 06:34:26] <_rc_> I think what fim may be after is scope.lookupvar("var_#{othervar}") | ||
| [2011/12/02 06:34:33] <fim> Well, ok. Let me describe what I'm doing. I am looking up a url using extlookup and this URL contains puppet variables inside. So it could be ""${arch}.domain.com/repo" or var="repo.domain.com/${architecture}" as defined in the host's extlookup file | ||
| [2011/12/02 06:35:03] <fim> I want to use expand the puppet var inside my URL within the template | ||
| [2011/12/02 06:35:09] <_rc_> so you want to double-expand based on what's in extdata? | ||
| [2011/12/02 06:35:19] <fim> _rc_: yep | ||
| [2011/12/02 06:35:32] <_rc_> doesn't sound like a smart idea at all | ||
| [2011/12/02 06:35:37] <tremble> Eeek that is heading in the direction of inline_template inside template | ||
| [2011/12/02 06:35:46] <Spads> sup dawg | ||
| [2011/12/02 06:36:14] <fim> So, can anyone recommend a better way to handle this? :) | ||
| [2011/12/02 06:36:17] @ Quit: pugnacity: Quit: Leaving. | ||
| [2011/12/02 06:36:29] @ Quit: duerF: Ping timeout: 252 seconds | ||
| [2011/12/02 06:36:38] <tremble> fim : Do you control the contents of what your extlookup returns, or does it need to be fixed for other reasons? | ||
| [2011/12/02 06:36:50] <_rc_> step back, rethink so you avoid layering magic on top of magic | ||
| [2011/12/02 06:37:06] * tremble agrees with _rc_ | ||
| [2011/12/02 06:37:12] <robinbowes> Use yum vars in the URL, and rely on themy being expanded by yum | ||
| [2011/12/02 06:37:29] <fim> tremble: I control them in the sense that I write the extdata. The domains though I can't control (not everytime) | ||
| [2011/12/02 06:37:35] @ Quit: mlout: Quit: Leaving. | ||
| [2011/12/02 06:38:09] <ashp> pile more evil in until it works, that's my approach | ||
| [2011/12/02 06:38:48] <tremble> fim: So either you need to go with robinbowes suggestion or you need to look at returning values that inline_template can parse ('<%= somevar %>' instead of '${somevar}') | ||
| [2011/12/02 06:39:45] <tremble> and personally I'd say you probably want to expand the variables with inline_template before you get as far as the template. | ||
| [2011/12/02 06:40:01] <tremble> If you go in that direction | ||
| [2011/12/02 06:40:11] <tremble> robinbowes suggestion is much cleaner | ||
| [2011/12/02 06:40:16] <robinbowes> When I start tying myself in knots like this, I stop and re-evaluate. I'm probably doing it wrong | ||
| [2011/12/02 06:41:17] <fim> well the problem I have with puppet is that I usually need only to write a manifest to support my environment but I get caught up in an effort to make it as abstract as possible | ||
| [2011/12/02 06:41:36] <_rc_> yeah; stop doing that. | ||
| [2011/12/02 06:41:41] <fim> and this is the case here. This will probably never be used but *what if* :P | ||
| [2011/12/02 06:41:50] <_rc_> simplest thing that can work and then move on | ||
| [2011/12/02 06:42:42] <chrischris> Puppet "solo" question: if I want to manage the source of my modules -- are there any tools? (chef has an interesting tool called librarian...) -- or is the best practice approach to submodule? | ||
| [2011/12/02 06:42:57] <_rc_> otherwise when it breaks later, and it will, you'll spend ages tracing the logic through | ||
| [2011/12/02 06:43:59] <fim> _rc_: or when you need to expand you actually have to rewrite everything | ||
| [2011/12/02 06:44:04] <_rc_> chrischris: I think this is what rump attempts to be, that or puppet-module-tool | ||
| [2011/12/02 06:45:02] <_rc_> depends exactly what parts of librarian you're looking to emulate | ||
| [2011/12/02 06:45:13] @ Quit: flatline_: Read error: Operation timed out | ||
| [2011/12/02 06:46:17] <chrischris> _rc_: specifically the "cheffile" (gemfile) behavior -- where I can target a git repo and version of a puppet module have it download into my modules | ||
| [2011/12/02 06:48:00] <chrischris> _rc_: I am hoping to use some generalized puppet modules (e.g. apache, php, etc...) then extend the functionality I need in my own modules -- but I do not want to manage the versions in my repo locally -- I would rather those be dependencies (that maybe a rake task could pull in?) | ||
| [2011/12/02 06:48:25] <_rc_> seems odd to have that in a chef specific tool, rather than just using submodules/braid | ||
| [2011/12/02 06:48:47] <_rc_> unless there's extra logic it manages | ||
| [2011/12/02 06:48:56] @ rcrowley joined channel #puppet | ||
| [2011/12/02 06:50:24] <chrischris> _rc_: well - from what I understand (note: we did some brief experimentation with chef before settling on puppet...): chef-server manages your cookbooks (version, etc...) but to version it yourself (e.g. in git) you can submodule, but that gets tricky | ||
| [2011/12/02 06:51:24] <chrischris> _rc_: mainly b/c you cannot just include say opscode/cookbooks (as not all cookbooks are properly compatible with each other) or another source like <third party>/bookname | ||
| [2011/12/02 06:52:48] <chrischris> _rc_: so librarian serves very much like a gemfile to point to cookbooks and even lock their versions to run and go out to all the different sources and pull them in locally (either to version or at runtime) ... from there you could commit them to an "app specific" repo or push them to chef-server to be versioned | ||
| [2011/12/02 06:53:31] <chrischris> _rc_: seems like `puppet-module-tool` is definitely a step in the direction I was looking! | ||
| [2011/12/02 06:54:12] @ flatline_ joined channel #puppet | ||
| [2011/12/02 06:54:18] <_rc_> I'm not sure. I just looked at libraraian and it reminded me very much of bundler | ||
| [2011/12/02 06:54:34] <_rc_> where p-m-t is more like gem | ||
| [2011/12/02 06:55:23] <_rc_> the difference in my mind being one's for one-off bringing a module down for you to use, and the other suite is for tracking all your relationships | ||
| [2011/12/02 06:55:59] @ abetterlie joined channel #puppet | ||
| [2011/12/02 06:57:13] @ Quit: zahna: Ping timeout: 244 seconds | ||
| [2011/12/02 06:57:22] @ zahna_ joined channel #puppet | ||
| [2011/12/02 06:57:50] @ Stenbryggen joined channel #puppet | ||
| [2011/12/02 06:58:17] @ cliff-hm joined channel #puppet | ||
| [2011/12/02 06:59:30] <chrischris> _rc_: yes, I can see that: `puppet-module install myuser-mymodule --version=0.0.1` this is what piqued my interest - where I could line those up in a rake task; but as you said for tracking all it would be nice to have say a yaml with `myuser-mymodule => 0.0.1` for example | ||
| [2011/12/02 06:59:49] @ Quit: apollo13: Ping timeout: 240 seconds | ||
| [2011/12/02 07:04:16] @ apollo13 joined channel #puppet | ||
| [2011/12/02 07:04:46] @ Quit: kolla: Remote host closed the connection | ||
| [2011/12/02 07:08:11] @ Quit: Meier: Quit: Leaving. | ||
| [2011/12/02 07:08:12] @ jodok joined channel #puppet | ||
| [2011/12/02 07:08:13] @ Quit: jodok: Changing host | ||
| [2011/12/02 07:08:13] @ jodok joined channel #puppet | ||
| [2011/12/02 07:08:58] @ Quit: fronlius: Quit: fronlius | ||
| [2011/12/02 07:12:21] @ Quit: kermit: Quit: Leaving. | ||
| [2011/12/02 07:12:46] @ kermit joined channel #puppet | ||
| [2011/12/02 07:14:19] <pednyc> hey guys, im running into a syntax error in one of my templates. From what I gather this is probably relating to a missing end quote, but I do not see one in my template | ||
| [2011/12/02 07:14:20] <pednyc> new-host-2 templates (master) $ erb -x -T '-' sudoers.erb | ruby -c | ||
| [2011/12/02 07:14:20] <pednyc> -:18: syntax error, unexpected tCONSTANT, expecting $end | ||
| [2011/12/02 07:14:42] <robinbowes> pastie sudoers.erb | ||
| [2011/12/02 07:14:49] @ ghg joined channel #puppet | ||
| [2011/12/02 07:15:42] <pednyc> this is the relevant parts http://pastebin.com/fHyBdzpf | ||
| [2011/12/02 07:16:44] <robinbowes> Which is line 18 ? | ||
| [2011/12/02 07:16:58] <pednyc> same as pastebin shows | ||
| [2011/12/02 07:17:00] <pednyc> %wheel ALL=(ALL) ALL | ||
| [2011/12/02 07:17:08] @ kangman joined channel #puppet | ||
| [2011/12/02 07:17:33] <ashp> I really need to come up with a decent generic puppet framework to have master/slave environments of various things (focusing purely on rsyncing from one box to n number of slave nodes) | ||
| [2011/12/02 07:17:49] <_rc_> % in column 0 means something to erb | ||
| [2011/12/02 07:17:50] @ sascha_ joined channel #puppet | ||
| [2011/12/02 07:17:54] <ashp> Anyone do this currently? An example is we have a box that has the contents of the RHEL isos for foreman to consume - my boss wants me to push that to each datacenter | ||
| [2011/12/02 07:17:56] <robinbowes> Can't see anything obvious, I'm afraid | ||
| [2011/12/02 07:18:02] @ Quit: vincent_vdk: Quit: leaving | ||
| [2011/12/02 07:18:08] <ashp> I am also going to be building a bunch of repos that I'd like to automatically mirror | ||
| [2011/12/02 07:18:20] @ vincent_vdk joined channel #puppet | ||
| [2011/12/02 07:18:28] @ Quit: paroneayea: Remote host closed the connection | ||
| [2011/12/02 07:18:37] <_rc_> pednyc: it's taking "wheel ALL=(ALL) ALL" as literal ruby | ||
| [2011/12/02 07:19:08] @ key joined channel #puppet | ||
| [2011/12/02 07:19:18] @ asteve joined channel #puppet | ||
| [2011/12/02 07:19:19] @ Quit: asteve: Changing host | ||
| [2011/12/02 07:19:19] @ asteve joined channel #puppet | ||
| [2011/12/02 07:19:23] <pednyc> _rc_: any idea how I can modify that so it still works for sudo but doesnt cause erb to complain? | ||
| [2011/12/02 07:19:44] <kjetilho> <%%>% | ||
| [2011/12/02 07:19:46] <_rc_> http://en.wikipedia.org/wiki/ERuby "Alternatively, lines starting with a % sign are interpreted as Ruby as well:" | ||
| [2011/12/02 07:19:51] <_rc_> just %% I think | ||
| [2011/12/02 07:19:55] <pednyc> Ah thank you! | ||
| [2011/12/02 07:20:21] <_rc_> also that @customer is probably wrong | ||
| [2011/12/02 07:20:37] <_rc_> you pobably just want customer | ||
| [2011/12/02 07:21:08] <pednyc> Okay, looks like the %% fixed the original issue. thanks for that. I will check on the customer var. Thanks for your help | ||
| [2011/12/02 07:21:25] @ paroneayea joined channel #puppet | ||
| [2011/12/02 07:22:53] <Evolution> I'm starting up with puppet, and trying to get a decent master set up for a deployment similar to this: global-master -> site-local-master -> clients | ||
| [2011/12/02 07:23:19] <chjohnst> ok | ||
| [2011/12/02 07:23:20] <Evolution> I'm reading up on storeconfigs but I'm not 100% on the concept yet. | ||
| [2011/12/02 07:23:28] <Evolution> would they help in this instance, or get in the way | ||
| [2011/12/02 07:24:04] <chjohnst> I dont use them, I have 36 puppet master (maybe more now), but each pair of masters controls a single site and gets a replica from the gold server | ||
| [2011/12/02 07:24:06] <_rc_> it depends on what you wanted to do | ||
| [2011/12/02 07:24:26] @ Quit: robinbowes: Quit: leaving | ||
| [2011/12/02 07:24:44] @ robinbowes joined channel #puppet | ||
| [2011/12/02 07:25:10] <_rc_> if you want to use exported resources from any node to any node then you need to make your storeconfigs db a central thing; if you don't you don't | ||
| [2011/12/02 07:25:12] <SFox> any idea why tagmail.conf only send a blank email when trigger ? | ||
| [2011/12/02 07:25:34] <Evolution> chjohnst: thanks | ||
| [2011/12/02 07:26:11] <Evolution> _rc_: could you elaborate on that a bit? I plan to use custom facts to determine what gets installed/pushed to a client | ||
| [2011/12/02 07:26:29] <Evolution> but those facts are just for that client, not shared among them. | ||
| [2011/12/02 07:26:42] <_rc_> Evolution: http://docs.puppetlabs.com/guides/exported_resources.html | ||
| [2011/12/02 07:27:16] <_rc_> if you're not using exported resources, you probably don't want to use storeconfigs at all | ||
| [2011/12/02 07:28:33] <adedommelin> /c | ||
| [2011/12/02 07:28:37] <adedommelin> ooops | ||
| [2011/12/02 07:29:11] @ ccaum joined channel #puppet | ||
| [2011/12/02 07:29:13] @ ldlework joined channel #puppet | ||
| [2011/12/02 07:31:58] @ Quit: joe-mac: Remote host closed the connection | ||
| [2011/12/02 07:34:59] @ ldleworker joined channel #puppet | ||
| [2011/12/02 07:35:14] @ Quit: ldlework: Read error: Connection reset by peer | ||
| [2011/12/02 07:36:39] @ Quit: ocherno: Quit: Leaving | ||
| [2011/12/02 07:37:27] @ miche joined channel #puppet | ||
| [2011/12/02 07:38:36] @ pll` is now known as pll | ||
| [2011/12/02 07:41:19] @ highclasshole joined channel #puppet | ||
| [2011/12/02 07:42:20] <highclasshole> I'm getting intermittent errors on client nodes, "err: Could not retrieve catalog from remote server: end of file reached" | ||
| [2011/12/02 07:42:55] <highclasshole> anyone ever have similar issue? | ||
| [2011/12/02 07:43:10] <_rc_> highclasshole: how are you running your puppetmaster? standalone? | ||
| [2011/12/02 07:43:28] <highclasshole> yeah | ||
| [2011/12/02 07:43:38] <_rc_> how many nodes? | ||
| [2011/12/02 07:44:09] <highclasshole> 230ish | ||
| [2011/12/02 07:44:20] @ wilmoore joined channel #puppet | ||
| [2011/12/02 07:44:41] <_rc_> you're going to want to look at passenger or similar to scale your master | ||
| [2011/12/02 07:44:50] <_rc_> http://docs.puppetlabs.com/guides/passenger.html | ||
| [2011/12/02 07:45:13] <_rc_> as you're hitting contention using webrick | ||
| [2011/12/02 07:45:49] <highclasshole> ahhhh | ||
| [2011/12/02 07:45:50] <highclasshole> thanks | ||
| [2011/12/02 07:46:24] <_rc_> we started seeing regular problems past about 20 nodes | ||
| [2011/12/02 07:47:49] @ Quit: artkrz: Quit: Linkinus - http://linkinus.com | ||
| [2011/12/02 07:51:11] @ natewalck joined channel #puppet | ||
| [2011/12/02 07:52:02] @ Quit: EL84: Quit: Leaving. | ||
| [2011/12/02 07:55:38] @ Quit: miche: Quit: Computer has gone to sleep. | ||
| [2011/12/02 07:57:09] @ Quit: wilmoore: Remote host closed the connection | ||
| [2011/12/02 08:02:22] @ Quit: sascha_: Quit: sascha_ | ||
| [2011/12/02 08:02:30] <pll> Does the a user{} type auto-create a group when 'group => <gid>' is specified ? Or must that group already exist ? | ||
| [2011/12/02 08:03:00] <tremble> pll: Group needs to be created or managed by puppet | ||
| [2011/12/02 08:03:12] <tremble> pll: But it will automatically get the creation order right. | ||
| [2011/12/02 08:03:19] @ xoveruk joined channel #puppet | ||
| [2011/12/02 08:03:45] <xoveruk> how can i check from the server which i my clients successfully updated? | ||
| [2011/12/02 08:04:33] <pll> tremble: so, if <gid> doesn't already exist, and isn't managed by puppet anywhere else, puppet will error out with the "group <gid> does not exist" error? | ||
| [2011/12/02 08:05:01] <tremble> pll: puppet will throw and error. | ||
| [2011/12/02 08:05:02] @ kaptk2 joined channel #puppet | ||
| [2011/12/02 08:05:09] <tremble> pll: don't remember which. | ||
| [2011/12/02 08:05:25] <pll> Okay, that's what I'm seeing, just trying to make sure I understand the root cause. | ||
| [2011/12/02 08:05:32] <pll> thanks for the clarification! | ||
| [2011/12/02 08:05:35] <tremble> np | ||
| [2011/12/02 08:06:59] <_rc_> xoveruk: if you have reporting enabled you can look at those | ||
| [2011/12/02 08:08:27] <clustermagnet> mkay, guys… who has 34 seconds to pound some sense into me? :) as all of you, i am at a point to implement hiera/extlookup into my puppet setup… i am struggling finding examples so i can fully understand it :( | ||
| [2011/12/02 08:09:03] <clustermagnet> what i need, is, to deploy dozens of websites, svn hook scripts, etc etc, and dont like putting all that into variables inside manifest :) | ||
| [2011/12/02 08:09:18] <_rc_> clustermagnet: http://www.devco.net/archives/2009/08/31/complex_data_and_puppet.php | ||
| [2011/12/02 08:09:29] <clustermagnet> _rc_: awesome, dont tell me more, let me read | ||
| [2011/12/02 08:09:32] <_rc_> that's the basics for extlookup | ||
| [2011/12/02 08:09:55] <clustermagnet> _rc_: ive read that.... | ||
| [2011/12/02 08:10:03] <pll> Where's the page in the docs that lists the hierarchy of the /etc/puppet dir ? | ||
| [2011/12/02 08:10:07] <clustermagnet> extlookup, is it included in puppet by default? | ||
| [2011/12/02 08:10:15] <_rc_> from 2.6 onwards, yes | ||
| [2011/12/02 08:10:22] <_rc_> ealier you have to install it | ||
| [2011/12/02 08:10:33] @ Quit: sdog: Remote host closed the connection | ||
| [2011/12/02 08:10:43] @ Wolfspyre joined channel #puppet | ||
| [2011/12/02 08:10:52] <clustermagnet> _rc_: perfect… now… where should i define my variables? | ||
| [2011/12/02 08:11:07] <_rc_> clustermagnet: you said you'd read the article. | ||
| [2011/12/02 08:11:23] <clustermagnet> _rc_: reading and understanding are 2 different things… let me re-read :) | ||
| [2011/12/02 08:11:34] <clustermagnet> also, i am trying to understand the difference of hiera and extlookup | ||
| [2011/12/02 08:11:45] <_rc_> tthere's also a lot of documentation in http://www.devco.net/code/extlookup.rb | ||
| [2011/12/02 08:11:47] <clustermagnet> and why are they using csvs… and where do those go :) | ||
| [2011/12/02 08:12:00] <_rc_> hiera is the similar, with structured data | ||
| [2011/12/02 08:13:28] <ryancoleman> clustermagnet: and if you're starting from scratch, go with hiera. it's more or less a successor to extlookup and gives you more flexibility. | ||
| [2011/12/02 08:13:56] <clustermagnet> ryancoleman: thanks! .. by back-ends… it simply refers to config file types? | ||
| [2011/12/02 08:14:04] @ trickyf left channel #puppet () | ||
| [2011/12/02 08:14:04] <_rc_> one's simple to grok and aleady installed, the other is hiera | ||
| [2011/12/02 08:14:05] @ Quit: fir31ight: Remote host closed the connection | ||
| [2011/12/02 08:14:33] @ trickyf joined channel #puppet | ||
| [2011/12/02 08:14:37] @ Quit: dim0: Ping timeout: 244 seconds | ||
| [2011/12/02 08:15:14] @ duerF joined channel #puppet | ||
| [2011/12/02 08:15:26] <ryancoleman> clustermagnet: back-ends are what solution you use hiera with, like puppet. | ||
| [2011/12/02 08:15:28] <clustermagnet> how about this… i'll try both | ||
| [2011/12/02 08:15:37] <ryancoleman> http://www.devco.net/archives/2011/06/06/puppet_backend_for_hiera.php | ||
| [2011/12/02 08:15:52] <ryancoleman> That article goes over hiera and compares it to extlookup | ||
| [2011/12/02 08:16:41] <clustermagnet> maybe there i can find some keynote speach on youtube :) | ||
| [2011/12/02 08:18:30] @ Quit: jsm: Quit: jsm | ||
| [2011/12/02 08:19:11] <xoveruk> _rc_: where do the reports log to? | ||
| [2011/12/02 08:19:40] <_rc_> xoveruk: it depends which one | ||
| [2011/12/02 08:19:46] @ Quit: Razique: Quit: Razique | ||
| [2011/12/02 08:20:30] <_rc_> xoveruk: http://docs.puppetlabs.com/references/stable/report.html | ||
| [2011/12/02 08:20:36] @ axisys joined channel #puppet | ||
| [2011/12/02 08:21:47] @ l0n joined channel #puppet | ||
| [2011/12/02 08:23:10] <pll> Is there any documentation on managing users, the Puppet_Best_Practice_2 page seems overly brief and lacking in examples | ||
| [2011/12/02 08:23:16] @ n8 joined channel #puppet | ||
| [2011/12/02 08:23:31] @ Quit: raphdg: Ping timeout: 260 seconds | ||
| [2011/12/02 08:23:35] <_rc_> users are a pain | ||
| [2011/12/02 08:24:04] <pll> I know, life would be much easier without them :) | ||
| [2011/12/02 08:24:11] @ f0 joined channel #puppet | ||
| [2011/12/02 08:25:04] @ tblamer joined channel #puppet | ||
| [2011/12/02 08:25:13] @ PedroGomes joined channel #puppet | ||
| [2011/12/02 08:28:43] @ Quit: jaytaph: Quit: jaytaph | ||
| [2011/12/02 08:31:08] @ Quit: paroneayea: Remote host closed the connection | ||
| [2011/12/02 08:32:07] @ Quit: thestash: Ping timeout: 265 seconds | ||
| [2011/12/02 08:32:58] @ paroneayea joined channel #puppet | ||
| [2011/12/02 08:36:58] <pll> if setting up user {jdoe: uid=> 500, gid=> 500}, is it necessary to previously declare group { jdoe: gid => 500 } ? Or, will leaving the gid attribute off dtrt ? | ||
| [2011/12/02 08:37:07] <pll> i.e. what's the common/best practice here ? | ||
| [2011/12/02 08:37:16] <robinbowes> I create the group too | ||
| [2011/12/02 08:37:30] <pll> okay. | ||
| [2011/12/02 08:37:33] <robinbowes> I have a define I use that creates both user and group | ||
| [2011/12/02 08:38:11] <pll> that sounds reasonable... I'll probably eventually migrate this class into using virtuals at some point. | ||
| [2011/12/02 08:38:40] @ Quit: tblamer: Ping timeout: 248 seconds | ||
| [2011/12/02 08:38:53] <xoveruk> what is the default update value for puppetd ? | ||
| [2011/12/02 08:39:15] <_rc_> what? | ||
| [2011/12/02 08:39:16] @ neonflux joined channel #puppet | ||
| [2011/12/02 08:39:36] <xoveruk> do i need to specify when to update? | ||
| [2011/12/02 08:39:42] <robinbowes> Guessing... every 30 minutes? | ||
| [2011/12/02 08:39:52] <_rc_> do you mean how often it polls for changes? | ||
| [2011/12/02 08:39:53] <xoveruk> i have no changes since the 14th nov | ||
| [2011/12/02 08:40:08] <xoveruk> when i did a pupet agent -t it updated the files. | ||
| [2011/12/02 08:40:46] <xoveruk> does the client pull or the server push? | ||
| [2011/12/02 08:40:58] <_rc_> http://docs.puppetlabs.com/references/stable/configuration.html#runinterval | ||
| [2011/12/02 08:41:03] <_rc_> the client pulls | ||
| [2011/12/02 08:41:24] @ tblamer joined channel #puppet | ||
| [2011/12/02 08:41:25] <_rc_> unless you set up the puppet-kick stuff | ||
| [2011/12/02 08:43:10] @ Quit: jgaret: Quit: jgaret | ||
| [2011/12/02 08:43:35] <_rc_> that sounds like you weren't running the agent | ||
| [2011/12/02 08:43:59] @ ludicruz joined channel #puppet | ||
| [2011/12/02 08:44:21] @ andreaturli joined channel #puppet | ||
| [2011/12/02 08:44:22] @ wilmoore joined channel #puppet | ||
| [2011/12/02 08:44:33] <xoveruk> what is the overhead of running this continuously? | ||
| [2011/12/02 08:45:09] <_rc_> it eats memory like it's candy. you're better off running puppetd --onetime out of cron | ||
| [2011/12/02 08:46:15] @ seanh-ansca joined channel #puppet | ||
| [2011/12/02 08:48:02] @ cyflea joined channel #puppet | ||
| [2011/12/02 08:49:01] @ trickyf left channel #puppet () | ||
| [2011/12/02 08:49:12] <pednyc> hi again. I'm still trying to figure out the proper way to escape these lines in my sudoers.erb template so that they are printed literally to the /etc/sudoers file | ||
| [2011/12/02 08:49:14] <pednyc> %wheel ALL=(ALL) ALL | ||
| [2011/12/02 08:49:17] <pednyc> %%wheel ALL=(ALL) ALL | ||
| [2011/12/02 08:49:25] <pednyc> that ends up printing that literally in the file | ||
| [2011/12/02 08:49:47] <pednyc> i need it to pass tests and show up correctly in the file. not sure what the correct syntax is to escape the special character % | ||
| [2011/12/02 08:50:07] <_rc_> you double it | ||
| [2011/12/02 08:50:22] * Spads just uses lots of little files in sudoers.d and skips templating altogether | ||
| [2011/12/02 08:50:36] <pednyc> when i double it, it is added to /etc/sudoers literally as %%wheel ALL=(ALL) ALL | ||
| [2011/12/02 08:50:43] <pednyc> it passes tests but breaks sudo | ||
| [2011/12/02 08:50:48] <ashp> lavaman: Public callout, you wiped out 156 .pp files by running lint on them and commited it, giving me a _heart attack_! | ||
| [2011/12/02 08:50:57] <mgw> *also uses little files* | ||
| [2011/12/02 08:51:13] <xoveruk> how do i change columns with top | ||
| [2011/12/02 08:51:16] @ Quit: RhysMorgan: Quit: RhysMorgan | ||
| [2011/12/02 08:51:23] <xoveruk> when i press left or right with shift nothing happens, just get 'unknown command' | ||
| [2011/12/02 08:51:24] <pednyc> I'm on centos, not sure i have sudoers.d | ||
| [2011/12/02 08:51:27] <Spads> xoveruk: < and > | ||
| [2011/12/02 08:51:58] <_rc_> so long as it's over 5.4 I think you have it | ||
| [2011/12/02 08:52:35] <Spads> xoveruk: I like to < over to the state column and hit shift-R to reverse so that D and R states are at the top | ||
| [2011/12/02 08:52:41] <pednyc> i will look into it. any idea of something else to try in the meantime via the template before i can implement that? | ||
| [2011/12/02 08:52:50] <_rc_> pednyc: did you move it out of the 1st column? | ||
| [2011/12/02 08:53:05] <_rc_> you only need to double it if it's in the 1st column | ||
| [2011/12/02 08:53:07] <pednyc> i did not, so you think just spacing it over may work? will try that | ||
| [2011/12/02 08:53:15] @ maxmartin joined channel #puppet | ||
| [2011/12/02 08:53:20] <mgw> pednyc: sudo >= 1.7.2p2-3 has sudoers.d (but you might have to make the directory) | ||
| [2011/12/02 08:53:33] <_rc_> http://pastie.org/2955752 | ||
| [2011/12/02 08:53:35] <pednyc> mgw: thank you | ||
| [2011/12/02 08:53:48] <xoveruk> wow, it uses loads of memory | ||
| [2011/12/02 08:53:52] <_rc_> you'll need to make the directory and add the directive | ||
| [2011/12/02 08:53:57] <pednyc> _rc_: thank you sir | ||
| [2011/12/02 08:54:31] @ sascha_ joined channel #puppet | ||
| [2011/12/02 08:56:25] @ bosszaru joined channel #puppet | ||
| [2011/12/02 08:59:13] @ labrown joined channel #puppet | ||
| [2011/12/02 08:59:26] @ Quit: negronjl: Remote host closed the connection | ||
| [2011/12/02 08:59:41] @ nfagerlund joined channel #puppet | ||
| [2011/12/02 08:59:49] @ negronjl joined channel #puppet | ||
| [2011/12/02 09:00:30] @ rharrison joined channel #puppet | ||
| [2011/12/02 09:02:59] @ Quit: sascha_: Quit: sascha_ | ||
| [2011/12/02 09:03:13] @ Quit: Ramonster: Quit: So Long and thanks for all the fish | ||
| [2011/12/02 09:03:25] @ Quit: Spads: Ping timeout: 240 seconds | ||
| [2011/12/02 09:03:43] @ Quit: the_ramink: Ping timeout: 259 seconds | ||
| [2011/12/02 09:05:04] @ Quit: cwerki: Quit: leaving | ||
| [2011/12/02 09:06:07] @ Quit: ali3n0: Ping timeout: 244 seconds | ||
| [2011/12/02 09:08:39] @ Quit: d3c: Quit: leaving | ||
| [2011/12/02 09:09:05] <fixxxermet> Any of you guys using the puppet-zenoss module (https://github.com/mamba/puppet-zenoss) ? | ||
| [2011/12/02 09:09:19] <fixxxermet> Wondering if that is still useful as it hasn't been updated since 09 | ||
| [2011/12/02 09:10:02] @ otaku2 joined channel #puppet | ||
| [2011/12/02 09:10:07] @ gniks joined channel #puppet | ||
| [2011/12/02 09:10:10] @ otaku2 left channel #puppet () | ||
| [2011/12/02 09:12:02] <PedroGomes> Hi, I'm trying to build to use a external node interface in puppet but my clients throw errors like: "err: Could not retrieve catalog from remote server: Error 400 on SERVER: Could not find class ntp for 3d8.foreman.lsd.com on node 3d8.foreman.lsd.com" | ||
| [2011/12/02 09:13:24] @ sascha_ joined channel #puppet | ||
| [2011/12/02 09:13:26] <_rc_> that's not really an enc issue, it's that the autoloader can't find your ntp module | ||
| [2011/12/02 09:14:12] <_rc_> http://docs.puppetlabs.com/guides/modules.html#module-autoloading | ||
| [2011/12/02 09:15:05] <PedroGomes> nvm, yep I add a bad module path | ||
| [2011/12/02 09:15:14] <PedroGomes> _rc_: thanks anyway | ||
| [2011/12/02 09:15:54] @ Quit: brtl: Ping timeout: 245 seconds | ||
| [2011/12/02 09:16:09] @ Quit: inecas: Quit: Leaving | ||
| [2011/12/02 09:16:20] <PedroGomes> _rc_: btw, I can't open any puppet related pages from puppetlab, is the problem on my side? | ||
| [2011/12/02 09:16:49] @ devonstclair joined channel #puppet | ||
| [2011/12/02 09:16:50] @ brtl joined channel #puppet | ||
| [2011/12/02 09:16:53] <_rc_> I'd guess so | ||
| [2011/12/02 09:16:58] @ ali3n0 joined channel #puppet | ||
| [2011/12/02 09:17:01] <PedroGomes> strange | ||
| [2011/12/02 09:17:13] @ Quit: devonstclair: Remote host closed the connection | ||
| [2011/12/02 09:17:23] @ devonstclair joined channel #puppet | ||
| [2011/12/02 09:17:40] @ Quit: maxmartin: Quit: Computer has gone to sleep. | ||
| [2011/12/02 09:17:49] @ fronlius joined channel #puppet | ||
| [2011/12/02 09:17:59] @ Quit: devonstclair: Remote host closed the connection | ||
| [2011/12/02 09:20:18] @ adimania joined channel #puppet | ||
| [2011/12/02 09:21:28] @ alanna joined channel #puppet | ||
| [2011/12/02 09:22:01] @ Quit: andreaturli: Remote host closed the connection | ||
| [2011/12/02 09:22:37] @ xoveruk left channel #puppet () | ||
| [2011/12/02 09:23:18] <shortbus-> if i have a 2 execs, one of while is subscribed to a file resource, and the other requires the same file resource, which happens first; the subscribe to the requires? | ||
| [2011/12/02 09:23:52] <shortbus-> er, damn typos. that should be the subscribe or the require? | ||
| [2011/12/02 09:25:38] @ tilgovi joined channel #puppet | ||
| [2011/12/02 09:25:38] @ Quit: tilgovi: Changing host | ||
| [2011/12/02 09:25:38] @ tilgovi joined channel #puppet | ||
| [2011/12/02 09:27:37] @ Quit: cyflea: Quit: rcirc on GNU Emacs 24.0.90.1 | ||
| [2011/12/02 09:29:02] <walkeran> shortbus-: There's no answer to that. If there's no relationship between the two execs, it's indeterminate which one will happen first | ||
| [2011/12/02 09:29:32] <walkeran> shortbus-: Wait...let me read your question again... :P | ||
| [2011/12/02 09:29:33] @ Quit: f0: Ping timeout: 244 seconds | ||
| [2011/12/02 09:29:54] @ alfism joined channel #puppet | ||
| [2011/12/02 09:30:18] <shortbus-> walkeran: thx. i can pastebin the manifest if you'd like, may make it a little easier | ||
| [2011/12/02 09:30:48] <walkeran> shortbus-: Ok... I've sorted out the muddled mess which is my brain, today... I'm sticking with that answer | ||
| [2011/12/02 09:31:18] <walkeran> shortbus-: There's no way you can guarantee that one exec will fire off before the other, in that case, if there's no relationship between them | ||
| [2011/12/02 09:31:30] <shortbus-> that's fine, i can add a require to the one i want to go last, no biggie. i was just worried about creating a loop of requires | ||
| [2011/12/02 09:31:36] @ ckauhaus joined channel #puppet | ||
| [2011/12/02 09:31:39] @ Quit: ckauhaus: Client Quit | ||
| [2011/12/02 09:31:52] <walkeran> Ah..gotcha. You'll know if you create a loop :) Things don't work so good | ||
| [2011/12/02 09:32:03] @ Quit: rcrowley: Quit: Computer has gone to sleep. | ||
| [2011/12/02 09:32:29] <shortbus-> yup yup. and i get some of that lovely pink text --test creates | ||
| [2011/12/02 09:33:10] @ jsurfer joined channel #puppet | ||
| [2011/12/02 09:33:23] @ f0 joined channel #puppet | ||
| [2011/12/02 09:33:25] @ Quit: ccaum: Ping timeout: 240 seconds | ||
| [2011/12/02 09:34:10] <shortbus-> btw, whoever had the idea to make errors a horrible color that stands out like a sore thumb when testing is a genius. being able to just watch the output and look for a color instead of reading it all makes catching errors sooooo much easier | ||
| [2011/12/02 09:34:10] @ illsci joined channel #puppet | ||
| [2011/12/02 09:34:33] @ alanevans joined channel #puppet | ||
| [2011/12/02 09:35:54] <illsci> I have some interrelated modules and I want to enforce their ordering... Can you have something like Class["apple::params"] -> Class["apple::install"] -> Class["orange"] -> Class["pear"] in a apple module? | ||
| [2011/12/02 09:36:20] <illsci> or do you have to list out every subclass of the orange and pear modules.... | ||
| [2011/12/02 09:36:28] <shortbus-> the main class should be fine | ||
| [2011/12/02 09:36:29] <illsci> I have their ordering defined inside of those modules already | ||
| [2011/12/02 09:36:33] <illsci> hmm | ||
| [2011/12/02 09:37:07] <illsci> it's saying can't find resource and it's capitalizing the name of the class... | ||
| [2011/12/02 09:37:10] <illsci> Could not find resource 'Class[Newrelic_sysmon]' | ||
| [2011/12/02 09:37:21] <illsci> let me keep looking | ||
| [2011/12/02 09:39:09] <bodepd> masterzen: ping | ||
| [2011/12/02 09:41:02] @ jonhattan joined channel #puppet | ||
| [2011/12/02 09:42:48] @ Quit: lwhalen: Ping timeout: 244 seconds | ||
| [2011/12/02 09:43:03] @ lwhalen joined channel #puppet | ||
| [2011/12/02 09:43:56] @ Quit: choonming: Read error: Connection timed out | ||
| [2011/12/02 09:44:14] @ Quit: lavaman: Ping timeout: 258 seconds | ||
| [2011/12/02 09:44:18] @ maxmartin joined channel #puppet | ||
| [2011/12/02 09:45:54] <hobbeswalsh> hey puppeteers -- is there a way for me to have a directory full of .erb templates and ask puppet to recursively transfer the entire directory o' them to a node... while also interpolating all the templates and filling in variables? | ||
| [2011/12/02 09:46:29] <hobbeswalsh> like file { "/tmp/foo": ensure => "present", source => "blah", recurse => "true" } | ||
| [2011/12/02 09:46:34] <_rc_> not without writing it yourself | ||
| [2011/12/02 09:46:42] <hobbeswalsh> okay, thanks D: | ||
| [2011/12/02 09:47:31] @ ccaum joined channel #puppet | ||
| [2011/12/02 09:53:05] @ scsinutz joined channel #puppet | ||
| [2011/12/02 09:53:49] @ Quit: labrown: Ping timeout: 258 seconds | ||
| [2011/12/02 09:53:50] @ ezmobius joined channel #puppet | ||
| [2011/12/02 09:53:56] @ scsinutz left channel #puppet () | ||
| [2011/12/02 09:58:05] @ raphdg joined channel #puppet | ||
| [2011/12/02 09:58:14] @ ianbrandt joined channel #puppet | ||
| [2011/12/02 09:58:33] <PedroGomes> _rc_: strange, when I put that doubt, I changed the module path and it seemed to work, but now the error is back "Could not find class ntp…" (the default class in puppet). The module path is correct, the class has the init.pp… | ||
| [2011/12/02 09:59:39] <_rc_> PedroGomes: passenger, puppet < 2.6.9? | ||
| [2011/12/02 09:59:40] @ althara joined channel #puppet | ||
| [2011/12/02 10:01:59] <_rc_> if so then I'd guess #5318, otherwise, I'd doublecheck paths | ||
| [2011/12/02 10:02:00] <gepetto> _rc_: #5318 is http://projects.puppetlabs.com/issues/show/5318 "Puppet - Bug #5318: puppet master behind passenger does not re-parse manifests correctly when running with default environment.. It has a status of Closed and is assigned to Daniel Pittman" | ||
| [2011/12/02 10:02:41] <PedroGomes> _rc_: passenger but puppet 2.7.1. I have the database adapter configured but that seems ok also. when generating yaml node classifiersfiers includes and such are not needed ? | ||
| [2011/12/02 10:02:59] <_rc_> I think 2.7.1 has that bug | ||
| [2011/12/02 10:03:46] @ the_ramink joined channel #puppet | ||
| [2011/12/02 10:05:27] @ avandendorpe joined channel #puppet | ||
| [2011/12/02 10:06:33] @ jsm joined channel #puppet | ||
| [2011/12/02 10:06:37] @ Quit: tilgovi: Ping timeout: 240 seconds | ||
| [2011/12/02 10:06:38] <_rc_> actually, 2.7.1 should too, nvm | ||
| [2011/12/02 10:06:53] <_rc_> oh no, I remember now, it won't | ||
| [2011/12/02 10:07:45] <_rc_> http://projects.puppetlabs.com/projects/1/wiki/Release_Notes#273 | ||
| [2011/12/02 10:09:18] @ Quit: key: Remote host closed the connection | ||
| [2011/12/02 10:09:37] <_rc_> PedroGomes: your ENC returns a set of classes which are in effect included (or possibly turned into class resources) | ||
| [2011/12/02 10:11:05] @ Quit: jodok: Quit: jodok | ||
| [2011/12/02 10:12:22] @ comptona joined channel #puppet | ||
| [2011/12/02 10:12:45] <PedroGomes> _rc_: mh ok, still can't access the link if it was supposed to help…. | ||
| [2011/12/02 10:12:45] @ langworthy joined channel #puppet | ||
| [2011/12/02 10:13:36] <_rc_> PedroGomes: it says that the fix for #5318 was part of 2.7.3 (ie 2.7.1 still has the bug) | ||
| [2011/12/02 10:13:37] <gepetto> _rc_: PedroGomes: #5318 is http://projects.puppetlabs.com/issues/show/5318 "Puppet - Bug #5318: puppet master behind passenger does not re-parse manifests correctly when running with default environment.. It has a status of Closed and is assigned to Daniel Pittman" | ||
| [2011/12/02 10:14:36] <_rc_> PedroGomes: this bug means you have to run your agents twice to pick up a manifest change, as there's an internal cache | ||
| [2011/12/02 10:14:39] @ Quit: leifmadsen: Ping timeout: 245 seconds | ||
| [2011/12/02 10:14:54] <_rc_> PedroGomes: maybe you can see this? https://github.com/puppetlabs/puppet/commit/68c106e3ef192d64eb5a1e8daa1e070774909728 | ||
| [2011/12/02 10:15:15] <PedroGomes> yep thanks | ||
| [2011/12/02 10:15:28] <pll> Can manifests have hyphens in the name ? | ||
| [2011/12/02 10:15:35] <pll> i.e. vixie-cron.pp ? | ||
| [2011/12/02 10:15:40] <Volcane> pll: best avoided | ||
| [2011/12/02 10:15:46] <_rc_> they can, but with the 2.7 variable changes it'll bite you in the ass | ||
| [2011/12/02 10:15:53] <pll> Excellent! | ||
| [2011/12/02 10:15:54] <pll> thanks | ||
| [2011/12/02 10:16:00] <pabelanger> wow, I think I injected some bad data into storedconfigs sqlite3 db. Now my puppetmaster just crashes when a client connects. Rebuilding the db to see if that is thehe cas | ||
| [2011/12/02 10:16:01] <ptman> underscores? | ||
| [2011/12/02 10:16:49] @ fir31ight joined channel #puppet | ||
| [2011/12/02 10:17:38] <_rc_> what did your last testing environment die of? | ||
| [2011/12/02 10:17:57] <pabelanger> indeed, that was the case | ||
| [2011/12/02 10:18:34] @ Quit: jsurfer: Quit: jsurfer | ||
| [2011/12/02 10:18:37] @ leifmadsen joined channel #puppet | ||
| [2011/12/02 10:19:22] @ robsweet joined channel #puppet | ||
| [2011/12/02 10:19:39] @ Quit: flatline_: Ping timeout: 244 seconds | ||
| [2011/12/02 10:21:15] <robsweet> Anybody know why turning on stored configs (backed by mysql) would start my puppet master looking at /var/lib/puppet for things? I turned that on and started getting permissions errors trying to create a reports directory. I shut storedconfigs off and the errors go away. I was assuming that if it was backed by mysql, it would keep all of the data there. I know it was using mysql because it created and started populating the tables | ||
| [2011/12/02 10:21:34] <PedroGomes> _rc_: it seems so, running things twice seems to the the trick... | ||
| [2011/12/02 10:21:57] @ st33v joined channel #puppet | ||
| [2011/12/02 10:22:05] <Volcane> robsweet: reports dont go in that db | ||
| [2011/12/02 10:22:10] <dpittman> robsweet: stored configs only stores the, er, resource stuff in the database. The rest of the content and state remains in the same place. | ||
| [2011/12/02 10:22:42] <PedroGomes> is there a java module for unbutu for the oracle version that doesn't come in the packages anymore | ||
| [2011/12/02 10:22:48] <dpittman> robsweet: what surprises me is that it works without sc enabled; where would you expect that content to go normally? | ||
| [2011/12/02 10:23:35] <robsweet> that doesn't appear to be the case (WRT 'the same place'). I've got the following lines in my puppet.conf already and reports are being stored there. It's like storeconfigs starts ignoring my reports/reportdir directives. | ||
| [2011/12/02 10:23:38] @ Quit: f0: Ping timeout: 252 seconds | ||
| [2011/12/02 10:23:42] <robsweet> reports = store, http reportdir = /mnt/u01/puppet_reports | ||
| [2011/12/02 10:24:39] <dpittman> robsweet: Agreed; that is what surprises me, because storeconfigs should *not* change anything else. Can you file a bug report on this, and attach your configurations with and without storeconfigs enabled? | ||
| [2011/12/02 10:25:19] <_rc_> are you sure you're not breaking your puppet.conf in some subtle way when you turn on the storeconfigs? | ||
| [2011/12/02 10:25:59] @ Quit: lkoranda: Quit: Ex-Chat | ||
| [2011/12/02 10:26:22] @ Quit: ezryx: Remote host closed the connection | ||
| [2011/12/02 10:26:32] <robsweet> I'm fairly certain but not 100% (obviously, or I wouldn't be asking, right?) The config lines are straight from the Using Stored Configuration page. I'm not trying to use async yet either. | ||
| [2011/12/02 10:26:47] <_rc_> pastie your config? | ||
| [2011/12/02 10:28:18] @ Quit: st33v: Quit: st33v | ||
| [2011/12/02 10:28:18] @ Quit: ianbrandt: Quit: ianbrandt | ||
| [2011/12/02 10:28:23] <robsweet> http://pastie.org/2956167 | ||
| [2011/12/02 10:28:32] @ st33v joined channel #puppet | ||
| [2011/12/02 10:28:38] @ ianbrandt joined channel #puppet | ||
| [2011/12/02 10:29:44] <_rc_> and the before? | ||
| [2011/12/02 10:30:04] <robsweet> take out the block w/ storeconfigs and the db* stuff | ||
| [2011/12/02 10:30:05] @ Quit: natewalck: Read error: Connection reset by peer | ||
| [2011/12/02 10:30:14] @ ghanima joined channel #puppet | ||
| [2011/12/02 10:30:15] @ natewalck joined channel #puppet | ||
| [2011/12/02 10:30:34] <_rc_> so it's been 2.6+ all the time? | ||
| [2011/12/02 10:30:51] <robsweet> yep | ||
| [2011/12/02 10:30:54] <robsweet> no version change | ||
| [2011/12/02 10:31:02] @ Quit: natewalck: Client Quit | ||
| [2011/12/02 10:31:49] @ Quit: leifmadsen: Read error: Connection reset by peer | ||
| [2011/12/02 10:31:58] <_rc_> beats me, looks valid | ||
| [2011/12/02 10:32:04] @ flatline_ joined channel #puppet | ||
| [2011/12/02 10:32:18] <robsweet> That's what I was afraid of. Well, thanks for looking. | ||
| [2011/12/02 10:32:21] <_rc_> and I don't see anything in the code that looks even likely | ||
| [2011/12/02 10:32:40] <robsweet> I've got a meeting. I guess I'll look further after that. | ||
| [2011/12/02 10:34:35] @ cwarden joined channel #puppet | ||
| [2011/12/02 10:36:08] <cwarden> should chained resource collections work? e.g. https://gist.github.com/1424321 | ||
| [2011/12/02 10:37:08] @ scsinutz joined channel #puppet | ||
| [2011/12/02 10:38:33] @ Quit: ghanima: Quit: Leaving. | ||
| [2011/12/02 10:40:42] @ luckman212 joined channel #puppet | ||
| [2011/12/02 10:43:50] @ abourg joined channel #puppet | ||
| [2011/12/02 10:44:40] @ leifmadsen joined channel #puppet | ||
| [2011/12/02 10:46:18] @ jsurfer joined channel #puppet | ||
| [2011/12/02 10:48:53] @ ken_barber joined channel #puppet | ||
| [2011/12/02 10:49:19] @ shadoi joined channel #puppet | ||
| [2011/12/02 10:49:50] @ Quit: raphdg: Remote host closed the connection | ||
| [2011/12/02 10:51:37] <Volcane> cwarden: no | ||
| [2011/12/02 10:52:16] @ sdog joined channel #puppet | ||
| [2011/12/02 10:53:20] @ ghanima joined channel #puppet | ||
| [2011/12/02 10:53:38] @ Quit: sdog: Changing host | ||
| [2011/12/02 10:53:38] @ sdog joined channel #puppet | ||
| [2011/12/02 10:53:49] @ labrown joined channel #puppet | ||
| [2011/12/02 10:54:33] <Phibs> anyone here using puppet to install the netbackup agent ? | ||
| [2011/12/02 10:54:40] @ Quit: BRMatt: Ping timeout: 248 seconds | ||
| [2011/12/02 10:55:43] @ BRMatt joined channel #puppet | ||
| [2011/12/02 10:55:56] <Phibs> (I know, a dirty request) ;0 | ||
| [2011/12/02 10:56:30] @ sduchesneau joined channel #puppet | ||
| [2011/12/02 10:56:48] @ Quit: nottings: Ping timeout: 248 seconds | ||
| [2011/12/02 10:57:54] @ Kazuhiro joined channel #puppet | ||
| [2011/12/02 10:58:12] @ Quit: skm: Ping timeout: 244 seconds | ||
| [2011/12/02 10:58:54] @ skm joined channel #puppet | ||
| [2011/12/02 11:01:12] @ erthad left channel #puppet () | ||
| [2011/12/02 11:04:34] @ Quit: jimi_c: Remote host closed the connection | ||
| [2011/12/02 11:04:59] <rharrison> So I've got a variable scoping question. If I have a variable defined in a class contained withing one of my initial manifest files can I refer it in a module with the form $classname::variablename ? | ||
| [2011/12/02 11:05:42] <rharrison> I'm getting the following error when the master compiles | ||
| [2011/12/02 11:05:43] <rharrison> (Scope(Class[Cobbler])) Could not | ||
| [2011/12/02 11:05:43] <rharrison> look up qualified variable 'general::smtp_server'; class general has not been | ||
| [2011/12/02 11:05:43] <rharrison> evaluated | ||
| [2011/12/02 11:05:44] @ Quit: astrostl: | ||
| [2011/12/02 11:07:09] @ mrwacky42 joined channel #puppet | ||
| [2011/12/02 11:07:16] <Volcane> u need to have already included the class | ||
| [2011/12/02 11:09:11] <rharrison> I did in the node definition. | ||
| [2011/12/02 11:09:24] @ bmahe joined channel #puppet | ||
| [2011/12/02 11:09:30] <Volcane> order matters | ||
| [2011/12/02 11:09:46] <rharrison> does it also need to be included in the module to ensure its present? | ||
| [2011/12/02 11:09:49] <Volcane> so make sure its like in the file above when you include this other line | ||
| [2011/12/02 11:09:54] <Volcane> it wont harm to do that | ||
| [2011/12/02 11:10:00] <robsweet> rharrison: Make sure your include for general happens before the include chain of the thing that uses it. | ||
| [2011/12/02 11:10:09] @ Quit: PhilK: Excess Flood | ||
| [2011/12/02 11:10:41] @ jbartus joined channel #puppet | ||
| [2011/12/02 11:10:53] <Volcane> variables in other classes is generally just a bad idea | ||
| [2011/12/02 11:11:00] <Volcane> cos its so funky and order dependant etc | ||
| [2011/12/02 11:11:12] @ Quit: nfagerlund: Remote host closed the connection | ||
| [2011/12/02 11:11:15] <Volcane> and creates nasty coupling and unsharable code | ||
| [2011/12/02 11:12:25] <rharrison> What is the best way to handle variables that would be used by multiple modules? | ||
| [2011/12/02 11:12:34] <Volcane> extlookup or hiera | ||
| [2011/12/02 11:12:40] <Volcane> or an ENC | ||
| [2011/12/02 11:13:09] <mrwacky42> or get_var ;) | ||
| [2011/12/02 11:13:23] @ Quit: lwhalen: Ping timeout: 244 seconds | ||
| [2011/12/02 11:13:31] @ PhilK joined channel #puppet | ||
| [2011/12/02 11:14:28] @ st33v left channel #puppet () | ||
| [2011/12/02 11:14:34] <rharrison> hmmm, I did want to look into hiera... | ||
| [2011/12/02 11:19:57] @ lwhalen joined channel #puppet | ||
| [2011/12/02 11:20:29] @ Quit: ccaum: Quit: ccaum | ||
| [2011/12/02 11:20:57] @ natewalck joined channel #puppet | ||
| [2011/12/02 11:26:38] @ nfagerlund joined channel #puppet | ||
| [2011/12/02 11:27:45] @ tilgovi joined channel #puppet | ||
| [2011/12/02 11:27:46] @ Quit: tilgovi: Changing host | ||
| [2011/12/02 11:27:46] @ tilgovi joined channel #puppet | ||
| [2011/12/02 11:27:47] <tobert> weird .. a fact in one module, e.g. local/lib/facter/foo.rb should be available in other modules right? e.g. hadoop/manifests/lzo.pp uses $foo | ||
| [2011/12/02 11:28:16] <agaffney> yes | ||
| [2011/12/02 11:28:29] <agaffney> the origin of the fact should have no bearing on its scope | ||
| [2011/12/02 11:30:46] <tobert> agaffney: ok I'll look elsewhere then | ||
| [2011/12/02 11:33:21] @ johnston___ joined channel #puppet | ||
| [2011/12/02 11:36:14] @ Quit: natewalck: Quit: Outta here | ||
| [2011/12/02 11:37:17] @ Quit: ryancoleman: Quit: Leaving. | ||
| [2011/12/02 11:37:26] <lisa> in the first 50 pages of Pro Puppet i feel as if i've wasted the past 18 months of my Puppet existence on frivelous spaghetti manifests :( | ||
| [2011/12/02 11:37:30] @ Quit: nfagerlund: Remote host closed the connection | ||
| [2011/12/02 11:38:45] <agaffney> lisa: heh, I've been through at least 2 "overhauls" of my puppet style in the 3 or so years I've been using it | ||
| [2011/12/02 11:38:54] <lisa> Same | ||
| [2011/12/02 11:39:07] <the_ramink> lisa: I think nearly everyone has that moment somewhere in the puppet experience | ||
| [2011/12/02 11:39:14] <lisa> agaffney: i never really knew of any style guide so i just made shit up. turns out: i chose...poorly. | ||
| [2011/12/02 11:39:17] @ Quit: ianbrandt: Quit: ianbrandt | ||
| [2011/12/02 11:40:20] <agaffney> lisa: my typical style was half-way to the style guide teachings, so it wasn't a bit stretch for me to "update" my style | ||
| [2011/12/02 11:40:22] @ ianbrandt joined channel #puppet | ||
| [2011/12/02 11:40:34] @ BManojlovic joined channel #puppet | ||
| [2011/12/02 11:40:39] @ Quit: BManojlovic: Changing host | ||
| [2011/12/02 11:40:39] @ BManojlovic joined channel #puppet | ||
| [2011/12/02 11:40:41] <agaffney> the first overhaul was the switch from manifests/ to modules/ a couple of months after I started using puppet | ||
| [2011/12/02 11:41:27] @ eightyeight joined channel #puppet | ||
| [2011/12/02 11:42:47] <lisa> Ahh, we already use modules/ | ||
| [2011/12/02 11:43:02] @ Quit: johnston___: Ping timeout: 265 seconds | ||
| [2011/12/02 11:43:02] @ Quit: ludicruz: Ping timeout: 265 seconds | ||
| [2011/12/02 11:43:03] @ Quit: rocketeer: Ping timeout: 265 seconds | ||
| [2011/12/02 11:43:13] <lisa> i shove everything in init.pp, lots of classes, and chewing gum. i have a lot of work ahead of me. | ||
| [2011/12/02 11:43:25] <jsm> just finished my puppet re-write into paramaterized classes.. saved my life :) | ||
| [2011/12/02 11:43:27] <jsm> and hiera too :) | ||
| [2011/12/02 11:43:31] @ Quit: robsweet: Ping timeout: 265 seconds | ||
| [2011/12/02 11:43:40] <lisa> jsm: i'm looking forward to that | ||
| [2011/12/02 11:43:57] <walkeran> I wonder what percentage of the people in this room are currently in the process of a "puppet re-write" :) | ||
| [2011/12/02 11:44:00] <jsm> lisa: i'm working on a blog entry for it | ||
| [2011/12/02 11:44:29] @ Quit: johnston: Ping timeout: 265 seconds | ||
| [2011/12/02 11:44:33] <jsm> it will also talk about using puppet with amazon's cloudformation | ||
| [2011/12/02 11:44:53] @ choonming joined channel #puppet | ||
| [2011/12/02 11:45:38] <agaffney> walkeran: I'm always in the process of a puppet rewrite :P | ||
| [2011/12/02 11:46:30] <walkeran> agaffney: Very true :) Just when I think I've got my proposed structure figured out, such that everything will work nicely in our environment, I learn something new that drastically changes the way I think about everything | ||
| [2011/12/02 11:46:47] <walkeran> It's an interesting cycle to say the least | ||
| [2011/12/02 11:47:01] <jsm> my next re-write is to use as many existing modules as possible | ||
| [2011/12/02 11:47:01] <lisa> walkeran: i feel your pain | ||
| [2011/12/02 11:47:19] <lisa> we'll be using puppetlabs-firewall shortly, which will be our first third party module. | ||
| [2011/12/02 11:47:32] <agaffney> walkeran: it keeps things interesting :P | ||
| [2011/12/02 11:47:33] <jsm> there are a few legacy modules i've written before puppet had such a big collection | ||
| [2011/12/02 11:47:36] <lisa> it will be a blast to incorporate it into our existing modules. | ||
| [2011/12/02 11:47:55] <jsm> i'm looking forward to iptables, nfs, postgres | ||
| [2011/12/02 11:48:15] <jsm> that will pull out a lot of croft i've written hopefully | ||
| [2011/12/02 11:48:29] <agaffney> I prefer to use my own modules instead of these overly-generalized "commodity" modules | ||
| [2011/12/02 11:48:48] <lisa> we have an in-house nfs one. it's pretty janky. | ||
| [2011/12/02 11:48:54] <jsm> agaffney: i think it's good to evaluate on a per-module basis, but yeah, some of them are overly generalized | ||
| [2011/12/02 11:49:09] <lisa> agaffney: agreed. some of the 3rd party modules really suck :P | ||
| [2011/12/02 11:49:18] <agaffney> lisa: *many* of them really suck | ||
| [2011/12/02 11:49:21] @ tangledhelix joined channel #puppet | ||
| [2011/12/02 11:49:31] <agaffney> I found 3 different gitolite modules this morning that are all drastically different | ||
| [2011/12/02 11:49:40] <jsm> my biggest accomplishment was getting all of my userdata/sshkeys in hiera.. that makes stuff much cleaner | ||
| [2011/12/02 11:49:42] <the_ramink> agaffney: I'm in agreement and I find them shallow. They'll support all sorts of odd cases, but very little flexibility in the things I care about. | ||
| [2011/12/02 11:49:51] <lisa> jsm: what is hiera? | ||
| [2011/12/02 11:50:07] <agaffney> lisa: souped-up extlookup, essentially | ||
| [2011/12/02 11:50:13] <lisa> oh? | ||
| [2011/12/02 11:50:14] <jsm> lisa: what he said :) | ||
| [2011/12/02 11:50:25] <jsm> lisa: and very easy to use | ||
| [2011/12/02 11:50:52] <agaffney> the idea of relying on some data external to puppet for each puppet run doesn't sit right with me | ||
| [2011/12/02 11:51:27] <lisa> agaffney: sounds like a dependency hell issue. extlookup is, at least, selfcontained, if limited. | ||
| [2011/12/02 11:51:40] @ Quit: neonflux: Quit: Leaving. | ||
| [2011/12/02 11:51:45] <jsm> I don't like mixing my puppet code and node/parameter definitions | ||
| [2011/12/02 11:51:51] <Volcane> lisa: hiera should be in puppet core soonish | ||
| [2011/12/02 11:52:11] <Volcane> lisa, jsm: and param classes should use hiera magically and seemlessly soonish too | ||
| [2011/12/02 11:52:12] <the_ramink> Volcane: is it going to make 2.7 or end up in 2.8? | ||
| [2011/12/02 11:52:21] <Volcane> hard to say | ||
| [2011/12/02 11:52:24] <agaffney> jsm: I'll eventually move to a ENC, but there's a lot of cleanup between here and there | ||
| [2011/12/02 11:52:28] <jsm> Volcane: looking forward to that.. | ||
| [2011/12/02 11:52:38] <walkeran> jsm, Volcane: +1 | ||
| [2011/12/02 11:53:10] <Volcane> agaffney: how does an ENC resolve your question re external data per run? | ||
| [2011/12/02 11:53:40] <agaffney> Volcane: it doesn't "solve" it...just that that's probably the only external data source I'd feel comfortable with | ||
| [2011/12/02 11:54:28] @ Quit: sascha_: Quit: sascha_ | ||
| [2011/12/02 11:54:40] <Volcane> agaffney: problem is that any layering logic or whatever u need in your ENC you have to code, maintain and design yourself, this is costly and often end up being something exactly what you're trying to avoid with puppet | ||
| [2011/12/02 11:55:15] <Volcane> agaffney: writing a bunch of stuff in something normal programming language to model your systems etc | ||
| [2011/12/02 11:55:35] <jsm> Volcane: would you mind taking a look at my blog post about hiera before I publish it? I'd like someone to point out everything that's wrong about it. :) I should have it ready early next week. | ||
| [2011/12/02 11:56:02] <agaffney> Volcane: I intend my ENC to support a very simple mapping between machine name and role, and nothing else | ||
| [2011/12/02 11:56:02] <Volcane> jsm: sure - but I'll be traveling early in the week so late week will be easier | ||
| [2011/12/02 11:56:34] <jsm> Volcane: sure, i'm in no rush. where would I find your email address ? | ||
| [2011/12/02 11:56:46] <Volcane> jsm: rip@devco.net | ||
| [2011/12/02 11:56:51] <jsm> thanks | ||
| [2011/12/02 11:57:02] <Volcane> agaffney: nods, sounds like just using hiera will save you a ton of effort since it can do exactly that | ||
| [2011/12/02 11:57:12] @ ezmobius is now known as ezmobius|zoning | ||
| [2011/12/02 11:57:43] <agaffney> Volcane: the source of information will likely be a machine DB that also has information about P/O, "owner", commission date, etc. | ||
| [2011/12/02 11:58:22] <Volcane> agaffney: kewl - thats why the hiera backend logic is pluggable, so you only write something to fetch the requested data but you get all the other benefits for free | ||
| [2011/12/02 11:58:32] <agaffney> hmm, interesting | ||
| [2011/12/02 11:58:38] <agaffney> I haven't looked into hiera at all | ||
| [2011/12/02 11:58:44] <agaffney> I may end up using it later | ||
| [2011/12/02 11:59:17] @ neek joined channel #puppet | ||
| [2011/12/02 11:59:32] <jsm> Volcane: are you or anyone else working on a db plugin for hiera? | ||
| [2011/12/02 12:00:23] @ scsinutz left channel #puppet () | ||
| [2011/12/02 12:00:27] <Volcane> jsm: there's some mongo ones knocking around - a mysql one would be very very easy to write | ||
| [2011/12/02 12:00:57] <Volcane> jsm: i kind of dont want to go down that route myself since everyone think they're a special case and have their own data models and shit | ||
| [2011/12/02 12:01:09] <jsm> that makes sense | ||
| [2011/12/02 12:01:11] <Volcane> jsm: so I'll rather just provide an easy to extend infra so the special ppl can be special | ||
| [2011/12/02 12:01:15] <gepetto> ::redmine:: Wiki edit: Downloading_Puppet (#127) @ http://projects.puppetlabs.com/projects/puppet/wiki/Downloading_Puppet?version=127 (by Nigel Kersten) | ||
| [2011/12/02 12:01:16] @ johnston_ joined channel #puppet | ||
| [2011/12/02 12:02:57] <Volcane> jsm: but with it integrated in puppet proper and with param classes just supporting it out the box without any additional effort on the module writers part means that it should greatly improve the module sharing situation | ||
| [2011/12/02 12:03:18] <Volcane> jsm: a module from the forge might suddenly be usable by more than one person since you can provide data from hiera in whatever way you want | ||
| [2011/12/02 12:03:33] @ aherzog joined channel #puppet | ||
| [2011/12/02 12:03:42] @ ccaum joined channel #puppet | ||
| [2011/12/02 12:03:46] <jsm> i think reductivelabs should sponsor a module bounty... zenoss did this a couple years ago with great success | ||
| [2011/12/02 12:03:58] <jsm> give away an ipad for the top 5 modules | ||
| [2011/12/02 12:04:00] <jsm> or whatever | ||
| [2011/12/02 12:04:29] <Volcane> yeah that'd be good - once there's a good solution to some of these data problems right now it would be painful | ||
| [2011/12/02 12:05:04] @ Quit: tblamer: Ping timeout: 248 seconds | ||
| [2011/12/02 12:08:37] <tobert> wtf why are my facts that return true actually return "true"? | ||
| [2011/12/02 12:08:44] @ tblamer joined channel #puppet | ||
| [2011/12/02 12:08:46] <tobert> pebkac? | ||
| [2011/12/02 12:09:07] <Volcane> puppet doesnt really have data types | ||
| [2011/12/02 12:09:14] <_rc_> facts are just strings | ||
| [2011/12/02 12:09:48] <tobert> but puppet does have a raw true/false hmm oh well I'll just use strings then thanks | ||
| [2011/12/02 12:09:55] @ nfagerlund joined channel #puppet | ||
| [2011/12/02 12:10:18] @ Quit: abourg: Quit: abourg | ||
| [2011/12/02 12:10:45] <jsm> Volcane: i think there's a big opportunity once hiera gets included for companies whole just write puppet modules all day... and could sell them with support or provide the code for free... one day things will be so mature in these infrastructure management languages that i doubt we'll be touching the OS by hand very often. We'll just use the best of breed management system and vetted, tested, qa'd modules. | ||
| [2011/12/02 12:13:26] <Volcane> yeah - once there's a solid way to do the data hopefully things will become better in many ways wrt code sharing | ||
| [2011/12/02 12:13:56] @ Quit: hobbeswalsh: Quit: hobbeswalsh | ||
| [2011/12/02 12:14:24] @ jaytaph joined channel #puppet | ||
| [2011/12/02 12:14:26] @ mao joined channel #puppet | ||
| [2011/12/02 12:14:50] <mao> hello | ||
| [2011/12/02 12:15:21] @ RhysMorgan joined channel #puppet | ||
| [2011/12/02 12:16:01] <mao> i was under the impression that when setting up a new puppet master and new ca, the certs automatically generated already include alternatednsnames | ||
| [2011/12/02 12:16:15] <mao> i am not finding this to be the case however | ||
| [2011/12/02 12:17:39] <mao> additionally: i'm not sure what changed, but a previous master i set up used CN=ca as the issuer, but my new one is using CN=masterhostname | ||
| [2011/12/02 12:18:56] @ neonflux joined channel #puppet | ||
| [2011/12/02 12:19:02] <mao> any tips on how i can regenerate these certs such that my nodes can reference the master via the puppet cname? | ||
| [2011/12/02 12:23:26] <the_ramink> mao: you're getting hit by #10739 and you need set dns_alt_names = etc etc in your puppet.conf | ||
| [2011/12/02 12:23:26] <gepetto> the_ramink: mao: #10739 is http://projects.puppetlabs.com/issues/show/10739 "Puppet - Bug #10739: An initial installation of 2.7.6 results in a default certificate without alternate names. It has a status of Merged - Pending Release and is assigned to Josh Cooper" | ||
| [2011/12/02 12:27:05] @ ryancoleman joined channel #puppet | ||
| [2011/12/02 12:27:23] <mao> the_ramink | gepetto: thanks, i thought i was going nuts! after adding this config to the puppet.conf on the master should i just go ahead and blow away the ca.pem and masterhostname.pem? | ||
| [2011/12/02 12:27:51] <mao> will restarting puppetmasterd regenerate those or will i have to puppet cert -g them? | ||
| [2011/12/02 12:28:00] @ Quit: PedroGomes: Ping timeout: 248 seconds | ||
| [2011/12/02 12:28:56] @ dingus9 joined channel #puppet | ||
| [2011/12/02 12:30:12] <dingus9> hey gents what is the dealio with Bug #2690? I have my site.pp in git and I'd like to symlink to the file rather then copy | ||
| [2011/12/02 12:30:12] <gepetto> dingus9: #2690 is http://projects.puppetlabs.com/issues/show/2690 "Puppet - Bug #2690: puppet insists on /etc/puppet/manifests being a folder. It has a status of Duplicate and is assigned to Luke Kanies" | ||
| [2011/12/02 12:30:52] @ nottings joined channel #puppet | ||
| [2011/12/02 12:31:52] <dingus9> shut up no one likes you anyway gepetto ;) | ||
| [2011/12/02 12:34:27] @ Quit: dfamorato_: Remote host closed the connection | ||
| [2011/12/02 12:35:07] @ c00p left channel #puppet () | ||
| [2011/12/02 12:35:19] <the_ramink> mao: I'm not sure myself, but you try it and use openssl x509 -text -noout -in puppet/var/ssl/certs/puppet.pem to verify | ||
| [2011/12/02 12:36:07] @ Quit: localhost: Remote host closed the connection | ||
| [2011/12/02 12:36:08] <mao> the_ramink: thanks, that fixed my issue. is there any problem with the issuer change? | ||
| [2011/12/02 12:37:02] <mao> i mean, as long as the master's cert is signed with that ca cert, i guess it may not matter | ||
| [2011/12/02 12:37:20] @ localhost joined channel #puppet | ||
| [2011/12/02 12:39:06] <jbartus> i'm trying to get a freshly installed puppet server to connect to itself as a puppet master (running both puppetd and puppetmaster on one host) | ||
| [2011/12/02 12:39:14] <jbartus> but i get this | ||
| [2011/12/02 12:39:18] <jbartus> err: Could not retrieve catalog from remote server: hostname was not match with the server certificate | ||
| [2011/12/02 12:39:31] <jbartus> when i try to do the initial request | ||
| [2011/12/02 12:39:34] <Hunner> jbartus: What does `puppet agent --configprint server` say? | ||
| [2011/12/02 12:39:54] <jbartus> puppet | ||
| [2011/12/02 12:39:58] @ Spads joined channel #puppet | ||
| [2011/12/02 12:40:21] <jbartus> i have puppet.my.domain.com cname'd to this server (cobbler.my.domain.com) | ||
| [2011/12/02 12:40:30] <Hunner> jbartus: and what does `puppet cert list <your hostname>` say? Does it list any dns_alt_names? | ||
| [2011/12/02 12:40:43] <jbartus> err: Could not call list: Could not find a certificate for puppet | ||
| [2011/12/02 12:41:07] <Hunner> Not puppet... the fqdn | ||
| [2011/12/02 12:41:17] <Hunner> or try `puppet cert list --all` | ||
| [2011/12/02 12:41:44] <jbartus> + cobbler.my.domain.com (35:B8:B5:F3:41:4B:F0:87:80:42:9E:16:61:DF:79:A6) | ||
| [2011/12/02 12:42:08] <jbartus> they both put out just that one entry | ||
| [2011/12/02 12:42:09] <Hunner> Okay, so the puppet agent daemon is trying to contact the puppet master daemon at the short dns name of 'puppet', but the certificate of the master is only signed as 'cobbler.my.domain.com' | ||
| [2011/12/02 12:42:34] <jbartus> i see | ||
| [2011/12/02 12:42:55] <Hunner> So you should regenerate the puppet master cert with the dns_alt_names of puppet,puppet.my.domain.name | ||
| [2011/12/02 12:43:20] <Hunner> then it will still have the cobbler... name, but will also be recognized as puppet, puppet...com, etc | ||
| [2011/12/02 12:43:40] <jbartus> dns_alt_names being a config option i set somehwere? | ||
| [2011/12/02 12:44:08] <Hunner> jbartus: Yeah, you can set dns_alt_names in puppet.conf [master] section | ||
| [2011/12/02 12:44:28] @ Quit: jsurfer: Read error: Operation timed out | ||
| [2011/12/02 12:44:48] <Hunner> Also, try these commands: `puppet cert clean cobbler.my.domain.com && puppet certificate generate cobbler.my.domain.com --dns-alt-names puppet.puppet.my.domain.com && service puppetmaster restart` | ||
| [2011/12/02 12:45:10] <Hunner> I think that will work... | ||
| [2011/12/02 12:45:36] <Hunner> Otherwise, you can do the `puppet cert clean <name> && service puppetmaster restart` after adding dns_alt_names to puppet.conf | ||
| [2011/12/02 12:46:09] <Hunner> `puppet master --configprint dns_alt_names` should show your alt names, and they should also show up in `puppet cert list --all` | ||
| [2011/12/02 12:46:15] <Hunner> Then your agent will trust it :) | ||
| [2011/12/02 12:47:38] <jbartus> getting a syntax error, tried to solve it myself, getting the rdoc error | ||
| [2011/12/02 12:47:42] <jbartus> brb while i solve that | ||
| [2011/12/02 12:47:53] <jbartus> thanks though this sounds like it | ||
| [2011/12/02 12:49:40] <Hunner> jbartus: Oh, I should have asked which puppet version you have... | ||
| [2011/12/02 12:49:53] <Hunner> These commands are relevant to puppet 2.7.something... | ||
| [2011/12/02 12:50:08] <Hunner> The process is pretty much the same for older versions, just different commands | ||
| [2011/12/02 12:50:21] <jbartus> 2.6.12, from EPEL | ||
| [2011/12/02 12:50:27] <jbartus> on rhel6.1 | ||
| [2011/12/02 12:50:47] <pll> can one have nested case statements ? | ||
| [2011/12/02 12:50:59] @ Quit: jonhattan: Quit: Ex-Chat | ||
| [2011/12/02 12:50:59] <Hunner> Okay, then you might not have the `puppet certificate generate` command. Just add dns_alt_names to the puppet.conf and clean the cert and restart the puppet master | ||
| [2011/12/02 12:51:14] <Hunner> pll: If you can't, it's probably a bug :) | ||
| [2011/12/02 12:51:21] <pll> thanks. | ||
| [2011/12/02 12:51:35] <mao> Hunner | jbartus: i think puppetca --generate would still work prior to 2.7 | ||
| [2011/12/02 12:52:23] <jbartus> invalid option: --dns-alt-names | ||
| [2011/12/02 12:52:40] <jbartus> ah | ||
| [2011/12/02 12:52:45] <jbartus> it worked when i made it --dns_alt_names | ||
| [2011/12/02 12:52:55] <Hunner> syntax shmintax :) | ||
| [2011/12/02 12:53:55] <mao> ok, related issue: | ||
| [2011/12/02 12:54:29] <jbartus> :( | ||
| [2011/12/02 12:54:32] <jbartus> still get the same error | ||
| [2011/12/02 12:54:35] <jbartus> err: Could not retrieve catalog from remote server: hostname was not match with the server certificate | ||
| [2011/12/02 12:54:45] <Hunner> There isn't any nice way to just *add* dns_alt_names unfortunately (yet), so you always have to generate a new cert... you can always use the `openssl x509` commands if you wanted to try though :) | ||
| [2011/12/02 12:55:00] <mao> wait...it went away | ||
| [2011/12/02 12:55:09] @ ezmobius|zoning is now known as ezmobius|zoning| | ||
| [2011/12/02 12:55:12] <Hunner> jbartus: does the cert list command show then names, and you restarted the puppet master? | ||
| [2011/12/02 12:55:39] <jbartus> yes i restarted | ||
| [2011/12/02 12:56:27] <jbartus> + cobbler.my.domain.com (92:56:74:46:42:3F:54:ED:62:24:79:12:AC:F9:E3:A5) (alt names: DNS:cobbler.my.domain.com, DNS:puppet.my.domain.com) | ||
| [2011/12/02 12:56:30] <jbartus> thats the list output | ||
| [2011/12/02 12:57:42] <Hunner> jbartus: Okay. You should also add the short name 'puppet' to dns_alt_names :) | ||
| [2011/12/02 12:57:56] <Hunner> puppet.my.domain.com != puppet | ||
| [2011/12/02 12:58:46] <Hunner> alternatively, if you ran `puppet agent --test --server puppet.my.domain.com` then it should trust it, since that name *is* on the list | ||
| [2011/12/02 12:58:50] <jbartus> whoo hoo | ||
| [2011/12/02 12:58:51] <jbartus> thanks | ||
| [2011/12/02 12:58:53] <jbartus> that did it | ||
| [2011/12/02 12:59:11] <Hunner> mao: haha | ||
| [2011/12/02 13:00:21] @ joe-mac joined channel #puppet | ||
| [2011/12/02 13:00:56] <mao> so does PE 2.0.0 suffer from the same dns_alt_names bug? | ||
| [2011/12/02 13:01:30] <Hunner> bug? | ||
| [2011/12/02 13:02:44] <Hunner> Any new puppet master must be told its dns_alt_names... it's not precient to peer into your DNS cnames and A records :P | ||
| [2011/12/02 13:03:20] <mao> i'm referring to #10739 | ||
| [2011/12/02 13:03:20] <gepetto> mao: #10739 is http://projects.puppetlabs.com/issues/show/10739 "Puppet - Bug #10739: An initial installation of 2.7.6 results in a default certificate without alternate names. It has a status of Merged - Pending Release and is assigned to Josh Cooper" | ||
| [2011/12/02 13:03:20] <Hunner> The PE installer has a dns_alt_names entry for the answer seed file for this | ||
| [2011/12/02 13:03:29] <jbartus> since the clients are configured to talk to 'puppet' by default | ||
| [2011/12/02 13:03:39] <jbartus> could puppetmaster be configured to generate a cert with that alt name by default | ||
| [2011/12/02 13:03:45] @ Quit: johnston_: Ping timeout: 265 seconds | ||
| [2011/12/02 13:03:46] <jbartus> so that everything would 'just work'... by default | ||
| [2011/12/02 13:03:55] <mao> jbartus: it used to do that | ||
| [2011/12/02 13:04:01] <Hunner> Well, you don't want any ol' puppet node to be trusted as 'puppet' because then people could mitm your puppet master | ||
| [2011/12/02 13:04:39] <jbartus> i actually don't care but i can see how the broader userbase would | ||
| [2011/12/02 13:04:56] <mao> Hunner: makes sense re: PE answer file | ||
| [2011/12/02 13:05:49] <Hunner> It looks like that's the approach that the 10739 answer is too. Default to security | ||
| [2011/12/02 13:05:57] @ jorhett1 joined channel #puppet | ||
| [2011/12/02 13:07:00] <Hunner> Actually, if a puppet master starts, it knows it's generating its own cert, so it could be smart enough to add it... dunno | ||
| [2011/12/02 13:07:23] @ Guest80646 joined channel #puppet | ||
| [2011/12/02 13:07:45] <mao> switching topics for a second: every once in a while i get this error on a puppet run: err: Could not retrieve catalog from remote server: Could not intern from pson: source did not contain any PSON! | ||
| [2011/12/02 13:07:52] <jorhett1> Gals and guys, has any of you seen puppetmasterd under passenger quietly lock up before? #11140 | ||
| [2011/12/02 13:07:53] <gepetto> jorhett1: #11140 is http://projects.puppetlabs.com/issues/show/11140 "Puppet - Bug #11140: Passenger servers lock up completely..... It has a status of Unreviewed and is assigned to -" | ||
| [2011/12/02 13:07:56] <mao> i run it again and it works | ||
| [2011/12/02 13:08:25] @ skippy joined channel #puppet | ||
| [2011/12/02 13:08:33] @ Quit: crs: Ping timeout: 258 seconds | ||
| [2011/12/02 13:08:33] <Hunner> Man, having that commit would have been really helpful to me in the past week | ||
| [2011/12/02 13:08:46] <jorhett1> mao: your puppetmaster was too busy to respond. Are you still using webrick? (default server) You might be overloading it. | ||
| [2011/12/02 13:09:01] <skippy> is it possible to have an ENC (puppet dashboard) enabled for one environment in a puppetmaster's puppet.conf, but disabled for another environment? | ||
| [2011/12/02 13:09:06] @ hobbeswalsh joined channel #puppet | ||
| [2011/12/02 13:09:59] <mrwacky42> ken_barber: https://github.com/puppetlabs/puppetlabs-firewall/pull/34/ | ||
| [2011/12/02 13:10:06] <mrwacky42> Updated with latest hotness | ||
| [2011/12/02 13:10:09] <Hunner> skippy: the node_terminus setting can be set per environment, I think | ||
| [2011/12/02 13:11:14] <mao> jorhett1: yes, it's an out of the box gem install. we're looking to move to PE soon so it may not be worth investigating further | ||
| [2011/12/02 13:11:50] <pll> Hmm, $lsbdistrelease and $lsbdistid aren't getting evaluated for some reason... | ||
| [2011/12/02 13:12:36] <Hunner> mao: Yeah, PE sets up the PM with apache/passenger/rack by default and should be good up to about 2000 nodes. The gem default goes to about 20-40 :) | ||
| [2011/12/02 13:13:07] <jorhett1> mao: passenger install is easier than the docs indicate and pretty painless | ||
| [2011/12/02 13:13:08] @ jmslagle joined channel #puppet | ||
| [2011/12/02 13:14:13] @ Quit: hleitzell: Read error: Connection reset by peer | ||
| [2011/12/02 13:14:16] <jorhett1> if you're on an el system, "yum install rubygem-rack rubygem-passenger", copy the apache configuration file into place, stop puppetmaster and start apache. 4 minutes tops :) | ||
| [2011/12/02 13:14:19] @ hleitzell joined channel #puppet | ||
| [2011/12/02 13:15:28] <skippy> Hunner: hrm. Testing suggests node_terminus and external_nodes are global, not per-environment. :( | ||
| [2011/12/02 13:19:19] @ Quit: bma: Quit: ZNC - http://znc.sourceforge.net | ||
| [2011/12/02 13:19:19] @ mutex_ joined channel #puppet | ||
| [2011/12/02 13:19:19] @ Quit: eshamow: Quit: ZNC - http://znc.sourceforge.net | ||
| [2011/12/02 13:19:20] @ Quit: mutex: Ping timeout: 252 seconds | ||
| [2011/12/02 13:19:20] @ Quit: hacim: Ping timeout: 248 seconds | ||
| [2011/12/02 13:19:21] @ Quit: jeremyb: Ping timeout: 248 seconds | ||
| [2011/12/02 13:19:22] @ Quit: ergonlogic: Ping timeout: 248 seconds | ||
| [2011/12/02 13:19:22] @ jeremyb joined channel #puppet | ||
| [2011/12/02 13:22:59] @ Joined channel #puppet | ||
| [2011/12/02 13:22:59] @ Topic is "Puppet Enterprise 2.0 is out: http://bit.ly/PE_2 | Important Security Announcement: http://bit.ly/AltNames | Puppet 2.7.6 and 2.6.12 available at downloads.puppetlabs.com | http://docs.puppetlabs.com | Bugs & Feature Requests http://bit.ly/dKVa0A | November Newsletter http://bit.ly/nl11-11" | ||
| [2011/12/02 13:22:59] @ Topic set by stahnma!~stahnma@fedora/stahnma on Wed Nov 30 10:47:21 -0800 2011 | ||
| [2011/12/02 13:23:12] @ Mode +Ccnt by asimov.freenode.net | ||
| [2011/12/02 13:23:32] @ miche joined channel #puppet | ||
| [2011/12/02 13:24:03] @ eshamow joined channel #puppet | ||
| [2011/12/02 13:24:21] @ Quit: maxmartin: Read error: Connection reset by peer | ||
| [2011/12/02 13:24:51] @ maxmartin joined channel #puppet | ||
| [2011/12/02 13:26:15] <LoRez> is there a less messy way to do this? http://pastebin.com/VeYSpE6t | ||
| [2011/12/02 13:29:03] @ Quit: tblamer: Read error: Connection reset by peer | ||
| [2011/12/02 13:29:40] @ crs joined channel #puppet | ||
| [2011/12/02 13:30:12] <dingus9> LoRez: I would do that all in an erg template | ||
| [2011/12/02 13:30:17] <dingus9> *erb | ||
| [2011/12/02 13:30:21] @ tblamer joined channel #puppet | ||
| [2011/12/02 13:30:24] @ Quit: Guest80646: Ping timeout: 258 seconds | ||
| [2011/12/02 13:30:32] @ Quit: eshamow: Quit: ZNC - http://znc.sourceforge.net | ||
| [2011/12/02 13:30:50] @ Quit: aglet: Quit: ZNC - http://znc.in | ||
| [2011/12/02 13:31:00] @ Quit: adimania: Ping timeout: 252 seconds | ||
| [2011/12/02 13:31:21] @ aglet joined channel #puppet | ||
| [2011/12/02 13:31:30] <pll> In a case statement, can there be multiple assignments or whatever on the RHS within the {...} ? | ||
| [2011/12/02 13:31:42] <eric0> pll yes | ||
| [2011/12/02 13:32:02] <pll> , or ; separated ? | ||
| [2011/12/02 13:32:24] <eric0> if you have assignments they can just be whitespace separated '$var=val1 $var2=val2' | ||
| [2011/12/02 13:32:32] @ Quit: jbartus: Ping timeout: 248 seconds | ||
| [2011/12/02 13:32:49] <pll> One is a notice() call, the other is an assignment. | ||
| [2011/12/02 13:33:14] @ jbartus joined channel #puppet | ||
| [2011/12/02 13:33:21] <pll> I've got: case $foo { bar : { notice("bar"), $biff = 'baz'}...} | ||
| [2011/12/02 13:33:26] <pll> is that valid syntax ? | ||
| [2011/12/02 13:34:34] @ eshamow joined channel #puppet | ||
| [2011/12/02 13:34:59] <pll> Ah, no , | ||
| [2011/12/02 13:35:01] <dingus9> pll: does bar need 'bar'? also not liking the , after notice | ||
| [2011/12/02 13:36:50] @ Quit: RhysMorgan: Quit: RhysMorgan | ||
| [2011/12/02 13:37:47] <pll> I figured it out, it was the , | ||
| [2011/12/02 13:38:01] @ Quit: maes_`: Read error: Connection reset by peer | ||
| [2011/12/02 13:38:54] <the_ramink> jorhett1: which version on passenger on your lockup? | ||
| [2011/12/02 13:39:46] @ maes_ joined channel #puppet | ||
| [2011/12/02 13:39:59] <jorhett1> rubygem-rake-0.8.7-2.el5 | ||
| [2011/12/02 13:39:59] <jorhett1> rubygems-1.3.1-1.el5 | ||
| [2011/12/02 13:39:59] <jorhett1> rubygem-rack-1.0.1-2.el5 | ||
| [2011/12/02 13:39:59] <jorhett1> rubygem-passenger-2.2.11-3.el5 | ||
| [2011/12/02 13:39:59] <jorhett1> rubygem-fastthread-1.0.7-1.el5 | ||
| [2011/12/02 13:40:25] @ christophermlee joined channel #puppet | ||
| [2011/12/02 13:40:30] <jorhett1> FYI, the system does not lock up, just goes very idle. I'm now alarming to catch this failure because it goes more idle than normal :) | ||
| [2011/12/02 13:40:49] <jorhett1> it seems like the puppet passengers just *stop* somewhere. | ||
| [2011/12/02 13:41:13] <jorhett1> I'm happy to keep diagnosing, just don't know where to look right now. Error logs are quiet. | ||
| [2011/12/02 13:43:29] @ Quit: ghg: Quit: Computer has gone to sleep. | ||
| [2011/12/02 13:45:23] @ d3c joined channel #puppet | ||
| [2011/12/02 13:45:25] <the_ramink> jorhett1: you may want to upgrade to Passenger 3.0.x. Overall I've had less issues with my RoR apps with it though it's odd that it just started. rpms here, http://blog.phusion.nl/2011/11/28/phusion-passenger-3-0-11-released/ | ||
| [2011/12/02 13:45:53] @ Quit: luckman212: Remote host closed the connection | ||
| [2011/12/02 13:48:31] @ jsurfer joined channel #puppet | ||
| [2011/12/02 13:49:07] @ Quit: tangledhelix: Quit: Textual IRC Client: http://www.textualapp.com/ | ||
| [2011/12/02 13:50:34] @ Quit: Kazuhiro: Quit: Leaving. | ||
| [2011/12/02 13:51:44] @ Quit: brtl: Quit: Leaving. | ||
| [2011/12/02 13:52:01] <ccaum> I'm loving Live Management's agent GUI. It's so awesome | ||
| [2011/12/02 13:52:21] @ Quit: illsci: Ping timeout: 244 seconds | ||
| [2011/12/02 13:52:33] <jorhett1> the_ramink: is this safe for ruby 1.8.5 or do I have to upgrade that first? | ||
| [2011/12/02 13:53:00] <jorhett1> I need to build an rpm environment here to build our 1.8.7 rpm soon. I know there's a lot of issues in 1.8.5 | ||
| [2011/12/02 13:54:44] <christophermlee> i am trying to figure out a way to have external data fill in variables in a puppet module. the first use is going to be configuring the additional IP addresses on a single interface. there could be multiple ips per interfaces. ext_lookup does not work as I need more than key/value... looking at hiera but need more examples, I can't grasp it. Anyone have examples of something like this? | ||
| [2011/12/02 13:56:36] @ Quit: jaytaph: Quit: jaytaph | ||
| [2011/12/02 13:58:01] <the_ramink> jorhett1: they have rhel 5 rpms so I assume they'll work without other changes | ||
| [2011/12/02 13:59:35] @ Dr_Memory joined channel #puppet | ||
| [2011/12/02 13:59:39] <Volcane> christophermlee: how do you need more than key/value? | ||
| [2011/12/02 13:59:46] <Dr_Memory> what should I make of this: err: Failed to apply catalog: Parameter ip failed: Invalid IP address | ||
| [2011/12/02 14:00:06] <Dr_Memory> no pointer to a manifest or anything :( | ||
| [2011/12/02 14:00:13] <Volcane> Dr_Memory: look at your host resources | ||
| [2011/12/02 14:00:22] @ ckauhaus joined channel #puppet | ||
| [2011/12/02 14:00:24] <christophermlee> Volcane: I am thinking someting like this for the values I need to import... http://pastie.org/2957087 | ||
| [2011/12/02 14:00:57] <christophermlee> I need to process the data for each interface with a single module | ||
| [2011/12/02 14:01:17] <Volcane> christophermlee: ok - yeah so extlookup is just strings or arrays of strings. hiera lets you store hashes etc | ||
| [2011/12/02 14:02:04] <christophermlee> this is just the first step, next I need to add application information associated with the ip address, and multiple apps may be on one server | ||
| [2011/12/02 14:02:32] <christophermlee> so maybe hiera is the answer but I am having issues grasphing the concept, was hoping to find some more complete examples | ||
| [2011/12/02 14:02:42] <Volcane> nod, thats probably quite complex and in the realms of doing something with an ENC | ||
| [2011/12/02 14:02:55] <Volcane> hiera is just like extlookup | ||
| [2011/12/02 14:03:24] <Volcane> except you can store hashes and choose where to store your data, json or yaml etc | ||
| [2011/12/02 14:03:24] <christophermlee> i was hoping to avoid going to ENC, I am not ready for that yet | ||
| [2011/12/02 14:03:29] <Volcane> the basic concept is the same | ||
| [2011/12/02 14:03:41] <Volcane> you could totally though break your problem down | ||
| [2011/12/02 14:03:45] <Volcane> and say get a list of interfaces | ||
| [2011/12/02 14:03:48] <matchaw_> hi, is anyone familair with https://github.com/deanwilson/puppet-aptsourced ? | ||
| [2011/12/02 14:03:55] <Volcane> then have another piece of data that you lookup based on interfac name | ||
| [2011/12/02 14:04:10] <Volcane> you dont need to think about the data as one blob of data, rather a selection of related information | ||
| [2011/12/02 14:04:22] @ Quit: brandon_dacrib: Ping timeout: 252 seconds | ||
| [2011/12/02 14:05:55] @ dwalu joined channel #puppet | ||
| [2011/12/02 14:06:42] <christophermlee> volcane: makes since, my google foo is failing as I can't find a complete hiera example. I thought there was a blog post about it where I first read about it but can't find now. | ||
| [2011/12/02 14:07:35] <Dr_Memory> Volcane: aha, thanks. a generate() was failing, leading to a Bad Thing getting fed into ip => | ||
| [2011/12/02 14:07:38] <Volcane> christophermlee: there's a good example in the hiera-puppet git repo | ||
| [2011/12/02 14:08:41] @ skippy left channel #puppet () | ||
| [2011/12/02 14:09:50] @ Quit: uphillian: Quit: Leaving. | ||
| [2011/12/02 14:10:03] <d3c> under 2.7.6, I can't my nodes to update using mco puppetd runall n. they'll only run one stage of the manifest and stop. no errors. if I run it via puppet agent -o on the nodes themselves, it works like a charm. changes are picked up. I was told earlier today that upgrading to 2.7.2 would fix it but it hasn't. what could be wrong then? I added ignorecache to both puppetmaster and node configs. | ||
| [2011/12/02 14:10:20] <d3c> 2.7.6, sorry | ||
| [2011/12/02 14:11:04] <Volcane> d3c: no idea what could cause that. | ||
| [2011/12/02 14:12:25] <d3c> Volcane: I have an 'init' stage that runs only 'yum clean all'. that's the only thing that actually succeeds when running via mco puppetd. don't know if that tells you anything though | ||
| [2011/12/02 14:13:19] @ f0 joined channel #puppet | ||
| [2011/12/02 14:13:42] @ Quit: f0: Client Quit | ||
| [2011/12/02 14:14:05] <Volcane> d3c: do you run the puppetd at all or just from mc? | ||
| [2011/12/02 14:14:25] <Volcane> in one case mc will start a new run-once puppet in the other it will just signal the puppetd to do a run | ||
| [2011/12/02 14:15:27] @ Quit: aherzog: Quit: aherzog | ||
| [2011/12/02 14:16:26] <d3c> Volcane: on my puppetmaster, I do mco puppetd runall 2. on any node, I'll look at /var/log/messages to see what it does and it only runs that one stage. I even modified the agent slightly to enable debugging mode but that doesn't show any new info really. | ||
| [2011/12/02 14:19:03] @ jongrocho joined channel #puppet | ||
| [2011/12/02 14:20:11] @ PedroGomes joined channel #puppet | ||
| [2011/12/02 14:20:36] <Dr_Memory> okay, so | ||
| [2011/12/02 14:20:52] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2011/12/02 14:20:52] <Dr_Memory> I'm in the process of (finally!) upgrading from 0.24 to 2.7.6 | ||
| [2011/12/02 14:20:53] @ Quit: PedroGomes: Client Quit | ||
| [2011/12/02 14:20:58] <Dr_Memory> and am running into something weird | ||
| [2011/12/02 14:21:20] <Dr_Memory> every time I make _any_ change to my manifests, I have to run puppet three times before it will actually apply the changes: | ||
| [2011/12/02 14:21:23] <Dr_Memory> http://pastebin.com/GA2cjDGs | ||
| [2011/12/02 14:21:29] <Dr_Memory> and the errors it comes up with the first two times make very little sense | ||
| [2011/12/02 14:21:49] @ windowsrefund joined channel #puppet | ||
| [2011/12/02 14:21:52] <windowsrefund> greets | ||
| [2011/12/02 14:22:21] <Volcane> Dr_Memory: the 2nd is a bad define so show the code for that | ||
| [2011/12/02 14:22:38] @ Quit: kaptk2: Quit: Leaving. | ||
| [2011/12/02 14:22:49] <Volcane> Dr_Memory: the first is possibly an indication that you dont have the correct module layout somewhere and rely on import instead | ||
| [2011/12/02 14:23:36] <Dr_Memory> Volcane: the 2nd is actually a problem I asked about here before. http://pastebin.com/BqXkYQr1 | ||
| [2011/12/02 14:24:08] <Dr_Memory> I can go into some detail if you like, but AFAICT there's nothing wrong with the define. and, as noted, it only gives the error once: the next time I run puppet agent, it works. | ||
| [2011/12/02 14:25:20] <windowsrefund> I'm running into some odd behavior where when I use an exec to run ssh. Despite every attempt and the fact the fact that I can run the exact command manually, my private key is not used. http://pastebin.com/EHb1DAAC | ||
| [2011/12/02 14:25:22] <Volcane> stuff like that only happens cos you either use import or have code node in class { } boundaries tc | ||
| [2011/12/02 14:25:42] <Dr_Memory> well, I definitely use import a lot :) | ||
| [2011/12/02 14:25:50] <Volcane> dont | ||
| [2011/12/02 14:26:06] <Volcane> you should only need import if you store node {} blocks in a file other than site.pp | ||
| [2011/12/02 14:26:09] <windowsrefund> I'd love to see if anyone could reproduce that | ||
| [2011/12/02 14:26:21] <Volcane> if you use it anywhere else you've designed your modules to work against how puppet is designed | ||
| [2011/12/02 14:26:27] <Dr_Memory> root@puppet:/etc/puppet# grep -r import modules/|wc -l | ||
| [2011/12/02 14:26:27] <Volcane> and unsurprisingly that doesnt work well | ||
| [2011/12/02 14:26:28] <Dr_Memory> 24 | ||
| [2011/12/02 14:26:30] <Dr_Memory> sigh | ||
| [2011/12/02 14:26:48] <Dr_Memory> (disclaimer: I inherited this mess from my predecessor) | ||
| [2011/12/02 14:27:17] <d3c> Volcane: if I run the *exact* same command (puppetd --onetime) directly on the node, it works. just tried that. it's only when running via mco from the master that it does not work. any ideas on how to debug this?... | ||
| [2011/12/02 14:27:43] @ emergion joined channel #puppet | ||
| [2011/12/02 14:28:00] <Dr_Memory> that said, site.pp is in fact the only place where there is a node{} block defined | ||
| [2011/12/02 14:28:18] <Volcane> d3c: no, you dont answer the questions I ask so really dont know sorry | ||
| [2011/12/02 14:28:20] @ larstobi joined channel #puppet | ||
| [2011/12/02 14:28:40] @ ghg joined channel #puppet | ||
| [2011/12/02 14:30:01] @ Quit: labrown: Quit: Leaving | ||
| [2011/12/02 14:30:41] <jorhett1> dc3: are you sure you aren't stumbling onto #10418 ? It appears ruby uses different filehandles when run with —test versus in the background, and that might be affecting why —test works. | ||
| [2011/12/02 14:30:41] <gepetto> jorhett1: dc3: #10418 is http://projects.puppetlabs.com/issues/show/10418 "Puppet - Bug #10418: "Caught TERM; calling stop" with state/puppetdlock left in place. It has a status of Investigating and is assigned to -" | ||
| [2011/12/02 14:31:35] <nlew> windowsrefund: Does it work if you manually run it with "su akosmin -c /usr/bin/ssh [...]" ? | ||
| [2011/12/02 14:33:17] <eric0> Dr_Memory: damn, still getting that 'could not find class user::deploy' eh? | ||
| [2011/12/02 14:33:37] <windowsrefund> nlew: the manual command works | ||
| [2011/12/02 14:34:53] <Dr_Memory> eric0: yeah. although at least this time it's transient, so, uh, progress? | ||
| [2011/12/02 14:34:54] <d3c> jorhett1: that might be. will look into it, thanks | ||
| [2011/12/02 14:35:07] @ Av is now known as avleen | ||
| [2011/12/02 14:35:50] @ Quit: alanna: Remote host closed the connection | ||
| [2011/12/02 14:36:31] <eric0> Dr_Memory: so freaky. you using webrick or a rack/webserver stack? | ||
| [2011/12/02 14:36:44] <Dr_Memory> webrick. | ||
| [2011/12/02 14:36:46] @ Quit: mcarpent: Quit: Thanks for the fish. | ||
| [2011/12/02 14:36:49] @ sascha_ joined channel #puppet | ||
| [2011/12/02 14:36:51] <Dr_Memory> which is another thing I need to fix, but one disaster at a time :) | ||
| [2011/12/02 14:36:58] <Volcane> eric0: he also likes import'ing things - all bets are off :) | ||
| [2011/12/02 14:37:28] <Dr_Memory> I didn't say I liked it, just that I did it. :) | ||
| [2011/12/02 14:37:36] <Dr_Memory> (and FSVO "I") | ||
| [2011/12/02 14:38:00] <Volcane> so no doubt in the process of removing al the import and stuff and getting it to just work | ||
| [2011/12/02 14:38:04] <Dr_Memory> btw, IIRC the official puppetmaster training suggests using "import *" in class files. :( | ||
| [2011/12/02 14:38:06] <eric0> aye yep | ||
| [2011/12/02 14:38:07] <Volcane> you'll at least get the bug to behave predictably | ||
| [2011/12/02 14:38:43] <Dr_Memory> anyway, yeah, trying to see how much of this I can just rip out without my stuff blowing up | ||
| [2011/12/02 14:38:45] <windowsrefund> I use import in site.pp | ||
| [2011/12/02 14:38:51] <windowsrefund> is that wrong now? | ||
| [2011/12/02 14:39:07] <agaffney> windowsrefund: that's about the only place that it's still "allowed" | ||
| [2011/12/02 14:39:12] <windowsrefund> oh | ||
| [2011/12/02 14:39:14] <windowsrefund> ok | ||
| [2011/12/02 14:39:21] <windowsrefund> that's the only place I've bothered with it | ||
| [2011/12/02 14:39:25] <agaffney> there's no other way to pull in a separate nodes.pp or the like | ||
| [2011/12/02 14:39:30] <windowsrefund> right | ||
| [2011/12/02 14:39:39] @ Quit: dcarley: Ping timeout: 245 seconds | ||
| [2011/12/02 14:39:40] <agaffney> but with modules, import is unnecessary | ||
| [2011/12/02 14:39:42] * Volcane did point out nodes outside of site.pp should be the only use :P | ||
| [2011/12/02 14:39:51] <agaffney> as long as your structure it correctly for the autoloader | ||
| [2011/12/02 14:39:54] <eric0> but really if you feel like that stuff is complicated enough you want to split it out, spend the time to simplify it insetad | ||
| [2011/12/02 14:40:34] @ Quit: blentz: Ping timeout: 244 seconds | ||
| [2011/12/02 14:41:01] @ Quit: tremble: Ping timeout: 240 seconds | ||
| [2011/12/02 14:42:36] <dwalu> hrm. is it possible to capture a regex in a node definition and use it? | ||
| [2011/12/02 14:43:06] @ Quit: asteve: Quit: asteve | ||
| [2011/12/02 14:43:55] <dwalu> like node /^dwalu-\(\d+\.\d+\) { $version = $1} | ||
| [2011/12/02 14:43:55] <Dr_Memory> huh. so the one thing that's really being include'ed is a module called "common", which seems to have a bunch of parser plugins | ||
| [2011/12/02 14:43:55] <dwalu> ? | ||
| [2011/12/02 14:44:10] <Dr_Memory> in one of them, I see: | ||
| [2011/12/02 14:44:12] <Dr_Memory> # For further help contact Volcane on #puppet | ||
| [2011/12/02 14:44:13] <Dr_Memory> :) | ||
| [2011/12/02 14:44:20] <Volcane> lols | ||
| [2011/12/02 14:44:25] <Dr_Memory> j'accuse! | ||
| [2011/12/02 14:44:26] @ Quit: lborda: Quit: Ex-Chat | ||
| [2011/12/02 14:46:20] <eric0> # For further help leave a briefcase full of cash in the LHR terminal B mens room | ||
| [2011/12/02 14:46:56] <Dr_Memory> I'll get right on that! | ||
| [2011/12/02 14:47:21] <Dr_Memory> okay, looks like half of this stuff is just here for my predecessor's crack-addled oncall rotator script | ||
| [2011/12/02 14:47:25] <Dr_Memory> delete delete delete delete | ||
| [2011/12/02 14:47:39] <dwalu> dr_memory: pagerduty.com for the win? | ||
| [2011/12/02 14:49:15] @ tinhead joined channel #puppet | ||
| [2011/12/02 14:49:26] @ Quit: tinhead: Changing host | ||
| [2011/12/02 14:49:26] @ tinhead joined channel #puppet | ||
| [2011/12/02 14:49:37] <the_ramink> dwalu: you'd probably want to do a case statement around fqdn w/ regex because you can only match node once if I'm understanding things correctly. | ||
| [2011/12/02 14:50:43] @ kelseyhi1htower joined channel #puppet | ||
| [2011/12/02 14:50:47] @ Quit: sascha_: Quit: sascha_ | ||
| [2011/12/02 14:51:12] @ nadavoid joined channel #puppet | ||
| [2011/12/02 14:52:21] @ Quit: tblamer: Quit: Going offline, see ya! (www.adiirc.com) | ||
| [2011/12/02 14:52:59] @ Quit: kelseyhightower: Ping timeout: 245 seconds | ||
| [2011/12/02 14:53:25] <tinhead> when using the zookeeper module here https://github.com/hstack/puppet.git, this manifest http://pastebin.com/WRMK3x6f keeps throwing this error http://pastebin.com/trLHXKK7 | ||
| [2011/12/02 14:53:42] <dwalu> the_ramink: Maybe that might not be so bad in this case? What I want to do is write a script which takes a version of our software as a parameter and then passes that into a userdata script for provisioning an ec2 node, and that version will get passed into the certname/nodename paramters for puppet agent when it's run .. then puppet should be able to strip that version and pass it into my classes, thus letting developers bring u | ||
| [2011/12/02 14:53:43] <dwalu> staging servers at will | ||
| [2011/12/02 14:54:01] <Dr_Memory> dwalu: got it in one | ||
| [2011/12/02 14:54:39] @ Quit: kelseyhi1htower: Client Quit | ||
| [2011/12/02 14:54:43] <eric0> dwalu: i do not think node regxp support captures/backreferences | ||
| [2011/12/02 14:54:50] @ kelseyhightower joined channel #puppet | ||
| [2011/12/02 14:55:05] @ nul0op joined channel #puppet | ||
| [2011/12/02 14:56:06] <nlew> dwalu: eric0: I believe there's an (old) open ticket to set $1 in regex nodes | ||
| [2011/12/02 14:56:22] <dwalu> @nlew: heh I think I just found it. http://projects.puppetlabs.com/issues/11133 ? | ||
| [2011/12/02 14:56:37] <nlew> Other than that, you might get $name set in the node block to the part that matched (but no captures) | ||
| [2011/12/02 14:56:41] <tinhead> there is, however, a file resource zookeeper_home defined in init.pp, any pointers? | ||
| [2011/12/02 14:57:30] <nlew> dwalu: Ah well that too. I was thinking of #2628 heh | ||
| [2011/12/02 14:57:31] <gepetto> nlew: dwalu: #2628 is http://projects.puppetlabs.com/issues/show/2628 "Puppet - Feature #2628: It would be useful if node name regexps set $1. It has a status of Accepted and is assigned to Brice Figureau" | ||
| [2011/12/02 14:57:42] <dwalu> fail | ||
| [2011/12/02 14:58:45] <torrancew> tinhead: that is not resource named zookeeper_home, that is a file located at the path stored in the *value* of zookeeper_home | ||
| [2011/12/02 14:59:00] @ Quit: i0n: Ping timeout: 244 seconds | ||
| [2011/12/02 14:59:16] <tinhead> torrancew: err, so it should not have been referenced in the erb? | ||
| [2011/12/02 14:59:28] <dwalu> that would have been super slick too | ||
| [2011/12/02 15:00:20] <tinhead> torrancew: nm, I declared zookeeper_home and now it complains about file paths that are not fully qualified | ||
| [2011/12/02 15:00:58] <dwalu> what does node actually match against anyways? does it match against fqdn or certname or hostname? | ||
| [2011/12/02 15:01:29] <torrancew> tinhead: you'll have to define a few parameters for zookeeper, it seems | ||
| [2011/12/02 15:01:34] <dwalu> nevermind, found it in the docs | ||
| [2011/12/02 15:02:12] <torrancew> tinhead: likely anything it lists as "parameters" in the comments of zookeeper/manifests/init.pp should be pre-declared | ||
| [2011/12/02 15:02:25] <dwalu> thanks all, and whoever suggested the case, that makes the most sense | ||
| [2011/12/02 15:02:47] @ newtMcKerr joined channel #puppet | ||
| [2011/12/02 15:03:03] <tinhead> torrancew: oh, the parameters mentioned in comments on init.pp … let me do all that and get back with my results | ||
| [2011/12/02 15:04:38] <Dr_Memory> okay, I think I'm down to the last set of import's that need attention, and it's the thorniest: | ||
| [2011/12/02 15:05:08] <Dr_Memory> ls modules/user/manifests/ | ||
| [2011/12/02 15:05:09] <Dr_Memory> core-groups.ppec2-users.ppinit.ppministry.pprepo.pprsync.ppvirtual.pp | ||
| [2011/12/02 15:05:14] <Dr_Memory> (and a bunch more) | ||
| [2011/12/02 15:05:20] @ Kazuhiro joined channel #puppet | ||
| [2011/12/02 15:05:24] <Dr_Memory> init.pp in that directory, of course, just calls "import '*'" | ||
| [2011/12/02 15:05:43] @ Kazuhiro1 joined channel #puppet | ||
| [2011/12/02 15:05:45] <Dr_Memory> should I just concat all of those files into init.pp and be done with it, or what? | ||
| [2011/12/02 15:05:45] <Volcane> and core-groups.pp doesnt have users::core-groups ? (you should avoid - in anything) | ||
| [2011/12/02 15:05:48] <jorhett1> the_ramink: interesting. I believe what's going on is that puppet passenger sessions are hanging, because the active session count keeps climbing but the number of clients processed don't change on some of them for excessively long times. So it's not a sudden hang, it's a slow acclimation until all 20 are hung. | ||
| [2011/12/02 15:06:07] <Dr_Memory> $ head -1 modules/user/manifests/core-groups.pp | ||
| [2011/12/02 15:06:08] <Dr_Memory> class user::core-groups { | ||
| [2011/12/02 15:06:28] <Volcane> then u dont need import for that class | ||
| [2011/12/02 15:06:43] <Volcane> when you say include user::core-groups puppet will go look for it in the right place on demand | ||
| [2011/12/02 15:06:55] @ Quit: Kazuhiro: Read error: No route to host | ||
| [2011/12/02 15:07:01] @ Quit: d3c: Ping timeout: 244 seconds | ||
| [2011/12/02 15:07:03] <Dr_Memory> hm, let me see if that's actually what happens | ||
| [2011/12/02 15:07:49] * Dr_Memory is quickly lost in a maze of include directives, all slightly different | ||
| [2011/12/02 15:07:52] <Dr_Memory> but... I think this works | ||
| [2011/12/02 15:08:00] @ Quit: rharrison: Quit: Leaving | ||
| [2011/12/02 15:08:06] <eric0> you'd just need to import those things if they didn't have all their code inside a class foo::bar { } stanza, where the file is named bar.pp | ||
| [2011/12/02 15:08:30] @ Quit: jsm: Quit: Leaving. | ||
| [2011/12/02 15:08:52] <Dr_Memory> ah, but the parser does that for me automatically? | ||
| [2011/12/02 15:09:00] <Volcane> yup | ||
| [2011/12/02 15:09:01] <dwalu> ooh, one more question. if I start puppet on a new node with puppet agent --node_name="dwalu", and I wanted to match gainst that fact in a nodes definition .. can I? | ||
| [2011/12/02 15:09:14] <Dr_Memory> huh. did it not used to do that, or was my predecessor just an idiot? | ||
| [2011/12/02 15:09:41] <Volcane> Dr_Memory: in the old days it didnt always no, now it does and it doesnt like if you think you're more clever than it | ||
| [2011/12/02 15:10:01] <Dr_Memory> (I should be less of an asshole: dude was a crackerjack coder, just kinda in over his head here) | ||
| [2011/12/02 15:10:08] <Dr_Memory> (where "here" = ops stuff) | ||
| [2011/12/02 15:10:14] <dwalu> where my "node name" is definitely going to be different than my hostname or fqdn, because amazon is a pain | ||
| [2011/12/02 15:10:16] <Volcane> Dr_Memory: mostly not using import and having it all works generally tell us you followed the basic best approach and so when we look at errors u get we can know its not down to some stupidity related to that | ||
| [2011/12/02 15:10:29] @ Quit: the_ramink: Read error: Connection reset by peer | ||
| [2011/12/02 15:10:49] @ the_ramink joined channel #puppet | ||
| [2011/12/02 15:11:05] <Volcane> Dr_Memory: people tend to import for all the wrong reasons - they did something stupid and think thatll fix it, generally thats naive cos of how importing relates to multiple runs | ||
| [2011/12/02 15:13:07] @ Quit: the_ramink: Read error: Connection reset by peer | ||
| [2011/12/02 15:13:08] @ Quit: kermit: Quit: Leaving. | ||
| [2011/12/02 15:13:23] @ kermit joined channel #puppet | ||
| [2011/12/02 15:13:57] @ the_ramink joined channel #puppet | ||
| [2011/12/02 15:14:22] @ Quit: ccaum: Remote host closed the connection | ||
| [2011/12/02 15:14:54] @ kelseyhi1htower joined channel #puppet | ||
| [2011/12/02 15:15:33] @ askreet joined channel #puppet | ||
| [2011/12/02 15:15:38] <the_ramink> jorhett1: interesting... I can't say I've ever seen anything like that. | ||
| [2011/12/02 15:15:56] <eric0> dwalu: you need to change the server 'node_name = facter', see https://gist.github.com/1182243 | ||
| [2011/12/02 15:16:04] <Dr_Memory> woohoo, I think my importectomy worked. | ||
| [2011/12/02 15:16:46] <askreet> So I have a cluster of systems that should receive a file push from a central system. Today we maintain a list of servers in a text file that is then read by an rsync script. I'm trying to find a way to do something similiar in Puppet and I'm coming up with exported/collected resources but I'd like to collect "lines in a file" -- am I missing something that makes generating this list easy? | ||
| [2011/12/02 15:17:03] <Volcane> Dr_Memory: restart the master completely to be 100% sure | ||
| [2011/12/02 15:17:35] @ Quit: kelseyhightower: Ping timeout: 255 seconds | ||
| [2011/12/02 15:18:14] <Dr_Memory> good point | ||
| [2011/12/02 15:18:51] <dwalu> eric0: if I make that node_name = facter change, will the rest of my nodes which are just hostname based, will they be screwed? | ||
| [2011/12/02 15:20:33] @ Quit: rutger_: Read error: Operation timed out | ||
| [2011/12/02 15:20:33] <Dr_Memory> grrrrr | ||
| [2011/12/02 15:20:40] <Dr_Memory> the user::deploy errors are gone | ||
| [2011/12/02 15:20:44] <Dr_Memory> but still getting this: | ||
| [2011/12/02 15:20:48] <Dr_Memory> err: Could not retrieve catalog from remote server: Error 400 on SERVER: Duplicate definition: File[undef] is already defined in file /etc/puppet/modules/apache2/manifests/init.pp at line 72; cannot redefine at /etc/puppet/modules/apache2/manifests/init.pp:80 on node puppet | ||
| [2011/12/02 15:21:15] <dwalu> eric0: or is it enough for me to make that change, and just on my new nodes to tell the agent to set node_name_fact = fqdn? | ||
| [2011/12/02 15:21:42] <Volcane> Dr_Memory: you have something like file{$foo: ... } and $foo is empty | ||
| [2011/12/02 15:22:13] <Dr_Memory> that's the thing, it's not, or at least I think it should not be: | ||
| [2011/12/02 15:22:14] <Dr_Memory> http://pastebin.com/EMBHGHpc | ||
| [2011/12/02 15:22:36] <Dr_Memory> note lines 11-14 versus lines 72 and 80 | ||
| [2011/12/02 15:23:45] <Volcane> u have stuff outside of class {} | ||
| [2011/12/02 15:23:47] <Volcane> dont do that | ||
| [2011/12/02 15:23:56] <Volcane> this is why idiot before you used import :) | ||
| [2011/12/02 15:24:03] <Volcane> cos it doesnt work like u think it does | ||
| [2011/12/02 15:24:57] <Dr_Memory> hm. so what's the issue: the base-scope variables don't necessarily get instantiated before the classes? | ||
| [2011/12/02 15:25:23] <Volcane> if u want anything outside a class, define or node | ||
| [2011/12/02 15:25:25] <Volcane> it has to be in site.pp | ||
| [2011/12/02 15:25:26] <eric0> dwalu: if you have nodes where the certificate name and the fqdn are the same, they will be OK without a change | ||
| [2011/12/02 15:25:46] <Volcane> else it will only work once and then not work again - on every puppet master start | ||
| [2011/12/02 15:25:59] <dwalu> eric0: Sweet, thanks! | ||
| [2011/12/02 15:26:07] * dwalu will out-clever himself, or will die trying. | ||
| [2011/12/02 15:27:24] <tinhead> torrancew: progressed to a new error: 'tar xzf zookeeper-3.4.0.tar.gz;' is not qualified and no path was specified. Please qualify the command or specify a path. | ||
| [2011/12/02 15:28:42] <torrancew> tinhead: you can add something like this: path => ['/usr/local/bin', '/usr/local/sbin', '/usr/bin', '/usr/sbin', '/bin', '/sbin' ] | ||
| [2011/12/02 15:28:46] <torrancew> to the exec | ||
| [2011/12/02 15:28:51] <eric0> tinhead: wow that is a very informative error, which tells you exactly what to do to fix it... | ||
| [2011/12/02 15:28:52] <torrancew> or you can fully qualify tar | ||
| [2011/12/02 15:29:19] <Dr_Memory> hm. moving the $apache_sites etc definitions into class apache2 has some interesting side-effects: | ||
| [2011/12/02 15:29:23] <Dr_Memory> err: Failed to apply catalog: Parameter path failed: File paths must be fully qualified, not '-available/nagios' at /etc/puppet/modules/apache2/manifests/init.pp:191 | ||
| [2011/12/02 15:29:45] <Volcane> Dr_Memory: yup, thats 2.7.x its forcing you to not do this kind of dumb thing :) | ||
| [2011/12/02 15:29:45] <tinhead> torrancew: aah, thanks. will be back with results after some more fiddling. | ||
| [2011/12/02 15:29:45] <joe-mac> you need to scope your variables | ||
| [2011/12/02 15:30:07] <Volcane> Dr_Memory: either put it in some class and reference it as $foo::bar which is $bar in class foo | ||
| [2011/12/02 15:30:10] @ Quit: alanevans: Remote host closed the connection | ||
| [2011/12/02 15:30:16] <Volcane> Dr_Memory: or pull it out into extlookup or something | ||
| [2011/12/02 15:31:53] <Dr_Memory> if I fix this for 2.7, will it blow up in 0.24? | ||
| [2011/12/02 15:32:08] <Dr_Memory> (working on 0.24 is not necessarily a requirement, but it'd be nice to be able to stage the rollout a bit) | ||
| [2011/12/02 15:32:50] <tinhead> torrancew: thanks for your help. now I am in an entirely different mess with my puppet server not serving tarballs, but that helped a lot. the zookeeper module does not some parameters not mentioned in that file, but I seem to have put that behind me now. | ||
| [2011/12/02 15:33:27] <joe-mac> anybody here have a response to my post about my forked sshkey type/provider posted on -dev? | ||
| [2011/12/02 15:33:41] <joe-mac> I really, really wanna finish that | ||
| [2011/12/02 15:34:54] @ illsci joined channel #puppet | ||
| [2011/12/02 15:35:11] <Volcane> Dr_Memory: shoud be ok but its too long ago now to remember the exact set of 0.24 bugs :P | ||
| [2011/12/02 15:35:45] <eric0> all the variable interpolation / manifest parsing stuff is done on the server though, so its not like the client is going to care | ||
| [2011/12/02 15:36:08] @ Quit: askreet: Quit: leaving | ||
| [2011/12/02 15:36:11] <Dr_Memory> eric0: my hope is to push a cleaned-up module set to all of my servers, and then roll through them and convert to 2.7 | ||
| [2011/12/02 15:36:14] <Volcane> yeah, but the 24 server didnt have all the auto import magic quite right :) | ||
| [2011/12/02 15:36:36] <Dr_Memory> but if that's not possible, that's not possible | ||
| [2011/12/02 15:37:45] <eric0> if by 'roll through' you mean 'update them all within one runinterval', i think thats fine. i really wouldnt want to run a mix of .24 and 2.7 servers for longer than that | ||
| [2011/12/02 15:37:50] <Dr_Memory> okay, fully qualifying all of those variables (${apache2::foo}) seems to be effective | ||
| [2011/12/02 15:37:55] <the_ramink> Dr_Memory: if you have a 2.7 master with a mix of clients you should be okay. | ||
| [2011/12/02 15:38:06] <Volcane> 24 cant talk to 2.7 master | ||
| [2011/12/02 15:38:35] <Dr_Memory> eric0: oh, heh, sorry, should have been clearer. for any given client, there's only one server. | ||
| [2011/12/02 15:39:04] <Dr_Memory> Volcane: orly? | ||
| [2011/12/02 15:39:10] <the_ramink> I didn't realize that... damn. | ||
| [2011/12/02 15:39:13] <Dr_Memory> that's... useful to know. | ||
| [2011/12/02 15:39:18] <Volcane> i believe so | ||
| [2011/12/02 15:39:35] <Volcane> i think the old XML-RPC interface is gone gone now, might be wrong. | ||
| [2011/12/02 15:39:37] <eric0> oh, nomore xmlrpc | ||
| [2011/12/02 15:39:39] <Volcane> but def o test it | ||
| [2011/12/02 15:42:22] @ Quit: ghg: Quit: Out of here. | ||
| [2011/12/02 15:44:33] @ Quit: kinsel: Ping timeout: 260 seconds | ||
| [2011/12/02 15:45:34] @ d3c joined channel #puppet | ||
| [2011/12/02 15:45:45] @ Quit: setient: Quit: leaving | ||
| [2011/12/02 15:46:04] @ kinsel joined channel #puppet | ||
| [2011/12/02 15:47:21] @ Quit: alfism: Quit: alfism | ||
| [2011/12/02 15:47:32] @ Quit: hobbeswalsh: Quit: hobbeswalsh | ||
| [2011/12/02 15:50:45] <Dr_Memory> hm: err: Could not retrieve catalog: Could not find class user::repo::add-private-key-to-root in namespaces apt-repo::client at /etc/puppet/modules/apt-repo/manifests/init.pp:99 on node ec2 | ||
| [2011/12/02 15:50:50] @ Quit: newtMcKerr: Quit: Leaving. | ||
| [2011/12/02 15:51:11] <Dr_Memory> but "class user::repo::add-private-key-to-root" exists in modules/user/manifests/repo.pp | ||
| [2011/12/02 15:51:20] <Volcane> so in the wrong place in other words? | ||
| [2011/12/02 15:52:09] <Dr_Memory> er? | ||
| [2011/12/02 15:52:21] <Volcane> if foo::bar goes in bar.pp | ||
| [2011/12/02 15:52:25] <Volcane> then..... | ||
| [2011/12/02 15:52:42] <Dr_Memory> ah, modules/user/manifests/repo/addblah.pp ? | ||
| [2011/12/02 15:52:51] <Volcane> bingo | ||
| [2011/12/02 15:52:55] <Dr_Memory> le sigh | ||
| [2011/12/02 15:54:24] @ Quit: emergion: Quit: later... | ||
| [2011/12/02 15:54:38] @ astrostl joined channel #puppet | ||
| [2011/12/02 15:56:23] <eric0> not sure if this makes it any more fun, but you are paying down principal on your site's technical debt | ||
| [2011/12/02 15:56:42] @ Quit: mgw: Quit: Leaving. | ||
| [2011/12/02 15:58:14] <Dr_Memory> yes indeedy | ||
| [2011/12/02 15:59:01] <Volcane> yeah, just look how suddenly people can immediately help you within about a minute of showing your error, just cos your site now works like all others | ||
| [2011/12/02 15:59:10] <Volcane> and suddenly docs apply to you | ||
| [2011/12/02 15:59:18] <Volcane> and references of what errors are do too etc | ||
| [2011/12/02 15:59:24] @ Quit: choonming: Read error: Connection timed out | ||
| [2011/12/02 16:01:21] @ sascha_ joined channel #puppet | ||
| [2011/12/02 16:02:02] <eric0> holy shit, ora safari is so spammy and awful. i havent used it in a couple of years and its somehow gotten worse | ||
| [2011/12/02 16:02:35] @ Quit: christophermlee: Ping timeout: 265 seconds | ||
| [2011/12/02 16:04:52] @ Quit: neonflux: Quit: Leaving. | ||
| [2011/12/02 16:05:27] @ Quit: d3c: Ping timeout: 240 seconds | ||
| [2011/12/02 16:07:01] @ Quit: spiette: Ping timeout: 240 seconds | ||
| [2011/12/02 16:07:21] @ labrown joined channel #puppet | ||
| [2011/12/02 16:07:30] @ Quit: labrown: Read error: Connection reset by peer | ||
| [2011/12/02 16:08:18] @ Quit: nfagerlund: Remote host closed the connection | ||
| [2011/12/02 16:08:33] @ hobbeswalsh joined channel #puppet | ||
| [2011/12/02 16:09:42] @ Quit: hobbeswalsh: Client Quit | ||
| [2011/12/02 16:10:31] @ Quit: lwhalen: Quit: Leaving | ||
| [2011/12/02 16:10:50] @ lwhalen joined channel #puppet | ||
| [2011/12/02 16:10:59] <dingus9> ora safari? | ||
| [2011/12/02 16:11:12] @ Quit: lwhalen: Client Quit | ||
| [2011/12/02 16:11:46] @ lwhalen joined channel #puppet | ||
| [2011/12/02 16:12:37] @ nfagerlund joined channel #puppet | ||
| [2011/12/02 16:15:45] @ newtMcKerr joined channel #puppet | ||
| [2011/12/02 16:17:10] @ Djelibeybi joined channel #puppet | ||
| [2011/12/02 16:18:01] <eric0> oreilly safaribookshelf | ||
| [2011/12/02 16:18:09] @ Quit: DJClean: Ping timeout: 260 seconds | ||
| [2011/12/02 16:19:10] @ alanevans joined channel #puppet | ||
| [2011/12/02 16:19:33] @ Quit: lwhalen: Quit: Leaving | ||
| [2011/12/02 16:21:54] @ Quit: ali3n0: Remote host closed the connection | ||
| [2011/12/02 16:24:15] @ Quit: verwilst: Ping timeout: 244 seconds | ||
| [2011/12/02 16:25:49] <mfridh> if i'm using run stages will puppet re-evalute, for example, a $sudoversion = generate('rpm -q --qf="%{version}"') in subsequent stages? | ||
| [2011/12/02 16:26:28] <Volcane> no | ||
| [2011/12/02 16:26:31] <mfridh> ok, thanks. | ||
| [2011/12/02 16:26:34] @ rubic joined channel #puppet | ||
| [2011/12/02 16:26:35] <jorhett1> mfridh: there are several open bugs that would suggest not. | ||
| [2011/12/02 16:26:51] <mfridh> I found out, after loosing root access, that I used unsupported syntax in sudoers :d | ||
| [2011/12/02 16:26:52] <Volcane> its not even a bug, its basic design | ||
| [2011/12/02 16:26:55] <mfridh> or rather, unsupported option | ||
| [2011/12/02 16:26:58] <jorhett1> in my experience puppet re-evaluates nothing, including nsswitch.conf changes not appearing until the puppet client is restarted | ||
| [2011/12/02 16:27:29] <mfridh> need sudo > 1.7.x for that specific option... wanted to make sure sudo is upgraded before putting the files in place. | ||
| [2011/12/02 16:27:34] <mfridh> does that make sense? | ||
| [2011/12/02 16:27:57] <Volcane> just set a require to the package | ||
| [2011/12/02 16:28:11] <Volcane> file{"sudoers": ..., require => Package["sudo"]} | ||
| [2011/12/02 16:28:25] <Volcane> and make sure package{"sudo": ensure => the.right.version} | ||
| [2011/12/02 16:28:33] <mfridh> yes but I can't use ensure => latest, only a specific version. | ||
| [2011/12/02 16:29:05] <mfridh> I guess that's a better compromise than requireing _two_ puppet runs for sudo to be setup :D | ||
| [2011/12/02 16:30:40] <leifmadsen> hey all, I'm setting up mcollective, and have some stuff working (I can do 'mco inventory <hostname>' and it returns results). But when I try to do 'mc-ping' I get no response, and on the server, I see the following in the logs (I've enabled debug level logging): http://pastebin.com/HyjTrQuS | ||
| [2011/12/02 16:30:56] <leifmadsen> from my google searching, I'm starting to think I'm missing some sort of filter on the servers | ||
| [2011/12/02 16:31:05] @ Quit: nfagerlund: Remote host closed the connection | ||
| [2011/12/02 16:31:35] <Volcane> leifmadsen: do you have discover.rb in the agents dir ? | ||
| [2011/12/02 16:31:41] <leifmadsen> yes, I did see that one | ||
| [2011/12/02 16:32:13] <leifmadsen> in /usr/share/mcollective/plugins/mcollective/agent/ | ||
| [2011/12/02 16:32:25] @ Quit: sduchesneau: Remote host closed the connection | ||
| [2011/12/02 16:32:28] <Volcane> and the inventory output shows it listed under agents? | ||
| [2011/12/02 16:32:31] @ Quit: nadavoid: Remote host closed the connection | ||
| [2011/12/02 16:32:44] <leifmadsen> yep, I have discovery and rpcutil | ||
| [2011/12/02 16:33:04] <Volcane> weird then, thats the whole log no lines removed? | ||
| [2011/12/02 16:33:21] <leifmadsen> yep, that's everything bit for bit after an mc-ping from the client | ||
| [2011/12/02 16:33:42] <Volcane> all the same version client and server? | ||
| [2011/12/02 16:33:56] <leifmadsen> let me double check, but they should all be 1.3.2 | ||
| [2011/12/02 16:34:05] <leifmadsen> but you never know what package management | ||
| [2011/12/02 16:34:11] <dwalu> when I run puppet agent with --node_name="somenode" should the puppetmaster then see the cert as "somenode" ? | ||
| [2011/12/02 16:34:21] <Volcane> if you're just getting going you should aim for 1.2.1 - thats hte prod code | ||
| [2011/12/02 16:34:26] <Volcane> but 1.3.2 shoud work :) | ||
| [2011/12/02 16:35:33] <leifmadsen> aha, I have a version mismatch | ||
| [2011/12/02 16:35:42] <leifmadsen> thanks for the thought | ||
| [2011/12/02 16:35:52] * leifmadsen waves his fist at dpkg | ||
| [2011/12/02 16:36:19] <Volcane> ok between 1.2.x and 1.3.x it doesnt work - i do keep the minor versions compatible | ||
| [2011/12/02 16:37:06] <leifmadsen> Volcane: yep, I was trying to install 1.3.2 all over with dpkg, but it looks like something on the client got installed incorrectly | ||
| [2011/12/02 16:37:12] <leifmadsen> I'll giv'r a shot again :) | ||
| [2011/12/02 16:37:19] <leifmadsen> thanks again, I'm sure that's my problem | ||
| [2011/12/02 16:37:23] <leifmadsen> otherwise, configuration wasn't so bad | ||
| [2011/12/02 16:37:57] <Volcane> kewl, ok gotta dash. | ||
| [2011/12/02 16:38:14] @ AK6L joined channel #puppet | ||
| [2011/12/02 16:38:27] <AK6L> anyone running Puppet Enterprise on CentOS 5.7? | ||
| [2011/12/02 16:38:54] <AK6L> i have a bunch of machines I just upgraded to 5.7; Puppet starts, runs once, and then never runs again | ||
| [2011/12/02 16:39:06] <AK6L> it gets stuck in some state where i thinks a run is always in progress. | ||
| [2011/12/02 16:39:11] <AK6L> not having much luck diagnosing it. | ||
| [2011/12/02 16:39:35] <haus> AK6L: what version of PE? | ||
| [2011/12/02 16:40:04] <AK6L> 1.2 | ||
| [2011/12/02 16:40:23] <haus> is cron(d) enabled? | ||
| [2011/12/02 16:40:40] <AK6L> yes, but i run puppet as a daemon, not under cron | ||
| [2011/12/02 16:40:50] <AK6L> are you suggesting cron as the solution? :) | ||
| [2011/12/02 16:41:08] <leifmadsen> Volcane: thanks again -- that was exactly the problem -- mco ping works great now | ||
| [2011/12/02 16:41:17] @ Quit: the_ramink: Ping timeout: 255 seconds | ||
| [2011/12/02 16:41:19] <AK6L> i bet it would fix the problem, but it's a drag that this upgrade causes this behavior | ||
| [2011/12/02 16:41:57] <haus> for sure | ||
| [2011/12/02 16:42:09] <haus> was it a 5.6=>5.7 upgrade? | ||
| [2011/12/02 16:42:15] <AK6L> yup | ||
| [2011/12/02 16:42:42] <haus> i'll try to replicate it and see if i can figure out what's going on | ||
| [2011/12/02 16:42:46] <AK6L> awesome | ||
| [2011/12/02 16:43:20] <AK6L> haus: if it helps at all, the stracing i've been doing seems to indicate it's trying to get a catalog and failing, but restarting and running puppet agent --test works | ||
| [2011/12/02 16:43:40] <AK6L> haus: there's also some indication that there are other problems caused by this upgrade, possibly around certificate verification | ||
| [2011/12/02 16:44:14] <haus> AK6L: if you open a bug at http://projects.puppetlabs.com/projects/puppet-enterprise i'll update it if i figure anything out | ||
| [2011/12/02 16:44:35] @ Quit: astrostl: Ping timeout: 244 seconds | ||
| [2011/12/02 16:44:43] <haus> since you have a good handle on what's going on | ||
| [2011/12/02 16:49:59] <jorhett1> AK6L: you probably got a compromised kernel. See #10418 | ||
| [2011/12/02 16:49:59] <gepetto> jorhett1: AK6L: #10418 is http://projects.puppetlabs.com/issues/show/10418 "Puppet - Bug #10418: "Caught TERM; calling stop" with state/puppetdlock left in place. It has a status of Investigating and is assigned to -" | ||
| [2011/12/02 16:50:20] <jorhett1> backdate your 5.7 kernels to 274.3.1 and the problem will disappear. | ||
| [2011/12/02 16:50:39] <AK6L> jorhett1: "compromised"? as in buggy, or has trojan code? | ||
| [2011/12/02 16:51:09] <jorhett1> compromised as in "they changed how calls to proc return" and this broke things like ruby that expect select() to work. | ||
| [2011/12/02 16:51:15] <AK6L> jorhett1: i'm using 2.6.18-274.7.1.el5 | ||
| [2011/12/02 16:51:29] <jorhett1> yep, that's the one. backdate to 274.3.1 and you'll be fine | ||
| [2011/12/02 16:51:32] <AK6L> ahhhhh | ||
| [2011/12/02 16:51:37] <AK6L> damn, i never would have found that. | ||
| [2011/12/02 16:52:16] <jorhett1> there's a redhat bug, they have fixed kernels — not sure when it will flow downstream into normal release | ||
| [2011/12/02 16:52:26] <AK6L> nod. | ||
| [2011/12/02 16:52:27] <jorhett1> redhat bug details in that ticket above | ||
| [2011/12/02 16:52:30] @ Quit: althara: Quit: Leaving. | ||
| [2011/12/02 16:54:07] <AK6L> hmm, redhat bugzilla wants a login. | ||
| [2011/12/02 16:54:50] <haus> yea, i couldn't view the bug report either | ||
| [2011/12/02 16:56:01] <AK6L> yeah even with a login it doesn't let me view the report. | ||
| [2011/12/02 16:56:06] <jorhett1> yeah, wish they wouldn't do that. I copied the only useful text out of there into the puppetlabs bug | ||
| [2011/12/02 16:56:14] <AK6L> time to work the social network for an authorized account.[ | ||
| [2011/12/02 16:56:22] <haus> sort of comforting that it's a known issue | ||
| [2011/12/02 16:58:02] <jorhett1> we found it weeks ago, when the kernel first shipped. Rebooting near a thousand systems to get them off that version is still ongoing :( | ||
| [2011/12/02 16:58:32] <haus> yea, that's a pain | ||
| [2011/12/02 16:58:35] <AK6L> jorhett1: ughh | ||
| [2011/12/02 16:58:40] <AK6L> jorhett1: can i ask where you work? | ||
| [2011/12/02 17:00:51] @ Quit: fronlius: Quit: fronlius | ||
| [2011/12/02 17:02:03] <jorhett1> Tango | ||
| [2011/12/02 17:02:20] <jorhett1> if you do video calling on your phone, you use us. We're the only thing that works on every platform. :) | ||
| [2011/12/02 17:02:28] @ Quit: zwi: Quit: Leaving. | ||
| [2011/12/02 17:02:31] <dingus9> jorhett1: still on rhel5? | ||
| [2011/12/02 17:02:59] @ Quit: newtMcKerr: Quit: Leaving. | ||
| [2011/12/02 17:03:13] @ the_ramink joined channel #puppet | ||
| [2011/12/02 17:04:02] @ Quit: tinhead: Quit: tinhead | ||
| [2011/12/02 17:04:14] @ Quit: dingus9: Quit: This computer has gone to sleep | ||
| [2011/12/02 17:05:14] @ Quit: miche: Disconnected by services | ||
| [2011/12/02 17:05:29] @ nfagerlund joined channel #puppet | ||
| [2011/12/02 17:09:58] @ mgw joined channel #puppet | ||
| [2011/12/02 17:10:58] @ Quit: jongrocho: Remote host closed the connection | ||
| [2011/12/02 17:13:01] @ Quit: tilgovi: Read error: Operation timed out | ||
| [2011/12/02 17:13:37] @ mgw1 joined channel #puppet | ||
| [2011/12/02 17:14:50] @ Quit: mgw: Read error: No route to host | ||
| [2011/12/02 17:16:08] @ mgw joined channel #puppet | ||
| [2011/12/02 17:16:39] @ cwarden left channel #puppet () | ||
| [2011/12/02 17:16:40] @ Quit: mgw: Read error: Connection reset by peer | ||
| [2011/12/02 17:17:02] @ mgw joined channel #puppet | ||
| [2011/12/02 17:17:03] @ Quit: rubic: Quit: Leaving. | ||
| [2011/12/02 17:17:25] @ rubic joined channel #puppet | ||
| [2011/12/02 17:17:33] @ Quit: bmahe: Quit: Leaving | ||
| [2011/12/02 17:17:42] @ Quit: nfagerlund: Remote host closed the connection | ||
| [2011/12/02 17:17:59] @ Quit: mgw1: Ping timeout: 244 seconds | ||
| [2011/12/02 17:19:53] <dwalu> to set node_name value I should just run puppet agent --node_name_value="mynodename" right? | ||
| [2011/12/02 17:20:10] @ Nolimits joined channel #puppet | ||
| [2011/12/02 17:20:14] <dwalu> for some reason this doesn't appear to be the case, and instead I see info: creating a new ssl certificat request for (myfqdn) | ||
| [2011/12/02 17:20:18] @ bmahe joined channel #puppet | ||
| [2011/12/02 17:20:18] @ Quit: bmahe: Changing host | ||
| [2011/12/02 17:20:18] @ bmahe joined channel #puppet | ||
| [2011/12/02 17:20:26] @ Nolimits left channel #puppet () | ||
| [2011/12/02 17:23:08] <the_ramink> dwalu: node_name only takes two values if I understand the docs correctly, facter and cert. If you're trying to change the name the cert uses try, --certname | ||
| [2011/12/02 17:23:25] <dwalu> the_ramink: I'm using node_name_value not node_name | ||
| [2011/12/02 17:23:31] <the_ramink> ah missed that | ||
| [2011/12/02 17:23:36] <dwalu> from the docs, node_name and node_name_value are mutually exclusive, so I'm confused | ||
| [2011/12/02 17:25:19] <the_ramink> you | ||
| [2011/12/02 17:25:36] <the_ramink> you're trying to do that single cert thing for a test env right? | ||
| [2011/12/02 17:26:07] <dwalu> sort of, biut not the single cert thing. | ||
| [2011/12/02 17:26:21] <dwalu> I guess I'm confused as to the difference between a cert name and a node name | ||
| [2011/12/02 17:26:37] <the_ramink> so the problem is node_name_value is strictly the node name, not the name on the cert | ||
| [2011/12/02 17:26:40] @ Quit: wilmoore: Remote host closed the connection | ||
| [2011/12/02 17:26:51] <the_ramink> certname defauls to fqdn. | ||
| [2011/12/02 17:27:10] @ labrown joined channel #puppet | ||
| [2011/12/02 17:27:38] <the_ramink> what's happening is cert generation is taking the same path it always does and your transaction is never getting to the node name value stuff | ||
| [2011/12/02 17:27:47] <the_ramink> make sense? | ||
| [2011/12/02 17:28:34] <dwalu> I think so. it seems that if I set certname and nodename from the commandline during bootstrap then I might be good to go | ||
| [2011/12/02 17:30:54] <the_ramink> Sounds reasonable. I've done some work where I can't control the fqdn, so I set --certname when I start Puppet and then set certname = <%= clientcert %> in puppet.conf to propagate it so I can do regex in my node definitions. | ||
| [2011/12/02 17:32:38] <the_ramink> using node_name_value is probably cleaner because if you hostnames are random enough then you don't have collisions... hmmm not you go me thinking about my own system. | ||
| [2011/12/02 17:36:25] @ Quit: mgw: Quit: Leaving. | ||
| [2011/12/02 17:37:34] @ Quit: dwalu: Quit: dwalu | ||
| [2011/12/02 17:38:56] @ Quit: ianbrandt: Quit: ianbrandt | ||
| [2011/12/02 17:41:41] @ Quit: bosszaru: Quit: leaving | ||
| [2011/12/02 17:46:53] @ Quit: fliplap: Read error: Connection reset by peer | ||
| [2011/12/02 17:47:05] @ fliplap joined channel #puppet | ||
| [2011/12/02 17:47:37] @ pll` joined channel #puppet | ||
| [2011/12/02 17:48:24] @ Quit: trespan: Ping timeout: 245 seconds | ||
| [2011/12/02 17:50:25] @ Quit: BManojlovic: Remote host closed the connection | ||
| [2011/12/02 17:50:32] @ Quit: seanh-ansca: Ping timeout: 244 seconds | ||
| [2011/12/02 17:51:06] @ Quit: alanevans: Ping timeout: 260 seconds | ||
| [2011/12/02 17:51:44] @ Quit: pll: Ping timeout: 245 seconds | ||
| [2011/12/02 18:04:12] @ alanevans joined channel #puppet | ||
| [2011/12/02 18:05:34] @ Quit: maxmartin: Quit: Computer has gone to sleep. | ||
| [2011/12/02 18:12:32] @ Quit: pednyc: Ping timeout: 258 seconds | ||
| [2011/12/02 18:13:05] @ Quit: mao: Ping timeout: 265 seconds | ||
| [2011/12/02 18:16:38] @ pednyc joined channel #puppet | ||
| [2011/12/02 18:19:41] @ Quit: comptona: Ping timeout: 260 seconds | ||
| [2011/12/02 18:21:01] @ Quit: pednyc: Ping timeout: 244 seconds | ||
| [2011/12/02 18:26:43] @ the_ramink_ joined channel #puppet | ||
| [2011/12/02 18:26:43] @ Quit: the_ramink: Read error: Connection reset by peer | ||
| [2011/12/02 18:29:43] @ Quit: labrown: Quit: Leaving | ||
| [2011/12/02 18:34:46] @ Quit: sascha_: Ping timeout: 258 seconds | ||
| [2011/12/02 18:40:01] @ Quit: Djelibeybi: Quit: Textual IRC Client: http://www.textualapp.com/ | ||
| [2011/12/02 18:42:10] @ Quit: skm: Ping timeout: 244 seconds | ||
| [2011/12/02 18:42:52] @ skm joined channel #puppet | ||
| [2011/12/02 18:43:45] @ Quit: mrwacky42: Ping timeout: 244 seconds | ||
| [2011/12/02 18:45:59] @ Quit: Kazuhiro1: Quit: Leaving. | ||
| [2011/12/02 18:47:52] @ wilmoore joined channel #puppet | ||
| [2011/12/02 18:56:25] @ Quit: bmahe: Ping timeout: 244 seconds | ||
| [2011/12/02 18:59:12] <jorhett1> dingus9: yes, and will be on rhel5 for the foreseeable future | ||
| [2011/12/02 19:02:51] @ Quit: langworthy: Quit: langworthy | ||
| [2011/12/02 19:03:40] @ Quit: skm: Quit: Computer has gone to sleep. | ||
| [2011/12/02 19:05:15] @ newtMcKerr joined channel #puppet | ||
| [2011/12/02 19:08:05] @ Quit: wilmoore: Remote host closed the connection | ||
| [2011/12/02 19:09:51] @ wilmoore joined channel #puppet | ||
| [2011/12/02 19:21:16] @ tilgovi joined channel #puppet | ||
| [2011/12/02 19:21:16] @ Quit: tilgovi: Changing host | ||
| [2011/12/02 19:21:16] @ tilgovi joined channel #puppet | ||
| [2011/12/02 19:25:21] @ jongrocho joined channel #puppet | ||
| [2011/12/02 19:25:26] @ Quit: jongrocho: Remote host closed the connection | ||
| [2011/12/02 19:27:26] @ Quit: ryancoleman: Quit: Leaving. | ||
| [2011/12/02 19:34:17] <AK6L> jorhett1: thank you, i owe you the drink of your choice if you are in or come to san francisco, your suggested kernel version fixed the issue. | ||
| [2011/12/02 19:43:54] @ Quit: marczak_: Ping timeout: 252 seconds | ||
| [2011/12/02 19:44:33] @ trespan joined channel #puppet | ||
| [2011/12/02 19:50:12] <jorhett1> I live in SF :) No worries, it's all good. Pay it forward. | ||
| [2011/12/02 19:50:41] <jorhett1> watch that bug, I'll update it when there's a fixed kernel pushed | ||
| [2011/12/02 19:52:30] @ Quit: cliff-hm: Read error: Operation timed out | ||
| [2011/12/02 19:56:25] @ Quit: tilgovi: Ping timeout: 258 seconds | ||
| [2011/12/02 19:59:52] @ Quit: joe-mac: Remote host closed the connection | ||
| [2011/12/02 20:00:48] @ choonming joined channel #puppet | ||
| [2011/12/02 20:01:15] @ Quit: druggo: Quit: 离开 | ||
| [2011/12/02 20:08:16] @ stillbourne joined channel #puppet | ||
| [2011/12/02 20:09:36] <jorhett1> the_ramink: corrupted yaml files. Removed them and all's well. | ||
| [2011/12/02 20:11:22] @ Quit: zahna_: Ping timeout: 258 seconds | ||
| [2011/12/02 20:14:07] @ sascha_ joined channel #puppet | ||
| [2011/12/02 20:17:35] @ Kazuhiro joined channel #puppet | ||
| [2011/12/02 20:31:13] @ Quit: ghanima: Ping timeout: 244 seconds | ||
| [2011/12/02 20:32:56] @ Quit: jorhett1: Quit: Leaving. | ||
| [2011/12/02 20:33:14] @ jorhett joined channel #puppet | ||
| [2011/12/02 20:33:18] <jorhett> yay, me without the 1 | ||
| [2011/12/02 20:35:21] @ Quit: jeremyb: Changing host | ||
| [2011/12/02 20:35:21] @ jeremyb joined channel #puppet | ||
| [2011/12/02 20:36:24] @ skm joined channel #puppet | ||
| [2011/12/02 20:41:09] <matchaw_> How can you ensure and exec(apt-get update) gets executed before any new package install? | ||
| [2011/12/02 20:42:52] @ Kazuhiro left channel #puppet () | ||
| [2011/12/02 20:42:57] <jorhett> require ;-) | ||
| [2011/12/02 20:43:22] <jorhett> require => Exec['apt-get-update'] | ||
| [2011/12/02 20:43:27] <matchaw_> Ah I think this will do it? Exec["apt-update"] -> Package <| |> | ||
| [2011/12/02 20:43:46] @ Quit: nevyn: Ping timeout: 260 seconds | ||
| [2011/12/02 20:43:51] <jorhett> or Package { require => Exec['apt-update'] } | ||
| [2011/12/02 20:44:02] <jorhett> sets a default for all package invocations in that block | ||
| [2011/12/02 20:44:47] @ asteve joined channel #puppet | ||
| [2011/12/02 20:44:48] @ Quit: asteve: Changing host | ||
| [2011/12/02 20:44:48] @ asteve joined channel #puppet | ||
| [2011/12/02 20:45:06] <matchaw_> ok, maybe I just use require on any package that I need to add a repo for first. | ||
| [2011/12/02 20:45:09] @ nevyn joined channel #puppet | ||
| [2011/12/02 20:47:55] <matchaw_> Can I schedule a exec to run immediatly? | ||
| [2011/12/02 20:48:04] <matchaw_> rather then it being "scheduled" | ||
| [2011/12/02 20:48:39] <matchaw_> i.e my exec resource imports a pgp key and then I want it to notify and immediately run the exec(apt-get-update) ? | ||
| [2011/12/02 20:54:58] <Volcane> why does it matter when it runs? if u add requires etc it will run before | ||
| [2011/12/02 20:55:14] <Volcane> there's no reason to say run it immediately as long as it runs before the things that has it as a requirement | ||
| [2011/12/02 20:56:22] <jorhett> puppet is not much with the exact-timing-of-things. You can play games by doing a lot of ordering work, but it's really not helpful. | ||
| [2011/12/02 20:59:37] <fenris02> use cron if you need something run at a specific time. | ||
| [2011/12/02 20:59:53] <gepetto> ::redmine:: Wiki edit: Puppet_Books (#3) @ http://projects.puppetlabs.com/projects/puppet/wiki/Puppet_Books?version=3 (by Grid Engine) | ||
| [2011/12/02 21:00:51] <fenris02> odd. why isnt pro puppet on that wiki page? | ||
| [2011/12/02 21:03:29] @ joe-mac joined channel #puppet | ||
| [2011/12/02 21:08:10] <matchaw_> hi, is anyone familair with https://github.com/deanwilson/puppet-aptsourced, as It seems the provider/type creates the file every time and hence causing reports to always show a change.. | ||
| [2011/12/02 21:14:40] @ mrwacky42 joined channel #puppet | ||
| [2011/12/02 21:15:00] @ Quit: stillbourne: Quit: Leaving | ||
| [2011/12/02 21:15:33] @ Quit: mrwacky42: Client Quit | ||
| [2011/12/02 21:24:00] @ Quit: abetterlie: Read error: Connection reset by peer | ||
| [2011/12/02 21:28:10] <jmslagle> Hrm | ||
| [2011/12/02 21:28:15] <jmslagle> Did AWS just melt? | ||
| [2011/12/02 21:34:22] @ matchaw___ joined channel #puppet | ||
| [2011/12/02 21:35:02] @ Quit: matchaw_: Ping timeout: 244 seconds | ||
| [2011/12/02 21:38:50] @ maxmartin joined channel #puppet | ||
| [2011/12/02 21:40:16] @ nitram_macair joined channel #puppet | ||
| [2011/12/02 21:47:19] @ Quit: nitram_macair: Quit: nitram_macair | ||
| [2011/12/02 21:50:47] @ Quit: maxmartin: Quit: Computer has gone to sleep. | ||
| [2011/12/02 22:02:23] @ Quit: l0n: | ||
| [2011/12/02 22:13:09] @ Quit: mauve_: Ping timeout: 252 seconds | ||
| [2011/12/02 22:28:40] @ Quit: jsurfer: Quit: jsurfer | ||
| [2011/12/02 22:35:12] @ tilgovi joined channel #puppet | ||
| [2011/12/02 22:35:12] @ Quit: tilgovi: Changing host | ||
| [2011/12/02 22:35:12] @ tilgovi joined channel #puppet | ||
| [2011/12/02 22:42:17] @ ocherno joined channel #puppet | ||
| [2011/12/02 22:46:17] @ dissipate_ joined channel #puppet | ||
| [2011/12/02 22:46:19] @ ali3n0 joined channel #puppet | ||
| [2011/12/02 22:47:22] @ Quit: tilgovi: Ping timeout: 244 seconds | ||
| [2011/12/02 22:50:46] @ tilgovi joined channel #puppet | ||
| [2011/12/02 22:50:46] @ Quit: tilgovi: Changing host | ||
| [2011/12/02 22:50:46] @ tilgovi joined channel #puppet | ||
| [2011/12/02 22:57:39] @ bmahe joined channel #puppet | ||
| [2011/12/02 22:57:39] @ Quit: bmahe: Changing host | ||
| [2011/12/02 22:57:39] @ bmahe joined channel #puppet | ||
| [2011/12/02 23:04:57] @ seanh-ansca joined channel #puppet | ||
| [2011/12/02 23:13:56] @ Quit: skm: Ping timeout: 244 seconds | ||
| [2011/12/02 23:14:36] @ Quit: ali3n0: Quit: Leaving. | ||
| [2011/12/02 23:14:38] @ skm joined channel #puppet | ||
| [2011/12/02 23:18:02] @ Quit: seanh-ansca: Quit: Leaving. | ||
| [2011/12/02 23:31:52] @ Quit: sdog: Quit: Leaving. | ||
| [2011/12/02 23:32:05] @ DJClean joined channel #puppet | ||
| [2011/12/02 23:35:57] @ sdog joined channel #puppet | ||
| [2011/12/02 23:37:36] @ Quit: n0ts: Ping timeout: 260 seconds | ||
| [2011/12/02 23:41:36] @ n0ts joined channel #puppet | ||
| [2011/12/02 23:42:29] @ ali3n0 joined channel #puppet | ||
| [2011/12/02 23:43:49] @ Quit: trespan: Ping timeout: 240 seconds | ||
| [2011/12/02 23:43:52] @ seanh-ansca joined channel #puppet | ||
| [2011/12/02 23:45:57] @ Quit: skm: Quit: Computer has gone to sleep. | ||
| [2011/12/02 23:49:47] @ skm joined channel #puppet | ||
| [2011/12/02 23:50:56] @ Quit: seanh-ansca: Quit: Leaving. | ||
| [2011/12/02 23:53:31] @ Quit: shadoi: Quit: Leaving. | ||
| [2011/12/02 23:53:33] @ Quit: sdog: Changing host | ||
| [2011/12/02 23:53:33] @ sdog joined channel #puppet |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!