| [2010/01/13 00:01:15] @ Log started by gepetto | ||
| [2010/01/13 00:01:15] @ nigelk joined channel #puppet | ||
| [2010/01/13 00:08:31] @ M- joined channel #puppet | ||
| [2010/01/13 00:09:41] @ ctso joined channel #puppet | ||
| [2010/01/13 00:11:01] @ Quit: artista: Read error: 110 (Connection timed out) | ||
| [2010/01/13 00:14:49] @ Quit: nigelk: | ||
| [2010/01/13 00:19:34] @ bgupta_ joined channel #puppet | ||
| [2010/01/13 00:33:43] @ yarihm joined channel #puppet | ||
| [2010/01/13 00:34:22] @ bug joined channel #puppet | ||
| [2010/01/13 00:36:36] @ Quit: bgupta: Connection timed out | ||
| [2010/01/13 00:36:37] @ bgupta_ is now known as bgupta | ||
| [2010/01/13 00:47:37] @ nigelk joined channel #puppet | ||
| [2010/01/13 00:48:22] @ nigelk_ joined channel #puppet | ||
| [2010/01/13 01:02:05] @ re__ joined channel #puppet | ||
| [2010/01/13 01:02:06] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 01:02:16] @ re__ is now known as re_ | ||
| [2010/01/13 01:02:53] @ re__ joined channel #puppet | ||
| [2010/01/13 01:02:53] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 01:03:02] @ re__ is now known as re_ | ||
| [2010/01/13 01:04:39] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/13 01:04:59] @ nigelk__ joined channel #puppet | ||
| [2010/01/13 01:05:51] @ Quit: nigelk__: Client Quit | ||
| [2010/01/13 01:05:57] @ Quit: nigelk: Read error: 110 (Connection timed out) | ||
| [2010/01/13 01:17:16] @ jbu joined channel #puppet | ||
| [2010/01/13 01:17:22] @ jbu is now known as tjoe | ||
| [2010/01/13 01:18:12] <tjoe> I have to use some define to clone a git repository, my problem is that this function is called at every puppet run | ||
| [2010/01/13 01:19:30] @ Quit: M-: "Leaving" | ||
| [2010/01/13 01:23:02] @ Quit: nigelk_: Read error: 110 (Connection timed out) | ||
| [2010/01/13 01:24:39] @ gebi joined channel #puppet | ||
| [2010/01/13 01:26:26] @ Quit: jfreeman: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 01:30:17] <jamesturnbull> tjoe: using an exec? | ||
| [2010/01/13 01:39:58] @ Quit: ctso: "Leaving." | ||
| [2010/01/13 01:42:43] @ rasputnik joined channel #puppet | ||
| [2010/01/13 01:47:42] @ Quit: Djelibeybi: "Leaving" | ||
| [2010/01/13 01:48:07] @ Quit: eric0: Remote closed the connection | ||
| [2010/01/13 01:54:08] @ eric0 joined channel #puppet | ||
| [2010/01/13 01:58:30] @ Quit: rasputnik: "This computer has gone to sleep" | ||
| [2010/01/13 01:59:35] @ Quit: kolla: Remote closed the connection | ||
| [2010/01/13 02:01:15] @ PaulWay[w] left channel #puppet ("I must go, there are too many tangrams on this channel.") | ||
| [2010/01/13 02:07:05] @ tufU joined channel #puppet | ||
| [2010/01/13 02:07:08] @ Quit: eight9three: Read error: 60 (Operation timed out) | ||
| [2010/01/13 02:12:37] @ Quit: tufU: "Ex-Chat" | ||
| [2010/01/13 02:15:09] @ hy is now known as _silver | ||
| [2010/01/13 02:18:20] @ rasputnik joined channel #puppet | ||
| [2010/01/13 02:18:38] @ Quit: Bass10: Connection timed out | ||
| [2010/01/13 02:21:34] @ Quit: teyo: | ||
| [2010/01/13 02:26:41] @ jes5 joined channel #puppet | ||
| [2010/01/13 02:36:55] @ Quit: rasputnik: "This computer has gone to sleep" | ||
| [2010/01/13 02:49:17] <tjoe> sorry james, I was hold back, that was with a define but we modified the declaration of the define and now it's working fine | ||
| [2010/01/13 02:49:54] @ mattwynne joined channel #puppet | ||
| [2010/01/13 02:51:43] @ Quit: mattwynne: Client Quit | ||
| [2010/01/13 02:55:46] @ giskard joined channel #puppet | ||
| [2010/01/13 02:59:22] <jamesturnbull> tjoe: okay - nw | ||
| [2010/01/13 03:01:23] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/13 03:01:38] @ giskard joined channel #puppet | ||
| [2010/01/13 03:02:56] @ Quit: jes5: "Leaving." | ||
| [2010/01/13 03:03:11] @ mvn071 joined channel #puppet | ||
| [2010/01/13 03:03:45] @ heydrick2 joined channel #puppet | ||
| [2010/01/13 03:10:40] @ kolla joined channel #puppet | ||
| [2010/01/13 03:15:41] @ Ramonster joined channel #puppet | ||
| [2010/01/13 03:17:18] @ Quit: heydrick1: Read error: 110 (Connection timed out) | ||
| [2010/01/13 03:19:39] @ Quit: giskard: Read error: 113 (No route to host) | ||
| [2010/01/13 03:25:49] @ francois joined channel #puppet | ||
| [2010/01/13 03:27:13] @ yarihm joined channel #puppet | ||
| [2010/01/13 03:32:08] @ giskard joined channel #puppet | ||
| [2010/01/13 03:34:06] @ skiold joined channel #puppet | ||
| [2010/01/13 03:38:31] @ tim|macbook joined channel #puppet | ||
| [2010/01/13 03:38:59] @ Quit: elementai: Read error: 60 (Operation timed out) | ||
| [2010/01/13 03:46:09] @ elementai joined channel #puppet | ||
| [2010/01/13 04:03:14] @ MattyM joined channel #puppet | ||
| [2010/01/13 04:05:36] @ Robbie joined channel #puppet | ||
| [2010/01/13 04:06:55] @ soot joined channel #puppet | ||
| [2010/01/13 04:08:05] @ artista-frustrad joined channel #puppet | ||
| [2010/01/13 04:09:08] @ _silver is now known as hy | ||
| [2010/01/13 04:12:57] @ poison joined channel #puppet | ||
| [2010/01/13 04:13:03] @ Quit: MattyM: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 04:13:43] @ MattyM joined channel #puppet | ||
| [2010/01/13 04:14:41] @ Quit: bug: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 04:14:59] @ bug joined channel #puppet | ||
| [2010/01/13 04:24:25] @ GioGio joined channel #puppet | ||
| [2010/01/13 04:24:33] <GioGio> hi everybody | ||
| [2010/01/13 04:25:25] @ artista joined channel #puppet | ||
| [2010/01/13 04:26:27] <GioGio> i want to know if it's possible to reload a service ? because on the documentation it only talks about start/stop and restart not reload, but for my apache service reload is important :) is it possible to do that please ? | ||
| [2010/01/13 04:26:46] @ Quit: artista-frustrad: Read error: 110 (Connection timed out) | ||
| [2010/01/13 04:26:47] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/13 04:29:01] <kjetilho> GioGio: you can special case the service with restart => "apache2ctl graceful" | ||
| [2010/01/13 04:29:32] <kjetilho> you may need "hasrestart => true" too | ||
| [2010/01/13 04:29:41] <kjetilho> not sure if that is implied when restart is set | ||
| [2010/01/13 04:31:03] @ Welsh_Dwarf joined channel #puppet | ||
| [2010/01/13 04:32:01] <GioGio> kjetilho: thanks :) | ||
| [2010/01/13 04:40:19] @ Quit: gebi: Read error: 60 (Operation timed out) | ||
| [2010/01/13 04:41:06] @ yarihm joined channel #puppet | ||
| [2010/01/13 04:42:47] @ Quit: bsund: Remote closed the connection | ||
| [2010/01/13 04:43:27] @ danielbln joined channel #puppet | ||
| [2010/01/13 04:43:36] @ artista-frustrad joined channel #puppet | ||
| [2010/01/13 04:44:58] @ Quit: artista: Read error: 110 (Connection timed out) | ||
| [2010/01/13 04:46:14] @ pacalm joined channel #puppet | ||
| [2010/01/13 04:54:04] @ verwilst joined channel #puppet | ||
| [2010/01/13 04:56:18] @ nasrat joined channel #puppet | ||
| [2010/01/13 04:58:15] @ Quit: yarihm: Read error: 110 (Connection timed out) | ||
| [2010/01/13 05:05:44] @ jab_doa joined channel #puppet | ||
| [2010/01/13 05:07:27] @ Quit: artista-frustrad: Read error: 110 (Connection timed out) | ||
| [2010/01/13 05:10:29] @ marcoecc joined channel #puppet | ||
| [2010/01/13 05:12:42] @ rasputnik joined channel #puppet | ||
| [2010/01/13 05:17:29] @ gebi joined channel #puppet | ||
| [2010/01/13 05:17:47] @ Quit: MattyM: "ta ta" | ||
| [2010/01/13 05:21:19] @ artista_frustrad joined channel #puppet | ||
| [2010/01/13 05:22:19] @ clusty joined channel #puppet | ||
| [2010/01/13 05:28:38] <clusty> been a while since i set up a puppet client. i just signed the client from the master and restart the client daemon. how can I make it start doing the recipes? | ||
| [2010/01/13 05:34:53] <z00dax> puppetd --test ? | ||
| [2010/01/13 05:40:06] <henk> clusty: it will every 30 minutes by default. and after restarting it. | ||
| [2010/01/13 05:40:08] <clusty> z00dax: though so myself: http://pastebin.org/75399 | ||
| [2010/01/13 05:40:39] <henk> clusty: have you even looked at line 4? | ||
| [2010/01/13 05:40:39] <clusty> henk: does it log in /var/log/puppet/... ? | ||
| [2010/01/13 05:40:49] <henk> clusty: no /var/log/syslog. at least on debian. | ||
| [2010/01/13 05:41:06] <clusty> henk: guess ubuntu is same as debian on this | ||
| [2010/01/13 05:41:14] <henk> probably... | ||
| [2010/01/13 05:41:39] <clusty> henk: i never edited the client stuff by hand. so i thought i do not need to do anything now | ||
| [2010/01/13 05:42:06] <henk> clusty: it's not client stuff... or do you have that file on your client: /etc/puppet/manifests/classes/baseservices.pp? | ||
| [2010/01/13 05:42:11] <clusty> wait | ||
| [2010/01/13 05:42:20] <clusty> i musyt have messed up baseservices rec | ||
| [2010/01/13 05:44:45] <clusty> that was it i think | ||
| [2010/01/13 05:44:51] @ artista-frustrad joined channel #puppet | ||
| [2010/01/13 05:44:58] <clusty> had a bracket missing. this also explains some other weird things :D | ||
| [2010/01/13 05:46:08] <henk> yep... always read error messages please | ||
| [2010/01/13 05:46:14] @ Quit: jab_doa: "Verlassend" | ||
| [2010/01/13 05:47:11] @ Quit: artista_frustrad: Read error: 110 (Connection timed out) | ||
| [2010/01/13 05:51:08] <clusty> henk: i am still finding puppet a bit strange and hard to digest | ||
| [2010/01/13 05:51:09] @ benlovell joined channel #puppet | ||
| [2010/01/13 05:52:17] @ Quit: clemens_: Read error: 110 (Connection timed out) | ||
| [2010/01/13 05:56:57] @ docelic joined channel #puppet | ||
| [2010/01/13 06:02:30] @ Quit: Juzzy: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 06:02:57] @ Juzzy joined channel #puppet | ||
| [2010/01/13 06:03:39] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/13 06:04:36] @ Quit: artista-frustrad: "Leaving" | ||
| [2010/01/13 06:04:50] @ artista_frustrad joined channel #puppet | ||
| [2010/01/13 06:05:04] @ artista-frustrad joined channel #puppet | ||
| [2010/01/13 06:10:16] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:10:29] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:12:45] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:12:51] @ yarihm joined channel #puppet | ||
| [2010/01/13 06:13:02] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:14:41] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:15:14] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:15:14] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:15:30] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:16:01] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:17:41] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:18:22] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 06:20:31] @ bsund joined channel #puppet | ||
| [2010/01/13 06:20:58] @ Quit: bsund: Client Quit | ||
| [2010/01/13 06:24:42] @ n0ts joined channel #puppet | ||
| [2010/01/13 06:29:02] @ Quit: rasputnik: "Leaving" | ||
| [2010/01/13 06:41:54] @ CopeEl joined channel #puppet | ||
| [2010/01/13 06:43:10] <CopeEl> I'm in the uncomfortable situation of having to install jmeter and ant from the latest binary tarballs; what's the best appraoch to this? Puppet is crap at serving big files... should I set up an rsync server or sth and rsync the tarballs and exec an untar? | ||
| [2010/01/13 06:45:02] <z00dax> rpm'ise :) | ||
| [2010/01/13 06:45:35] <z00dax> go the nosrc.rpm route if you need to | ||
| [2010/01/13 06:45:56] <CopeEl> I tried to - got stuck in a twsity maze if java shite | ||
| [2010/01/13 06:46:29] <CopeEl> nosrc is a good idea | ||
| [2010/01/13 06:46:52] @ ninjazjb left channel #puppet ("Leaving.") | ||
| [2010/01/13 06:50:33] <CopeEl> z00dax: | ||
| [2010/01/13 06:50:41] <kjetilho> CopeEl: or upgrade to 0.25.3 | ||
| [2010/01/13 06:50:49] <CopeEl> mistype | ||
| [2010/01/13 06:51:00] <CopeEl> kjetilho: how does that help? | ||
| [2010/01/13 06:51:14] * CopeEl is on 0.25.1 atm | ||
| [2010/01/13 06:51:17] <kjetilho> 0.25.3 doesn't transfer files over XML-RPC | ||
| [2010/01/13 06:51:32] <CopeEl> ah | ||
| [2010/01/13 06:51:33] <kjetilho> hmm. should be pretty efficient then, I thought | ||
| [2010/01/13 06:51:54] <CopeEl> yeah - i've not tested it - just assuming that puppet is shit at serving files, as it always used to be | ||
| [2010/01/13 06:52:27] <kjetilho> yeah, I guess it's the biggest improvement in 0.25.x | ||
| [2010/01/13 07:00:07] @ Quit: crdant: "Leaving." | ||
| [2010/01/13 07:00:13] @ darkweaver87 joined channel #puppet | ||
| [2010/01/13 07:00:14] @ giskard joined channel #puppet | ||
| [2010/01/13 07:00:19] <darkweaver87> hi all | ||
| [2010/01/13 07:02:30] @ Quit: lero: Remote closed the connection | ||
| [2010/01/13 07:02:31] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 07:02:33] @ re_ joined channel #puppet | ||
| [2010/01/13 07:03:11] <darkweaver87> is there anyway to execute a command onlyif a configuration have been modified ? e.g.: file{"/etc/ssh/sshd_config" ....} exec {"/etc/init.d/ssh reload"} | ||
| [2010/01/13 07:03:11] <darkweaver87> maybe a combination of notify and noop ? | ||
| [2010/01/13 07:03:12] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 07:03:15] @ re_ joined channel #puppet | ||
| [2010/01/13 07:04:20] <kjetilho> darkweaver87: exactly | ||
| [2010/01/13 07:04:45] <jamesturnbull> CopeEl: I'd try 0.25.1 with large files - I've seen significant performance improvements | ||
| [2010/01/13 07:04:50] <gepetto> ::trac:: Module Organisation edited @ http://reductivelabs.com/trac/puppet/wiki/ModuleOrganisation?version=51 | ||
| [2010/01/13 07:04:53] <jamesturnbull> although I am biases | ||
| [2010/01/13 07:04:58] <jamesturnbull> biased even | ||
| [2010/01/13 07:05:23] <jamesturnbull> darkweaver87: subscribe | ||
| [2010/01/13 07:08:31] <darkweaver87> jamesturnbull: the problem is I don't know in advance which file to subscribe to. I have defined, for instance, a web service by define web_service($vhost_name) and I use the metaparameter before for each file block in my define. The exec which gracefuls my web server is done externally of this define block. | ||
| [2010/01/13 07:08:39] @ Quit: skiold: | ||
| [2010/01/13 07:09:25] <Volcane> CopeEl: "Finished catalog run in 231.43 seconds" vs "Finished catalog run in 42.74 seconds" with lots of files | ||
| [2010/01/13 07:09:49] <Volcane> darkweaver87: so require / subscribe to the define | ||
| [2010/01/13 07:09:56] <Volcane> darkweaver87: or even the class its in | ||
| [2010/01/13 07:13:34] <kjetilho> why shouldn't he use notify? | ||
| [2010/01/13 07:13:37] * z00dax jealousifies | ||
| [2010/01/13 07:13:55] <z00dax> Finished catalog run in 4231.43 seconds | ||
| [2010/01/13 07:14:01] <Volcane> kjetilho: who knows, we havnt seen his code :P | ||
| [2010/01/13 07:14:09] <Volcane> z00dax: 25? | ||
| [2010/01/13 07:14:13] <z00dax> erm. 4231.34 | ||
| [2010/01/13 07:14:21] <z00dax> 23.something | ||
| [2010/01/13 07:14:23] <Volcane> heh | ||
| [2010/01/13 07:15:16] <darkweaver87> Volcane: same problem -> how can I know it ? | ||
| [2010/01/13 07:15:16] <darkweaver87> e.g: | ||
| [2010/01/13 07:15:16] <darkweaver87> modules/apache2/init.pp: define web_service($v_alias, $v_address, $v_maintenance, $v_enabled){ ...} exec {"apache2ctl graceful": onlyif => "apache2ctl -t"} | ||
| [2010/01/13 07:15:16] <darkweaver87> manifests/nodes.pp:apache2::web_service{"foo": ...} | ||
| [2010/01/13 07:15:16] <darkweaver87> Do you mean I can know all call of web_service in modules/apache2/init.pp ? | ||
| [2010/01/13 07:15:39] <Volcane> pastie your code | ||
| [2010/01/13 07:15:47] <Volcane> pastie.org | ||
| [2010/01/13 07:16:29] <darkweaver87> ok wait few minutes please | ||
| [2010/01/13 07:16:54] <kjetilho> hmm. do onlyif and notifyonly mix? | ||
| [2010/01/13 07:17:33] <kjetilho> I guess if the onlyif is consistent wrt. dependencies it's ok | ||
| [2010/01/13 07:17:57] <kjetilho> notifyonly will fire only once, but that's OK if onlyif depends on the configuration | ||
| [2010/01/13 07:18:15] <kjetilho> so that onlyif will fail until there is a new notification anyway | ||
| [2010/01/13 07:19:13] @ cynicismic joined channel #puppet | ||
| [2010/01/13 07:22:02] <darkweaver87> http://pastie.org/776330: manifests/nodes.pp | ||
| [2010/01/13 07:22:03] <darkweaver87> http://pastie.org/776331: modules/apache2/init.pp | ||
| [2010/01/13 07:22:03] <darkweaver87> http://pastie.org/776333: modules/apache2/front.pp | ||
| [2010/01/13 07:22:03] <darkweaver87> http://pastie.org/776336: manifests/templates.pp | ||
| [2010/01/13 07:25:00] @ PsychoSid joined channel #puppet | ||
| [2010/01/13 07:25:30] @ Quit: choc: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 07:26:46] @ Quit: rmiller4pi8: "Leaving." | ||
| [2010/01/13 07:27:30] <Volcane> so i am not following what the problem is, you're notifying the exec from inside the define right? so the exec will work, why do you need subscribe? | ||
| [2010/01/13 07:27:46] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/13 07:28:07] @ choc joined channel #puppet | ||
| [2010/01/13 07:29:35] @ Quit: choc: SendQ exceeded | ||
| [2010/01/13 07:30:40] <darkweaver87> Volcane: I don't need a subscribe jamesturnbull adviced me to do this ... what I need is my graceful done only if a file is changed. Nowadays, I get the following output: notice: //apache2::front/Exec[apache2ctl graceful]/returns: executed successfully whereas no config file have been touched. Consequently, my apache is gracefuled every 30 minutes. | ||
| [2010/01/13 07:31:13] @ choc joined channel #puppet | ||
| [2010/01/13 07:31:21] <darkweaver87> maybe a mistake from my code | ||
| [2010/01/13 07:32:04] <Volcane> why do you have the exec to do restarts? | ||
| [2010/01/13 07:32:19] <Volcane> you can tell service to use graceful when a reload is needed | ||
| [2010/01/13 07:32:31] <Volcane> but the problemw ith the exec is | ||
| [2010/01/13 07:32:40] <Volcane> without a refreshonly option it will run eery time | ||
| [2010/01/13 07:32:48] <Volcane> so all u probably have to do is add refreshonly => true to the exec | ||
| [2010/01/13 07:33:31] @ Quit: benlovell: | ||
| [2010/01/13 07:36:24] <darkweaver87> ok. I didn't see this option this is what I wanted neither the service restart/start/stop parameters. Thanks for your analyse. | ||
| [2010/01/13 07:37:48] @ darkweaver87 left channel #puppet () | ||
| [2010/01/13 07:47:39] @ whaley joined channel #puppet | ||
| [2010/01/13 07:48:22] @ clemens_ joined channel #puppet | ||
| [2010/01/13 07:51:59] @ Quit: maxagaz: "Ex-Chat" | ||
| [2010/01/13 07:52:33] @ skiold joined channel #puppet | ||
| [2010/01/13 07:58:13] <ohadlevy> telmo:pong | ||
| [2010/01/13 08:03:02] @ zombied_ joined channel #puppet | ||
| [2010/01/13 08:09:00] @ rmiller4pi81 joined channel #puppet | ||
| [2010/01/13 08:09:57] <robinbowes> I'm using exported resources to collect/distribute ssh keys | ||
| [2010/01/13 08:10:24] <robinbowes> I'm using pretty much the exact same code as the example here: | ||
| [2010/01/13 08:10:34] @ bodepd joined channel #puppet | ||
| [2010/01/13 08:10:40] <robinbowes> http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial#exported-resources | ||
| [2010/01/13 08:10:50] <robinbowes> except I'm using rsa keys | ||
| [2010/01/13 08:12:29] <robinbowes> What I find is that when I first deploy a node I can end up with an empty key in /etc/ssh/ssh_known_hosts | ||
| [2010/01/13 08:13:00] <robinbowes> (on other nodes) | ||
| [2010/01/13 08:13:08] <robinbowes> This breaks puppet and means I have to manually remove the line to get it to work | ||
| [2010/01/13 08:14:11] @ Quit: bug: | ||
| [2010/01/13 08:14:14] @ yure joined channel #puppet | ||
| [2010/01/13 08:15:16] <robinbowes> Hmm, I suspect this happens if the @@sshkey is evaluated before the $sshrsakey has been deployed to the node | ||
| [2010/01/13 08:16:29] @ bug joined channel #puppet | ||
| [2010/01/13 08:16:36] <robinbowes> Anyone seen anything like this? | ||
| [2010/01/13 08:16:48] <robinbowes> Got any idea how to fix it? | ||
| [2010/01/13 08:16:53] @ Quit: _elmata_: "up up and away" | ||
| [2010/01/13 08:17:17] @ Bass10 joined channel #puppet | ||
| [2010/01/13 08:17:48] @ brownj75 joined channel #puppet | ||
| [2010/01/13 08:23:28] @ Quit: bodepd: | ||
| [2010/01/13 08:27:13] @ Quit: rmiller4pi8: Read error: 110 (Connection timed out) | ||
| [2010/01/13 08:32:14] @ Quit: lilmatt: Client Quit | ||
| [2010/01/13 08:32:14] @ Quit: QMan: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 08:32:37] @ QMan joined channel #puppet | ||
| [2010/01/13 08:33:14] @ Quit: Bass10: "Leaving" | ||
| [2010/01/13 08:33:20] <zombied_> I got a handful of puppet clients that either aren't checking in with the puppet master or arent applying changes they see until puppet is restarted | ||
| [2010/01/13 08:33:25] <zombied_> Any idea why that may be the case? | ||
| [2010/01/13 08:33:50] @ Quit: brownj75: "Lost terminal" | ||
| [2010/01/13 08:36:14] @ Bass10 joined channel #puppet | ||
| [2010/01/13 08:37:18] @ Quit: Bass10: SendQ exceeded | ||
| [2010/01/13 08:37:47] @ Bass10 joined channel #puppet | ||
| [2010/01/13 08:37:54] @ Quit: zombied_: "Page closed" | ||
| [2010/01/13 08:39:49] @ nakano_ is now known as nakano | ||
| [2010/01/13 08:43:04] @ Quit: GioGio: Ping timeout: 180 seconds | ||
| [2010/01/13 08:46:44] @ Quit: docelic: "http://www.spinlocksolutions.com/" | ||
| [2010/01/13 08:47:45] @ tyll joined channel #puppet | ||
| [2010/01/13 08:47:57] @ benlovell joined channel #puppet | ||
| [2010/01/13 08:49:21] <ohadlevy> should facter on solaris report the dns domain name or the nis/ldap domain name? it seems a bit problematical to me | ||
| [2010/01/13 08:50:28] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 08:50:32] @ jab_doa joined channel #puppet | ||
| [2010/01/13 08:50:39] @ re_ joined channel #puppet | ||
| [2010/01/13 08:50:50] <ohadlevy> any Solaris gurus around? | ||
| [2010/01/13 08:51:37] <tyll> Hiyas, I noticed that the links on http://logcheck.org/download.html are either broken or contain only old content. Are there any tarballs somewhere else or is there only the git repository? | ||
| [2010/01/13 08:52:03] <henk> tyll: puppet != logcheck? | ||
| [2010/01/13 08:53:04] <tyll> henk: uh, sorry | ||
| [2010/01/13 08:55:39] @ nakano is now known as nakano_ | ||
| [2010/01/13 08:57:32] @ joe-mac joined channel #puppet | ||
| [2010/01/13 09:00:08] @ euclid joined channel #puppet | ||
| [2010/01/13 09:01:45] <CopeEl> I appear to have xinetd defined in a class somewhere; i've also included it in new_class; I guess I could abstract an xinetd module, or is this a case for virtual resources? | ||
| [2010/01/13 09:02:20] <fsweetser> for something like xinetd, I'd lean towards a separate module | ||
| [2010/01/13 09:05:30] <ohadlevy> I would make a define inside a xinetd module | ||
| [2010/01/13 09:06:11] @ Quit: gebi: Remote closed the connection | ||
| [2010/01/13 09:07:35] <kjetilho> do you guys remember if Puppet false maps to Ruby Boolean false or to the Ruby string "false"? | ||
| [2010/01/13 09:07:39] <kjetilho> in a template, that is | ||
| [2010/01/13 09:08:19] <kjetilho> ohadlevy: which fact do you mean? | ||
| [2010/01/13 09:08:44] <ohadlevy> kjetilho:domain | ||
| [2010/01/13 09:09:04] @ Newsome joined channel #puppet | ||
| [2010/01/13 09:09:18] <kjetilho> dnsdomain. it's not different from Linux, really | ||
| [2010/01/13 09:09:25] <kjetilho> getdomainname is NIS name there, too | ||
| [2010/01/13 09:09:54] <ohadlevy> kjetilho: the problem is that it will use the domainname command, which will return the nis domain | ||
| [2010/01/13 09:10:07] <kjetilho> oh | ||
| [2010/01/13 09:10:21] <ohadlevy> which has an impact on fqdn fact | ||
| [2010/01/13 09:10:35] <CopeEl> ok.... just to clarify my understanding; I have an xinetd module; node foo includes xinted::server; cool; I have a cobbler module, cobbler needs to ensure that xinetd is running; I have an rsync::server module, this also needs xinetd to be running; Do I include a service ensure => running in both modles, or do i trust puppet to get it right, because xinetd::server has an ensure => running? | ||
| [2010/01/13 09:10:42] <CopeEl> fuck that's a long sentence | ||
| [2010/01/13 09:10:58] <ohadlevy> the right way to solve it is by parsing /etc/resolv.conf and look for a domain statement | ||
| [2010/01/13 09:11:03] <ohadlevy> (i think at least) | ||
| [2010/01/13 09:12:03] <kjetilho> ohadlevy: domainname returns NIS domainname on Linux, too | ||
| [2010/01/13 09:12:13] <kjetilho> domainname(8) that is | ||
| [2010/01/13 09:12:28] <ohadlevy> kjetilho: yeah, but facter tries dnsdomainname first | ||
| [2010/01/13 09:12:40] <ohadlevy> on old solaris (e.g. 8) it doesnt exists | ||
| [2010/01/13 09:12:46] @ MattyM joined channel #puppet | ||
| [2010/01/13 09:12:51] <kjetilho> ok, but using domainname(8) is almost always broken, IMHO | ||
| [2010/01/13 09:12:54] <ohadlevy> lutter:ping | ||
| [2010/01/13 09:13:00] <ohadlevy> git blames lutter ;) | ||
| [2010/01/13 09:13:50] <ohadlevy> !time lutter | ||
| [2010/01/13 09:13:54] <gepetto> ohadlevy: lutter is an unknown time. | ||
| [2010/01/13 09:13:54] <kjetilho> I think the right way is to look up FQDN, then compare to nodename. if nodename is FQDN, use the part after first period | ||
| [2010/01/13 09:14:05] <kjetilho> if nodename is short, remove nodename from FQDN to make domainname | ||
| [2010/01/13 09:14:26] <ohadlevy> how do you look for fqdn? | ||
| [2010/01/13 09:14:27] <kjetilho> so that "$nodename.$domainname" == $fqdn | ||
| [2010/01/13 09:14:44] <kjetilho> ohadlevy: getent `uname -n` hosts | awk '{print $2}' :-) | ||
| [2010/01/13 09:15:21] <kjetilho> getent hosts `uname -n` | awk '{print $2}' | ||
| [2010/01/13 09:15:56] <ohadlevy> kjetilho: I would say the following order is the right imho: | ||
| [2010/01/13 09:16:06] <ohadlevy> 1. fqdn name (extract the domain from it) | ||
| [2010/01/13 09:16:19] <ohadlevy> 2. look in resolv.conf for domain statement | ||
| [2010/01/13 09:16:22] <ohadlevy> 3. dnsdomain name | ||
| [2010/01/13 09:16:34] <ohadlevy> 4. search statement in resolv.conf | ||
| [2010/01/13 09:16:46] <ohadlevy> and leave domainname out of it all together | ||
| [2010/01/13 09:17:02] <ohadlevy> its not so much of a problem with the domain fact by it self, rather that fqdn is composed of it | ||
| [2010/01/13 09:17:30] <ohadlevy> and doing something like host{$fqdn: ensure => $ipaddress} is really dangerous | ||
| [2010/01/13 09:17:59] <kjetilho> what is "extract domain" in step 1? | ||
| [2010/01/13 09:20:29] <ohadlevy> getent hosts `uname -n`|awk '{print $2}'|cut -d "." -f 2- | ||
| [2010/01/13 09:20:36] <ohadlevy> or ruby equiv | ||
| [2010/01/13 09:20:55] <kjetilho> ok. but if that's step 1, you don't really need the others | ||
| [2010/01/13 09:21:17] <kjetilho> I thought you wanted to cater for people who do domain = bigbank.com and nodename = www01.nyc | ||
| [2010/01/13 09:21:18] <ohadlevy> kjetilho: hostname can be in short format | ||
| [2010/01/13 09:21:48] <ohadlevy> but that domain is a nis/ldap domain isnt it? | ||
| [2010/01/13 09:21:58] <kjetilho> no | ||
| [2010/01/13 09:22:08] <kjetilho> you just need to set ndots in resolv.conf | ||
| [2010/01/13 09:22:16] @ mattwynne joined channel #puppet | ||
| [2010/01/13 09:22:17] <ohadlevy> ah right | ||
| [2010/01/13 09:22:30] <kjetilho> actually you don't need that either, it will try search path afterwards anyway | ||
| [2010/01/13 09:22:45] @ themurph joined channel #puppet | ||
| [2010/01/13 09:23:04] <ohadlevy> you can also use hostname -d (but I'm not sure if it exists on old systems) | ||
| [2010/01/13 09:23:41] <kjetilho> on Solaris, that will set the hostname to ... "-d" | ||
| [2010/01/13 09:23:45] <ohadlevy> yeah | ||
| [2010/01/13 09:23:53] <kjetilho> not uncommon to see Solaris hosts named "--fqdn" ;-) | ||
| [2010/01/13 09:24:09] <ohadlevy> neverthless, my point is that domain should not be set to domainname, thats purely nis/ldap domainname | ||
| [2010/01/13 09:24:15] <kjetilho> agreed | ||
| [2010/01/13 09:24:40] <ohadlevy> or atleast, change the order to make sure you cant find a dns domain name, and then add query the domainname | ||
| [2010/01/13 09:24:53] <ohadlevy> because now i need to change /etc/domainname just for facter | ||
| [2010/01/13 09:25:08] <ohadlevy> which I'm not sure what will it break for ldap | ||
| [2010/01/13 09:25:14] <kjetilho> that's bad, that will trigger ypbind to start | ||
| [2010/01/13 09:25:15] @ bjaspan joined channel #puppet | ||
| [2010/01/13 09:25:24] <ohadlevy> I'm guessing it uses for some maps lookup | ||
| [2010/01/13 09:25:28] <kjetilho> hmm. it needs /var/yp to exist, too | ||
| [2010/01/13 09:25:42] <kjetilho> of course it's just an unused process, but it's a bit ugly | ||
| [2010/01/13 09:26:28] <ohadlevy> i think that it would actully break some ldap lookups... the whole way that solaris store the ldap setting as a profile on the ldapserver itself | ||
| [2010/01/13 09:27:07] @ Quit: bjaspan: Client Quit | ||
| [2010/01/13 09:29:53] <kjetilho> I checked, Puppet false maps to Ruby false | ||
| [2010/01/13 09:36:19] <henk> what setup is recommended regarding the http stuff? apache as frontend? mongrel? nginx? | ||
| [2010/01/13 09:37:24] @ alfism joined channel #puppet | ||
| [2010/01/13 09:38:05] @ Quit: jason^: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 09:40:35] @ ctso joined channel #puppet | ||
| [2010/01/13 09:43:32] <ohadlevy> henk: anything goes, but most people use today apache with passenger | ||
| [2010/01/13 09:44:18] <henk> ohadlevy: ok, any idea which method is the best documented? | ||
| [2010/01/13 09:44:36] <ohadlevy> henk: probably apache + passenger / mongrel | ||
| [2010/01/13 09:44:42] <ohadlevy> but the docs are quite good also for nginx | ||
| [2010/01/13 09:45:57] <henk> ok, passenger is the "default" used by puppetmaster anyway, right? | ||
| [2010/01/13 09:46:17] @ ffrank joined channel #puppet | ||
| [2010/01/13 09:46:33] @ Quit: PsychoSid: Read error: 110 (Connection timed out) | ||
| [2010/01/13 09:48:42] @ cwebber joined channel #puppet | ||
| [2010/01/13 09:48:56] <ffrank> hi. i'd like a define to handle both arrays and scalar strings as values for one parameter. is there an obvious way i'm unaware of? because that seems rather difficult... | ||
| [2010/01/13 09:49:11] <henk> ohadlevy: thanks for the info! | ||
| [2010/01/13 09:50:05] @ pheezy joined channel #puppet | ||
| [2010/01/13 09:50:47] @ Quit: jab_doa: "Verlassend" | ||
| [2010/01/13 09:51:07] <jb> henk: no, webrick is. | ||
| [2010/01/13 09:51:16] <ffrank> i *could* let an inline_template look at "is_array?" and generate the appropriate string for further processing. that seems a little gratuitous though | ||
| [2010/01/13 09:51:21] <jb> (and the least scalable) | ||
| [2010/01/13 09:51:26] <henk> jb: ah, right, thanks! | ||
| [2010/01/13 09:55:26] @ Quit: rmiller4pi81: "Leaving." | ||
| [2010/01/13 10:01:48] @ re___ joined channel #puppet | ||
| [2010/01/13 10:01:49] @ Quit: re_: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 10:01:58] @ re___ is now known as re_ | ||
| [2010/01/13 10:05:24] @ bjaspan joined channel #puppet | ||
| [2010/01/13 10:05:27] @ vzctl__ joined channel #puppet | ||
| [2010/01/13 10:05:31] @ crdant joined channel #puppet | ||
| [2010/01/13 10:06:40] @ Quit: lex: Read error: 110 (Connection timed out) | ||
| [2010/01/13 10:14:54] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/13 10:20:25] @ Quit: elementai: Read error: 60 (Operation timed out) | ||
| [2010/01/13 10:23:19] @ Quit: lpmusic: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 10:28:11] @ elementai joined channel #puppet | ||
| [2010/01/13 10:29:36] @ lpmusic joined channel #puppet | ||
| [2010/01/13 10:31:06] @ kaptk2 joined channel #puppet | ||
| [2010/01/13 10:31:53] @ labrown joined channel #puppet | ||
| [2010/01/13 10:32:07] @ Quit: crdant: Read error: 110 (Connection timed out) | ||
| [2010/01/13 10:35:37] @ Quit: rmiller4pi8: "Leaving." | ||
| [2010/01/13 10:36:37] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/13 10:39:55] @ Quit: yarihm: Read error: 60 (Operation timed out) | ||
| [2010/01/13 10:41:29] <KarlHungus> i need to create a fact called snmpd that is true if snmpd is running and installed or false. how do i use puppets built in service status detection to do this? | ||
| [2010/01/13 10:42:32] <kjetilho> enable => $snmpd | ||
| [2010/01/13 10:46:18] @ Quit: giskard: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 10:46:24] @ giskard joined channel #puppet | ||
| [2010/01/13 10:47:13] @ bsund joined channel #puppet | ||
| [2010/01/13 10:48:07] @ Quit: tim|macbook: | ||
| [2010/01/13 10:54:47] <KarlHungus> kjetilho: i'm not sure i understand. i have modules (varnish is currently what i'm thinking of) that need to deploy configuration lines into snmpd.conf. but i don't want to deploy them if snmpd is not installed and running on that host | ||
| [2010/01/13 10:55:02] @ flakrat joined channel #puppet | ||
| [2010/01/13 10:55:03] @ Quit: ctso: "Leaving." | ||
| [2010/01/13 10:56:12] <kjetilho> well then, use if $snmpd { } | ||
| [2010/01/13 10:56:46] <kjetilho> or test for snmpd inside the template | ||
| [2010/01/13 10:57:20] <ffrank> kjetilho: is snmpd a fact? | ||
| [2010/01/13 10:57:36] <KarlHungus> i don't believe so... i think that is where i am becoming confused | ||
| [2010/01/13 10:58:10] <ffrank> KarlHungus: have you grepped through the facts on a snmpd-enabled machine? | ||
| [2010/01/13 10:58:16] <KarlHungus> ffrank: yes | ||
| [2010/01/13 10:58:48] <kjetilho> oh I misread "I need to create a fact" as "I've created a fact" | ||
| [2010/01/13 10:59:17] <ffrank> anyway, i don't think puppet's notion of services can help you here | ||
| [2010/01/13 10:59:45] <KarlHungus> i was going to write my own fact, but would like to use puppets built in service status detection instead of hardcoding 'foo=system 'service foo status' and locking the fact to an operating system | ||
| [2010/01/13 10:59:48] <ffrank> except you migrate all snmp nodes to having snmpd puppet managed (which sounds like the right thing to do here) | ||
| [2010/01/13 11:00:08] <KarlHungus> yes. snmpd is puppet managed | ||
| [2010/01/13 11:00:52] <ffrank> hmm. you could query that using defined(Service["snmpd"]), but that is order dependent | ||
| [2010/01/13 11:01:05] <ffrank> may still work in your case | ||
| [2010/01/13 11:01:16] <kjetilho> KarlHungus: if the snmpd class/module is only included on nodes which enable it, you can set a variable inside it | ||
| [2010/01/13 11:01:36] <kjetilho> say, $enabled = true. then you test for $snmpd::enabled | ||
| [2010/01/13 11:01:49] <KarlHungus> kjetilho: ahh. i like the sound of that let me test | ||
| [2010/01/13 11:02:32] <ffrank> kjetilho: won't that make puppet fail where class snmpd isn't included? | ||
| [2010/01/13 11:02:41] <kjetilho> hmm. perhaps :/ | ||
| [2010/01/13 11:03:01] <ffrank> but then, when checking only for existence not value...hmm... | ||
| [2010/01/13 11:03:07] <ffrank> he'll see :) | ||
| [2010/01/13 11:03:51] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 11:04:12] <kjetilho> "Class foo has not been evaluated so its variables cannot be referenced" | ||
| [2010/01/13 11:04:14] <kjetilho> oh well | ||
| [2010/01/13 11:04:40] <kjetilho> sorry about the bad suggestion | ||
| [2010/01/13 11:05:35] <KarlHungus> kjetilho: it seems to work fine for me | ||
| [2010/01/13 11:05:37] @ webminster joined channel #puppet | ||
| [2010/01/13 11:05:50] <ffrank> KarlHungus: have you tried on a non-snmp node? | ||
| [2010/01/13 11:06:07] <KarlHungus> ahh. but the snmpd class is available to puppet still | ||
| [2010/01/13 11:06:17] <KarlHungus> even though it isn't included on the test node | ||
| [2010/01/13 11:06:26] <ffrank> well, have you? :P | ||
| [2010/01/13 11:06:33] <KarlHungus> ffrank: yes | ||
| [2010/01/13 11:06:40] <ffrank> and it does work? which version? | ||
| [2010/01/13 11:07:15] <kjetilho> the error I pasted is from 0.24.8 | ||
| [2010/01/13 11:07:15] <KarlHungus> i did if $snmp::enabled { file { "/tmp/foo": ensure => "present", } } else { file { "/tmp/foo": ensure => "absent", } } | ||
| [2010/01/13 11:07:35] <KarlHungus> in 0.25.1 | ||
| [2010/01/13 11:07:40] <ffrank> interesting | ||
| [2010/01/13 11:07:59] @ billy_ran_away joined channel #puppet | ||
| [2010/01/13 11:08:00] <kjetilho> and the class snmp wasn't included or implicitly included by it? | ||
| [2010/01/13 11:08:04] @ lilmatt joined channel #puppet | ||
| [2010/01/13 11:08:16] <ffrank> i seem to recall 0.25.0 having problems there... | ||
| [2010/01/13 11:08:56] @ fluxdude joined channel #puppet | ||
| [2010/01/13 11:09:05] <KarlHungus> kjetilho: the test works whether or not the snmp class is available to puppet, and whether or not it is inclluded on the node | ||
| [2010/01/13 11:09:31] <ffrank> the former is indeed surprising | ||
| [2010/01/13 11:09:35] <robinbowes> KarlHungus: the "correct" way to do this is to use something like concatfile to build snmpd.conf from various file fragments | ||
| [2010/01/13 11:09:36] <kjetilho> very interesting :) | ||
| [2010/01/13 11:10:08] @ Quit: erm__: "leaving" | ||
| [2010/01/13 11:10:10] <KarlHungus> robinbowes: i know. but i don't want to write file fragments to /etc/snmp/conf.d/ or elsewhere if snmp isn't even installed | ||
| [2010/01/13 11:10:17] @ erm_ joined channel #puppet | ||
| [2010/01/13 11:10:19] <ffrank> obviously, puppet is going to be a lot less harsh. i'm not sure i'm pleased | ||
| [2010/01/13 11:11:54] <robinbowes> So add a test to the foo::snmpd class that only drops files in /etc/snmp/conf.d if /etc/snmp exists? | ||
| [2010/01/13 11:12:11] <robinbowes> Or some other such indicator of whether or not snmp is installed | ||
| [2010/01/13 11:12:21] <KarlHungus> robinbowes: that assumes /etc/snmp is the config dir for snmp | ||
| [2010/01/13 11:12:26] <ohadlevy> fsweetser: ping | ||
| [2010/01/13 11:12:37] <robinbowes> <robinbowes> Or some other such indicator of whether or not snmp is installed | ||
| [2010/01/13 11:12:50] <ffrank> robinbowes: that's not gaining anything. it still boils down to some sort of custom fact | ||
| [2010/01/13 11:13:04] <KarlHungus> robinbowes: my initial question was how to detect a services status inside of a fact using the built in puppet service status handler | ||
| [2010/01/13 11:13:32] <ohadlevy> KarlHungus: why do you want to do that? | ||
| [2010/01/13 11:14:01] <robinbowes> Do you control whether or not snmp is installed on machines? | ||
| [2010/01/13 11:14:29] <KarlHungus> ohadlevy: i need to export a file fragment for the snmp config directives needed to alloow net-snmp to view varnish internals. | ||
| [2010/01/13 11:14:40] <KarlHungus> robinbowes: yes. i include the snmp module i wrote | ||
| [2010/01/13 11:14:56] <ohadlevy> KarlHungus: so you need to query for a service status in your fact? | ||
| [2010/01/13 11:15:05] <KarlHungus> ohadlevy: that would be the cleanest way to do this, imho | ||
| [2010/01/13 11:15:26] <robinbowes> So, you know if snmp is installed on the machine already? | ||
| [2010/01/13 11:15:55] <KarlHungus> robinbowes: that is the crux of the problem... | ||
| [2010/01/13 11:16:03] <ffrank> robinbowes: the question revolves around how to pass that info to an otherwise unrelated part of the manifest | ||
| [2010/01/13 11:16:05] <robinbowes> ie. you question "is snmp installed" should be equivalent to "did I install snmp" ? | ||
| [2010/01/13 11:16:14] <ohadlevy> hKarlHungus: maybe this would be useful ttp://www.devco.net/archives/2009/10/19/reusing_puppet_providers.php | ||
| [2010/01/13 11:16:34] <KarlHungus> robinbowes: no. it isn't. the question is "does puppet think snmp has been installed." | ||
| [2010/01/13 11:16:37] <robinbowes> And this would be even more useful: http://www.devco.net/archives/2009/10/19/reusing_puppet_providers.php | ||
| [2010/01/13 11:16:39] <robinbowes> ;) | ||
| [2010/01/13 11:17:13] <KarlHungus> ahh. perfect | ||
| [2010/01/13 11:17:13] <robinbowes> So how are you deciding whether or not to install snmp on the machines? | ||
| [2010/01/13 11:17:27] <KarlHungus> robinbowes: in my nodes.pp file | ||
| [2010/01/13 11:18:03] <robinbowes> arbitrarily? | ||
| [2010/01/13 11:18:21] <KarlHungus> i could do "include snmp\n$snmp_enable = true" but that feels uneccesarily redundant | ||
| [2010/01/13 11:18:26] <KarlHungus> robinbowes: yes. arbitrarily | ||
| [2010/01/13 11:18:39] <webminster> hi, has anyone tried the mysql custom type like http://git.black.co.at/?p=module-mysql;a=summary and gotten it to work? I've spent a couple days on it, and while it seems the resources are recgnized, it doesn't apply any actions doign something like 'mysql_database{ "test": ensure => absent }'. have logging on the mysql database, and it doesn't show any connection attempts either | ||
| [2010/01/13 11:18:42] <KarlHungus> at least arbitrarily for the intents of this discussion | ||
| [2010/01/13 11:19:49] <Volcane> if defined(Class["snmp"]) { # do snmp stuff } | ||
| [2010/01/13 11:19:58] <robinbowes> ... that's where I was going | ||
| [2010/01/13 11:20:09] <KarlHungus> Volcane: that would have ordering problems, afaik | ||
| [2010/01/13 11:20:12] <Volcane> order dependant (I'd say its a puppet bug) | ||
| [2010/01/13 11:20:48] @ yarihm joined channel #puppet | ||
| [2010/01/13 11:20:55] <Volcane> KarlHungus: sure, but thats your only sane option really other than a simple fact that checks if the snmp config file exist - no need to bother with the service running or not | ||
| [2010/01/13 11:21:21] @ Quit: ffrank: "leaving" | ||
| [2010/01/13 11:21:26] <Volcane> an anway, you're installing some big daemon that supports snmp | ||
| [2010/01/13 11:21:34] <Volcane> just do 'include snmp' in the daemons class | ||
| [2010/01/13 11:21:37] <Volcane> cos bonus, why wouldnt you? | ||
| [2010/01/13 11:21:56] <Volcane> and then as others have said use snippets to build up the config for awesomeness | ||
| [2010/01/13 11:22:11] <KarlHungus> Volcane: placing $enabled = true in the snmp class and then checking $snmp::enabled in other classes seems to work well in 0.25.1 | ||
| [2010/01/13 11:22:23] <Volcane> why do u need to know its enabled? | ||
| [2010/01/13 11:22:25] <Volcane> i dont get it | ||
| [2010/01/13 11:22:36] <robinbowes> s/enabled/installed | ||
| [2010/01/13 11:22:53] <robinbowes> s/enabled/wibble/ | ||
| [2010/01/13 11:22:57] <robinbowes> It's just a var | ||
| [2010/01/13 11:22:58] <KarlHungus> i don't want to enforce the requirement for snmp on varnish | ||
| [2010/01/13 11:23:06] <KarlHungus> s/enabled/available/ | ||
| [2010/01/13 11:23:39] <Volcane> would you loose something if you did? cos really madly convoluted config management code vs varnish+snmp bonus | ||
| [2010/01/13 11:24:08] <KarlHungus> Volcane: snmp can be a security risk. i only want it installed on machines that include nagios::target | ||
| [2010/01/13 11:24:29] <KarlHungus> nagios::target will include snmp and ensure that it is secured | ||
| [2010/01/13 11:24:30] <Volcane> ah | ||
| [2010/01/13 11:24:51] <Volcane> so by default your snmp firewall ports are open | ||
| [2010/01/13 11:24:51] @ gebi joined channel #puppet | ||
| [2010/01/13 11:24:59] <Volcane> and so it would be insecure | ||
| [2010/01/13 11:25:01] <Volcane> shot | ||
| [2010/01/13 11:25:02] <robinbowes> So why not ensure snmp is installed *and* secured in the snmp module? | ||
| [2010/01/13 11:25:10] <KarlHungus> and checking for nagios::target in varnish is too heavily coupled for my tastes | ||
| [2010/01/13 11:25:37] <KarlHungus> robinbowes: i do. but whether or not snmp is installed doesnt depend on varnish. it depends on whether that node is configured as a nagios target | ||
| [2010/01/13 11:25:49] <KarlHungus> varnish has no dependency on snmp, nor should it | ||
| [2010/01/13 11:26:03] @ brownj75 joined channel #puppet | ||
| [2010/01/13 11:26:46] * robinbowes still not getting why snmp not on *all* nodes | ||
| [2010/01/13 11:27:03] <kjetilho> snmpd can go wild and use CPU and RAM | ||
| [2010/01/13 11:27:16] @ Quit: mvn071: "Leaving" | ||
| [2010/01/13 11:27:26] <KarlHungus> robinbowes: in my case, it currently is on all nodes. but that isn't always going to be true | ||
| [2010/01/13 11:27:35] @ Quit: erm_: "leaving" | ||
| [2010/01/13 11:27:48] <KarlHungus> so for the sake of this discussion, lets assume the decision to include snmp is arbitrary | ||
| [2010/01/13 11:28:08] <KarlHungus> for me, that decision is made when i include nagios::target | ||
| [2010/01/13 11:28:12] <Volcane> the only option you have is to set flag variables somewhere | ||
| [2010/01/13 11:28:20] <robinbowes> kjetilho: My snmpd run command: exec chpst -o ${SVOPENFILES} -d ${SVDATALIMIT} ${DAEMON} -f -Lo ${OPTIONS} | ||
| [2010/01/13 11:28:20] <Volcane> or to interrogate the node and set facts | ||
| [2010/01/13 11:28:45] <Volcane> if ordering issues is a problem - which really shoud just be filed as a puppet bug | ||
| [2010/01/13 11:28:52] <KarlHungus> Volcane: yeah. i think flag variables are the best way to go | ||
| [2010/01/13 11:29:21] @ Quit: lilmatt: Client Quit | ||
| [2010/01/13 11:29:29] <KarlHungus> thanks for the insight, everyone. i think i have a good solution =) | ||
| [2010/01/13 11:30:37] @ lilmatt joined channel #puppet | ||
| [2010/01/13 11:35:54] @ RageLink joined channel #puppet | ||
| [2010/01/13 11:39:56] <webminster> has anyone tried the mysql custom type like http://git.black.co.at/?p=module-mysql;a=summary and gotten it to work? I've spent a couple days on it, and while it seems the resources are recgnized, it doesn't apply any actions doign something like 'mysql_database{ "test": ensure => absent }'. have logging on the mysql database, and it doesn't show any connection attempts either | ||
| [2010/01/13 11:40:12] <webminster> is there any guidance on debugging execution of custom types? | ||
| [2010/01/13 11:42:30] @ Quit: Ramonster: Client Quit | ||
| [2010/01/13 11:43:54] @ Quit: clusty: "This computer has gone to sleep" | ||
| [2010/01/13 11:44:57] @ mattwynne joined channel #puppet | ||
| [2010/01/13 11:45:15] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 11:46:21] @ Quit: skiold: | ||
| [2010/01/13 11:46:52] @ mattwynne joined channel #puppet | ||
| [2010/01/13 11:47:36] @ Quit: bug: | ||
| [2010/01/13 11:48:22] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/13 11:50:19] @ herlo left channel #puppet () | ||
| [2010/01/13 11:51:41] @ Quit: pacalm: | ||
| [2010/01/13 12:10:27] @ Quit: kolla: Remote closed the connection | ||
| [2010/01/13 12:10:41] @ teyo joined channel #puppet | ||
| [2010/01/13 12:13:43] @ Quit: fsweetser: Remote closed the connection | ||
| [2010/01/13 12:14:48] @ Quit: Robbie: Remote closed the connection | ||
| [2010/01/13 12:19:22] @ Quit: benlovell: | ||
| [2010/01/13 12:22:21] @ Quit: Welsh_Dwarf: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 12:22:54] <lutter> ohadlevy: pong | ||
| [2010/01/13 12:23:18] @ ShiNboi joined channel #puppet | ||
| [2010/01/13 12:24:14] @ toi joined channel #puppet | ||
| [2010/01/13 12:25:46] @ Quit: gebi: Read error: 110 (Connection timed out) | ||
| [2010/01/13 12:26:55] @ Pirate_Hunter joined channel #puppet | ||
| [2010/01/13 12:26:56] @ Quit: malraid: Remote closed the connection | ||
| [2010/01/13 12:27:11] @ Pirate_Hunter left channel #puppet ("I'll be back Bigger, Badder n in HD View") | ||
| [2010/01/13 12:30:13] @ bodepd joined channel #puppet | ||
| [2010/01/13 12:32:58] @ BarnacleBob joined channel #puppet | ||
| [2010/01/13 12:34:50] <webminster> can anyone help with custom types debugging? | ||
| [2010/01/13 12:36:26] @ Quit: whaley: Remote closed the connection | ||
| [2010/01/13 12:38:09] @ Quit: cynicismic: ""datacenter time..."" | ||
| [2010/01/13 12:38:21] @ Quit: ShiNboi: Client Quit | ||
| [2010/01/13 12:40:35] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 12:43:28] @ mvn071 joined channel #puppet | ||
| [2010/01/13 12:44:21] @ francois left channel #puppet () | ||
| [2010/01/13 12:50:57] @ alexine_dsouza joined channel #puppet | ||
| [2010/01/13 12:58:08] @ Quit: MattyM: "ta ta" | ||
| [2010/01/13 13:01:40] @ Quit: tjoe: Read error: 110 (Connection timed out) | ||
| [2010/01/13 13:02:06] @ ambroff joined channel #puppet | ||
| [2010/01/13 13:02:40] @ ezmobius joined channel #puppet | ||
| [2010/01/13 13:03:31] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/13 13:04:20] @ Quit: maxp: Read error: 101 (Network is unreachable) | ||
| [2010/01/13 13:05:31] @ bug joined channel #puppet | ||
| [2010/01/13 13:10:23] @ mattwynne joined channel #puppet | ||
| [2010/01/13 13:10:54] <hggh> it it possible to use if, elsif and else at erb? | ||
| [2010/01/13 13:11:44] <duritong> hggh you can use any ruby | ||
| [2010/01/13 13:12:24] @ ctso joined channel #puppet | ||
| [2010/01/13 13:16:03] <hggh> duritong: ah ok. thanks | ||
| [2010/01/13 13:16:43] @ Quit: poison: Remote closed the connection | ||
| [2010/01/13 13:17:03] @ poison joined channel #puppet | ||
| [2010/01/13 13:17:38] @ Quit: poison: Remote closed the connection | ||
| [2010/01/13 13:21:55] @ bobbyz joined channel #puppet | ||
| [2010/01/13 13:27:51] @ Quit: Juzzy: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 13:28:07] @ Juzzy joined channel #puppet | ||
| [2010/01/13 13:37:25] @ pting joined channel #puppet | ||
| [2010/01/13 13:43:37] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 13:45:11] @ mattwynne joined channel #puppet | ||
| [2010/01/13 13:49:01] <KarlHungus> how do i prevent concatenated_file from refreshing a service everytime it runs? | ||
| [2010/01/13 13:49:58] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 13:52:05] <KarlHungus> according to the source in concatenated_file.pp it should only run when the file actually changes, but i'm finding that not to be the case | ||
| [2010/01/13 13:53:40] <Volcane> which concatenated_file.pp is this? mine? | ||
| [2010/01/13 13:54:45] <KarlHungus> Volcane: perhaps =) are you David? | ||
| [2010/01/13 13:54:52] <Volcane> nope | ||
| [2010/01/13 13:55:13] <Volcane> http://nephilim.ml.org/~rip/puppet/concatfile/ is the one i wrote fwiw | ||
| [2010/01/13 13:55:14] <KarlHungus> http://git.black.co.at/?p=module-common;a=blob;f=manifests/defines/concatenated_file.pp | ||
| [2010/01/13 13:56:33] @ Quit: ambroff: Read error: 60 (Operation timed out) | ||
| [2010/01/13 14:04:55] @ giskard joined channel #puppet | ||
| [2010/01/13 14:07:01] @ zombied_ joined channel #puppet | ||
| [2010/01/13 14:08:23] @ mattwynne joined channel #puppet | ||
| [2010/01/13 14:08:58] @ whaley joined channel #puppet | ||
| [2010/01/13 14:10:55] @ vzctl_ joined channel #puppet | ||
| [2010/01/13 14:12:03] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 14:13:17] @ mattwynne joined channel #puppet | ||
| [2010/01/13 14:13:51] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 14:19:47] @ Quit: marcoecc: Remote closed the connection | ||
| [2010/01/13 14:21:41] @ ambroff joined channel #puppet | ||
| [2010/01/13 14:22:42] @ Quit: vzctl__: Read error: 113 (No route to host) | ||
| [2010/01/13 14:23:58] @ incommon_ike joined channel #puppet | ||
| [2010/01/13 14:24:19] <incommon_ike> Anyone use RabbitMQ for the StoredConfigs with Puppet? | ||
| [2010/01/13 14:26:16] <Volcane> its stomp connector is a bit odd, suspect it might not work out the box for us | ||
| [2010/01/13 14:28:21] <incommon_ike> ok. thanks. I was looking for a non-java stomp implementation | ||
| [2010/01/13 14:28:53] <incommon_ike> I'll looking into QPID as well. | ||
| [2010/01/13 14:31:54] <chadh> hey Volcane, I just came across mcollective. That's yours, right? | ||
| [2010/01/13 14:33:26] <Volcane> sure is | ||
| [2010/01/13 14:33:58] <chadh> looks cool, but I have a question. How do you deal with/detect hosts that are down? | ||
| [2010/01/13 14:34:11] <chadh> other means? nagios, etc.? | ||
| [2010/01/13 14:34:31] <Volcane> well i dont spool requests to them and rn them at a later stage when those nodes come back | ||
| [2010/01/13 14:34:41] <Volcane> i cant imagine a sane way to do that | ||
| [2010/01/13 14:35:13] <Volcane> but to know if nodes are down I use the http://code.google.com/p/mcollective/wiki/Registration and the monitoring system that page mentions | ||
| [2010/01/13 14:35:16] <chadh> I guess I am more concerned about fact gathering | ||
| [2010/01/13 14:35:24] @ MattyM joined channel #puppet | ||
| [2010/01/13 14:35:38] <Volcane> basically every node can send in a message with whatever you like in it to any or all of the other nodes | ||
| [2010/01/13 14:36:03] <Volcane> but gathering data is real time | ||
| [2010/01/13 14:36:08] <Volcane> there's no database or anything | ||
| [2010/01/13 14:36:12] @ Quit: Juzzy: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 14:36:26] <Volcane> the registration is merely to enable building things like webapps or monitoring or to feed inventories or whatever | ||
| [2010/01/13 14:36:30] @ Juzzy joined channel #puppet | ||
| [2010/01/13 14:36:46] <zombied_> Question. I have a few puppet clients that aren't respecting a puppet class for whatever reason | ||
| [2010/01/13 14:36:54] <zombied_> But most puppet clients are working just fine with that class | ||
| [2010/01/13 14:37:05] <chadh> Volcane: yeah, I like that. | ||
| [2010/01/13 14:37:05] <zombied_> any idea why some puppet clients might ignore a class? | ||
| [2010/01/13 14:37:42] @ Quit: giskard: Success | ||
| [2010/01/13 14:38:33] <chadh> zombied_: we will probably need a little more info. Maybe pastie the error you see in the server logs? | ||
| [2010/01/13 14:38:49] <zombied_> I'm not seeing an error, though | ||
| [2010/01/13 14:38:52] <chadh> I see transient errors when my puppet master is overloaded sometimes | ||
| [2010/01/13 14:39:00] <chadh> oh, the class just isn't being applied? | ||
| [2010/01/13 14:39:16] <zombied_> Yep | ||
| [2010/01/13 14:39:28] <chadh> you sure you have it 'include'd? | ||
| [2010/01/13 14:39:29] <zombied_> basically I have a puppet exec class that I call cmdlauncher | ||
| [2010/01/13 14:39:36] <chadh> you can check the host yaml to see if it is specified there | ||
| [2010/01/13 14:39:41] <zombied_> Yep, because half the puppet clients can work it just fine | ||
| [2010/01/13 14:40:01] <chadh> same configuration on the clients? | ||
| [2010/01/13 14:40:08] @ mattwynne joined channel #puppet | ||
| [2010/01/13 14:40:44] @ crdant joined channel #puppet | ||
| [2010/01/13 14:41:39] @ Quit: dbristow: Read error: 110 (Connection timed out) | ||
| [2010/01/13 14:42:18] <zombied_> http://pastie.org/776886 | ||
| [2010/01/13 14:42:43] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 14:43:13] <chadh> zombied_: so you only want the command to run if the file changes, right? | ||
| [2010/01/13 14:43:27] <zombied_> Yep | ||
| [2010/01/13 14:43:43] <zombied_> And it worked on 2/3rds of the clients | ||
| [2010/01/13 14:43:53] <zombied_> But I got these 11 boxes that basically ignored it | ||
| [2010/01/13 14:43:56] <chadh> did all of them get the /etc/cmdlauncher file? | ||
| [2010/01/13 14:44:24] <zombied_> The ones that work, yes | ||
| [2010/01/13 14:44:32] <zombied_> the ones that aren't working have the old file | ||
| [2010/01/13 14:44:53] @ gebi joined channel #puppet | ||
| [2010/01/13 14:44:53] @ Quit: re_: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 14:45:06] @ re_ joined channel #puppet | ||
| [2010/01/13 14:45:14] <chadh> I know you said there were no errors, but if you overwhelmed the puppetmaster fileserver, that could explain it | ||
| [2010/01/13 14:45:26] <chadh> if you run puppet on those 11 hosts now, would it work? | ||
| [2010/01/13 14:45:38] <zombied_> Nope, I restarted puppet on a broken client manually | ||
| [2010/01/13 14:45:42] <zombied_> and it still doesn't get that file | ||
| [2010/01/13 14:46:04] <zombied_> keep in mind I have 57 clients | ||
| [2010/01/13 14:46:24] <chadh> did you run puppetd --test on one of the broken ones? Is it even trying to fetch the file? | ||
| [2010/01/13 14:46:34] @ mattwynne joined channel #puppet | ||
| [2010/01/13 14:47:50] @ WALoeIII joined channel #puppet | ||
| [2010/01/13 14:48:07] <zombied_> chad: running the test now, but keep in mind these "broken" ones DID get cron updates that I made | ||
| [2010/01/13 14:48:10] @ avleen left channel #puppet () | ||
| [2010/01/13 14:49:26] <chadh> zombied_: yeah, I'm just trying to catch up with you | ||
| [2010/01/13 14:49:28] @ Quit: MattyM: "ta ta" | ||
| [2010/01/13 14:49:36] @ Quit: WALoeIII: Client Quit | ||
| [2010/01/13 14:49:49] @ WALoeIII joined channel #puppet | ||
| [2010/01/13 14:50:18] <chadh> running puppet with --test (and possibly --debug) should be able to point out what the problem is, although maybe not why? :-) | ||
| [2010/01/13 14:53:11] @ Quit: incommon_ike: | ||
| [2010/01/13 14:53:16] <zombied_> running it now with debug | ||
| [2010/01/13 14:53:39] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 14:53:41] <zombied_> err: Configuration retrieval timed out | ||
| [2010/01/13 14:54:35] <zombied_> hrmm, wonder if a FW ys blocking some port connection | ||
| [2010/01/13 14:54:41] <zombied_> but that wouldn't explain why some stuff works | ||
| [2010/01/13 14:54:59] <chadh> zombied_: are you still using webrick? | ||
| [2010/01/13 14:55:37] <zombied_> webrick? | ||
| [2010/01/13 14:56:02] <jb> the default webserver.. | ||
| [2010/01/13 14:56:51] <BarnacleBob> puppetmaster by default runs a webrick webserver in it. there are various other ways of running puppet masters all of which support way more clients than webrick will | ||
| [2010/01/13 14:57:25] <zombied_> Not according to a process list | ||
| [2010/01/13 14:57:39] <Volcane> its internal, u wont see it | ||
| [2010/01/13 14:57:55] <zombied_> So then how can you tell? | ||
| [2010/01/13 14:57:58] <Volcane> wiki:UsingMongrel | ||
| [2010/01/13 14:57:58] <gepetto> Volcane: wiki: wiki:UsingMongrel is http://reductivelabs.com/trac/puppet/wiki/UsingMongrel | ||
| [2010/01/13 14:58:16] <Volcane> zombied_: if you didnt do anything special, like install apache infront of puppetmaster, you're using webrick | ||
| [2010/01/13 14:58:17] <chadh> zombied_: if you aren't sure, you are using webrick | ||
| [2010/01/13 14:58:33] <zombied_> k, I didn't set any of this stuff up | ||
| [2010/01/13 14:58:46] <zombied_> So all I'm learning right now is managing the existing configs, not how it was deployed in the first place | ||
| [2010/01/13 14:58:51] <Volcane> does ps show just one copy of puppetmasterd? | ||
| [2010/01/13 14:59:06] <zombied_> yep | ||
| [2010/01/13 14:59:10] <Volcane> then you're using webrick | ||
| [2010/01/13 14:59:17] <Volcane> and should expect problems around 20 hosts or so | ||
| [2010/01/13 15:01:13] <zombied_> So basically if you aren't using apache with mongrel | ||
| [2010/01/13 15:01:20] <zombied_> expect random problems with random puppet clients | ||
| [2010/01/13 15:01:29] <Volcane> or the other solutions mentioned on that page | ||
| [2010/01/13 15:01:37] <Volcane> like apache+passenger, or nginx+mongrel etcetc | ||
| [2010/01/13 15:02:11] <zombied_> is migration out of webrick easy? | ||
| [2010/01/13 15:02:13] @ skiold joined channel #puppet | ||
| [2010/01/13 15:02:24] <Volcane> zombied_: wiki:UsingMongrel | ||
| [2010/01/13 15:02:24] <gepetto> Volcane: zombied_: wiki:UsingMongrel is http://reductivelabs.com/trac/puppet/wiki/UsingMongrel | ||
| [2010/01/13 15:02:51] @ Quit: billy_ran_away: | ||
| [2010/01/13 15:02:59] <zombied_> Is there any way to confirm webrick is the problem? | ||
| [2010/01/13 15:03:00] @ bodepd left channel #puppet () | ||
| [2010/01/13 15:03:26] <Volcane> when your clients run do you see plenty of SYN_RECV sockets in netstat on the master? | ||
| [2010/01/13 15:04:49] <BarnacleBob> awesome | ||
| [2010/01/13 15:04:49] <zombied_> yep, on port 8140? | ||
| [2010/01/13 15:04:58] <BarnacleBob> i upgraded facter and its failing all over the place | ||
| [2010/01/13 15:05:23] <BarnacleBob> anyone able to help me with this facter error? http://pastie.org/776919 | ||
| [2010/01/13 15:05:41] <Volcane> yes | ||
| [2010/01/13 15:05:48] <Volcane> 8140 | ||
| [2010/01/13 15:06:07] <Volcane> BarnacleBob: how did you upgrade? | ||
| [2010/01/13 15:06:10] @ dbristow joined channel #puppet | ||
| [2010/01/13 15:07:14] <BarnacleBob> Volcane, well i built a new .deb using uupdate which did it cleanly | ||
| [2010/01/13 15:07:25] <BarnacleBob> then just a apt-get install facter from my custom repo | ||
| [2010/01/13 15:07:27] <BarnacleBob> *but* | ||
| [2010/01/13 15:07:35] <Volcane> ah good, youre already tons better than almost everyone who have issues with facter :P | ||
| [2010/01/13 15:07:41] <BarnacleBob> someone at some point turns out had manually installed a version of facter some where | ||
| [2010/01/13 15:07:45] <BarnacleBob> but i rm -r'ed it | ||
| [2010/01/13 15:07:47] <Volcane> oh that'll be it | ||
| [2010/01/13 15:07:53] <BarnacleBob> but its gone | ||
| [2010/01/13 15:07:56] <Volcane> ok | ||
| [2010/01/13 15:08:01] <BarnacleBob> at least slocate can't find any facter files | ||
| [2010/01/13 15:08:12] <Volcane> well it could be left overs, cold be that you didnt restart puppet or that your deb sux | ||
| [2010/01/13 15:08:44] <BarnacleBob> well the deb worked fine on several test machines (that didn't have facter manually installed, same with the place i built it) | ||
| [2010/01/13 15:09:58] <Volcane> somethings left over somewhere i'd say | ||
| [2010/01/13 15:11:23] <BarnacleBob> omg | ||
| [2010/01/13 15:11:32] <BarnacleBob> it was /root/facter-1.5.2 that was in its way | ||
| [2010/01/13 15:11:37] <BarnacleBob> *sigh* | ||
| [2010/01/13 15:12:57] <Volcane> heh | ||
| [2010/01/13 15:17:23] <BarnacleBob> i'm 100% positive it was way more work for this guy to manually install facter like this than just to build the fricken package | ||
| [2010/01/13 15:17:29] <BarnacleBob> why WHY WHYYYY!Y!!!!! | ||
| [2010/01/13 15:17:29] <BarnacleBob> :) | ||
| [2010/01/13 15:18:09] <jb> because he was lazy? | ||
| [2010/01/13 15:18:10] <jb> :) | ||
| [2010/01/13 15:18:40] <brd> s/lazy/ignorant/ | ||
| [2010/01/13 15:20:48] @ Djelibeybi joined channel #puppet | ||
| [2010/01/13 15:24:34] @ jes5 joined channel #puppet | ||
| [2010/01/13 15:26:22] <zombied_> Volcane: so all I have to do is a yum install httpd on the puppetmaster | ||
| [2010/01/13 15:26:26] <zombied_> and copy that config over? | ||
| [2010/01/13 15:26:58] <ashp> earlier I had 24 nodes in mc-ping, now 14 | ||
| [2010/01/13 15:27:00] <ashp> this is very weird. | ||
| [2010/01/13 15:28:30] <ashp> root 14204 0.0 0.2 66416 17312 ? Sl 11:10 0:00 /usr/bin/ruby /usr/sbin/mcollectived --pid=/var/run/mcollectived.pid --config=/etc/mcollective/server.cfg | ||
| [2010/01/13 15:28:35] <ashp> It's definitely running. Volcane, you here? | ||
| [2010/01/13 15:28:51] <tim|imac> hm... I'm reading that puppet is getting a ruby based DSL? | ||
| [2010/01/13 15:28:59] <Volcane> ashp: #mcollective | ||
| [2010/01/13 15:29:04] <ashp> join #mcollective | ||
| [2010/01/13 15:29:06] <ashp> argh | ||
| [2010/01/13 15:29:06] <Volcane> tim|imac: yes, an additional option | ||
| [2010/01/13 15:29:19] <tim|imac> ugh... that would be a bother... | ||
| [2010/01/13 15:29:27] <Volcane> u dont have to use it | ||
| [2010/01/13 15:29:44] <BarnacleBob> could be much nicer | ||
| [2010/01/13 15:29:48] <tim|imac> no, but if I ask for some ideas in here, people are bound to say something like "use the ruby based dsl" | ||
| [2010/01/13 15:29:53] <BarnacleBob> i'm using crazy hacks to use ruby in the dsl | ||
| [2010/01/13 15:30:31] <zombied_> Using mongrel wiki isn't that clear, vol | ||
| [2010/01/13 15:30:31] <BarnacleBob> this makes me sad face | ||
| [2010/01/13 15:30:31] <BarnacleBob> $minutes=split(inline_template("<% require 'ipaddr' %><%= IPAddr.new(ipaddress).to_i % 60 %>"),':') | ||
| [2010/01/13 15:30:52] <Volcane> tim|imac: nods, well the final structure of it isnt defined but its in the pipeline | ||
| [2010/01/13 15:31:08] <tim|imac> yeah, there are some cases in which i'd like some more options to represent stuff in my head, but i don't want to learn ruby to do it :/ | ||
| [2010/01/13 15:31:58] <tim|imac> the current dsl, although sometimes lacking some power, is usually powerful enough :/ | ||
| [2010/01/13 15:32:52] <Volcane> tim|imac: yeah i prolly wont be using it as an instead, but i tink there are cases where it will fill gaps betwen dsl based defines and types | ||
| [2010/01/13 15:33:09] <tim|imac> Volcane: any idea if they're supposed to be used next to each other? so you can supplement "normal" DSL with some ruby code? | ||
| [2010/01/13 15:33:17] <Volcane> tim|imac: apparently | ||
| [2010/01/13 15:33:25] <Volcane> tim|imac: thats what lak said anyway but | ||
| [2010/01/13 15:33:29] <Volcane> early days | ||
| [2010/01/13 15:33:36] <tim|imac> ok, I could live with that, I guess | ||
| [2010/01/13 15:34:07] <Volcane> like i asked can i drop in modname/manifests/foo.rb | ||
| [2010/01/13 15:34:11] <Volcane> and make a class/define there | ||
| [2010/01/13 15:34:13] @ yarihm joined channel #puppet | ||
| [2010/01/13 15:34:18] <Volcane> in addition to my other .pp files | ||
| [2010/01/13 15:34:21] <Volcane> he said thats the idea | ||
| [2010/01/13 15:34:24] <Volcane> but thats as much as i know :P | ||
| [2010/01/13 15:34:48] <Volcane> there's some mails on the dev list re it | ||
| [2010/01/13 15:35:15] <tim|imac> i'll check em out, thanks for the info :D | ||
| [2010/01/13 15:38:04] <zombied_> volcane: So all I have to do is a yum install httpd, edit the httpd.conf file so that it looks just like the config in the wiki you mentioned, replace puppetmaster with the IP/FQDN of the puppetmaster, and then start puppetmasterd with the switch --servertype mongrel | ||
| [2010/01/13 15:38:06] <zombied_> Is that correct? | ||
| [2010/01/13 15:38:31] <Volcane> zombied_: thats more or less what the wiki says - but really, the wiki is long, detailed and knows more than i can possibly answer on irc | ||
| [2010/01/13 15:40:07] <BarnacleBob> zombied_, you have to install mongrel also | ||
| [2010/01/13 15:40:50] @ Quit: pting: Read error: 110 (Connection timed out) | ||
| [2010/01/13 15:42:57] @ fsweetser joined channel #puppet | ||
| [2010/01/13 15:43:04] @ pting joined channel #puppet | ||
| [2010/01/13 15:45:15] <zombied_> BarnacleBob: Is it called something else in the yum repos? | ||
| [2010/01/13 15:45:20] <zombied_> No Match for argument: mongrel | ||
| [2010/01/13 15:45:31] <BarnacleBob> zombied_, yeah | ||
| [2010/01/13 15:45:38] <BarnacleBob> just do a yum list | grep -i mongrel | ||
| [2010/01/13 15:45:42] <BarnacleBob> its like ruby-mongrel or something | ||
| [2010/01/13 15:46:15] <zombied_> Didn't find it | ||
| [2010/01/13 15:46:49] <Volcane> EPEL has it | ||
| [2010/01/13 15:50:33] @ Quit: toi: "Ex-Chat" | ||
| [2010/01/13 15:50:54] <zombied_> You sure volcane? | ||
| [2010/01/13 15:51:05] <zombied_> I added the repo | ||
| [2010/01/13 15:51:09] <zombied_> And it didn't find it | ||
| [2010/01/13 15:51:57] <Volcane> http://download.fedora.redhat.com/pub/epel/5/i386/repoview/rubygem-mongrel.html | ||
| [2010/01/13 15:52:29] <zombied_> oh, im running centos 4.8 | ||
| [2010/01/13 15:53:14] <zombied_> Starting httpd: Syntax error on line 3 of /etc/httpd/conf/httpd.conf: Invalid command 'ProxyRequests', perhaps misspelled or defined by a module not included in the server configuration [FAILED] | ||
| [2010/01/13 15:53:18] <BarnacleBob> you can install it from gems but i highly recommend you don't go down that rabbit hole | ||
| [2010/01/13 15:53:31] <BarnacleBob> zombied_, you need to install modproxy also | ||
| [2010/01/13 15:55:26] <zombied_> man, EPEL doesn't have any of this stuff for 4.8 | ||
| [2010/01/13 15:55:31] @ Quit: mvn071: Remote closed the connection | ||
| [2010/01/13 15:56:13] @ Quit: bjaspan: | ||
| [2010/01/13 15:57:47] @ verwilst_ joined channel #puppet | ||
| [2010/01/13 15:59:50] @ poison joined channel #puppet | ||
| [2010/01/13 15:59:55] @ nakano_ is now known as nakano | ||
| [2010/01/13 16:01:18] <bobbyz> Hi guys, I have a quick question. I have a node that is includes two classes: baseclass and postfix::disabled, where baseclass includes a class postfix. I'm essentially trying to follow the example in http://reductivelabs.com/trac/puppet/wiki/PuppetBestPractice#overrides | ||
| [2010/01/13 16:01:35] <bobbyz> However, puppet fails the catalog with " Could not retrieve catalog: Puppet::Parser::AST::Resource failed with error ArgumentError: Duplicate definition: Package[postfix] is already defined in file /etc/puppet/modules/postfix/manifests/init.pp at line 3; cannot redefine at /etc/puppet/modules/postfix/manifests/init.pp:39 on node testprov.testdomain" | ||
| [2010/01/13 16:02:16] <bobbyz> Is this overriding something that was added post- pupper-0.24.8? | ||
| [2010/01/13 16:02:30] <zombied_> Doesn't look like there was a mongrel build for rhel 4 | ||
| [2010/01/13 16:02:41] <bobbyz> Or am I misunderstanding the overrides example in the best practices guide? | ||
| [2010/01/13 16:03:36] <Volcane> u can only override if you inherit | ||
| [2010/01/13 16:04:31] <Volcane> prolly | ||
| [2010/01/13 16:04:36] <Volcane> oops., wrong # | ||
| [2010/01/13 16:06:24] @ nevyn_ joined channel #puppet | ||
| [2010/01/13 16:07:07] <bobbyz> yeah, I have class postfix::disabled inherits postfix, and then the overriding Package def | ||
| [2010/01/13 16:13:19] @ mattwynne joined channel #puppet | ||
| [2010/01/13 16:16:25] @ mikerowehl joined channel #puppet | ||
| [2010/01/13 16:16:37] @ neek joined channel #puppet | ||
| [2010/01/13 16:17:32] @ jfreeman joined channel #puppet | ||
| [2010/01/13 16:17:39] @ Quit: nevyn: Read error: 101 (Network is unreachable) | ||
| [2010/01/13 16:18:11] <neek> ok, need a pointer. I need to install an rpm that isn't in any yum repo. can I just source => puppet:/// it somehow? | ||
| [2010/01/13 16:18:43] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 16:19:43] <whack> why not add it to your yum repo? | ||
| [2010/01/13 16:20:07] <neek> because I don't run any of them. :) | ||
| [2010/01/13 16:20:14] <chadh> neek: you have to either copy the rpm locally (ie, via file resource), or snag it from an http source via the package source attribute | ||
| [2010/01/13 16:20:42] @ Quit: jfreeman: Remote closed the connection | ||
| [2010/01/13 16:20:59] <neek> chadh: ok, so I stick in $tmp with a file {}. then tell package {... source => $tmp/my.rpm } ? | ||
| [2010/01/13 16:21:12] <whack> rpm can install via http | ||
| [2010/01/13 16:21:18] <whack> ie; rpm -Uvh http://foo/bar/baz.rpm | ||
| [2010/01/13 16:21:19] <whack> will work | ||
| [2010/01/13 16:21:21] <webminster> question: is it possible to subscribe to a "define" resource, or only to the resources the define would create? | ||
| [2010/01/13 16:21:34] <neek> I'd rather not depend on an outside http for the file... | ||
| [2010/01/13 16:21:42] <chadh> neek: yeah, although you don't want to put it somewhere temporary, or it will keep downloading it (or fail to install it if it goes away) | ||
| [2010/01/13 16:22:03] <whack> you can probably do package { "mypackage": source => "http://....", provider => rpm } | ||
| [2010/01/13 16:22:04] <neek> chadh: good point. could do an onlyif, but probably not worth it. | ||
| [2010/01/13 16:22:38] <chadh> neek: right. http or nfs sources are best, but you can probably hack something else together | ||
| [2010/01/13 16:22:46] <whack> webminster: yes, if you'r define is called 'foo' you can access the resource (for require, etc) as Foo["name"] | ||
| [2010/01/13 16:22:50] <chadh> neek: well, other than your own repo, of course ;-) | ||
| [2010/01/13 16:22:58] @ jfreeman joined channel #puppet | ||
| [2010/01/13 16:23:09] <whack> without http, you could do | ||
| [2010/01/13 16:23:37] <whack> file { "thing.rpm": source => "puppet:///..." } package { "thing": source => "/path/to/thing.rpm", require => File["thing.rpm"] } | ||
| [2010/01/13 16:24:05] <webminster> whack: thx, and if the define was mysql::user for instance, that could be Mysql::user? (cap letter prefix?) | ||
| [2010/01/13 16:24:11] <whack> webminster: Mysql::User | ||
| [2010/01/13 16:24:32] <whack> webminster: I have a daemontools::service define, and require using Daemontools::Service["thing"] | ||
| [2010/01/13 16:24:36] <whack> fwiw | ||
| [2010/01/13 16:25:18] <webminster> thx | ||
| [2010/01/13 16:26:41] <gepetto> ::trac:: Recipes/Authorized_keys edited @ http://reductivelabs.com/trac/puppet/wiki/Recipes/Authorized_keys?version=46 | ||
| [2010/01/13 16:29:44] @ Quit: skiold: | ||
| [2010/01/13 16:37:32] @ Quit: bsund: "Leaving" | ||
| [2010/01/13 16:39:23] @ Quit: Newsome: "Linux: Now with employee pricing!" | ||
| [2010/01/13 16:39:30] @ Quit: bug: Read error: 110 (Connection timed out) | ||
| [2010/01/13 16:48:33] @ jab_doa joined channel #puppet | ||
| [2010/01/13 16:50:46] @ sebas891 joined channel #puppet | ||
| [2010/01/13 16:52:26] @ giskard joined channel #puppet | ||
| [2010/01/13 16:53:06] <sebas891> hi folks, I'm having a problem on puppet clients, it hangs at "info: Loading fact mountpoints" for about 45 secs during runs. | ||
| [2010/01/13 16:53:32] <sebas891> I'm using the debian package in squeeze. | ||
| [2010/01/13 16:55:42] <gepetto> ::trac:: Recipes/ModuleSSHAuth edited @ http://reductivelabs.com/trac/puppet/wiki/Recipes/ModuleSSHAuth?version=15 | ||
| [2010/01/13 16:56:18] @ kolla joined channel #puppet | ||
| [2010/01/13 17:02:23] @ Quit: lilmatt: Client Quit | ||
| [2010/01/13 17:11:26] <neek> ...is there a --nodeps for provider => rpm? | ||
| [2010/01/13 17:11:46] * neek cant find it | ||
| [2010/01/13 17:12:12] @ webminster left channel #puppet () | ||
| [2010/01/13 17:12:45] <Volcane> no | ||
| [2010/01/13 17:12:53] <Volcane> why would puppet help you shoot yourself in the head? | ||
| [2010/01/13 17:13:51] @ Quit: labrown: "Leaving" | ||
| [2010/01/13 17:13:52] @ Quit: jaredrhine: Read error: 110 (Connection timed out) | ||
| [2010/01/13 17:14:16] <neek> Volcane: er. because sometimes rpms are just fucking broken? | ||
| [2010/01/13 17:15:49] @ jcape joined channel #puppet | ||
| [2010/01/13 17:16:24] <Volcane> so not installing broken rpms is a feature | ||
| [2010/01/13 17:17:14] <neek> I'm almost to the point of 'screw puppet for this one, I'm just going to install it and maintain it by hand" | ||
| [2010/01/13 17:17:38] <Volcane> why not fix the rpm? | ||
| [2010/01/13 17:18:00] <neek> no build environments. | ||
| [2010/01/13 17:18:17] <neek> and even if there were, I suck at speaking rpm. :) | ||
| [2010/01/13 17:20:36] <gepetto> ::trac:: Recipes/ModuleSSHAuth edited @ http://reductivelabs.com/trac/puppet/wiki/Recipes/ModuleSSHAuth?version=16 | ||
| [2010/01/13 17:29:22] @ Quit: taihen_: Read error: 113 (No route to host) | ||
| [2010/01/13 17:29:45] @ nevyn_ is now known as nevyn | ||
| [2010/01/13 17:30:22] @ taihen joined channel #puppet | ||
| [2010/01/13 17:30:43] @ nmi joined channel #puppet | ||
| [2010/01/13 17:31:02] @ joe-mac left channel #puppet () | ||
| [2010/01/13 17:32:28] @ Quit: Demosthenes: "leaving" | ||
| [2010/01/13 17:35:39] @ mattwynne joined channel #puppet | ||
| [2010/01/13 17:37:33] @ svend joined channel #puppet | ||
| [2010/01/13 17:37:54] <svend> What is the oldest version of Puppet that supports modules? | ||
| [2010/01/13 17:39:42] @ vorlon2261 joined channel #puppet | ||
| [2010/01/13 17:39:48] <BarnacleBob> so i just upgraded to 25.3 and i'm getting "-":6: bad command \n errors in crontab file, can't install. but only on one machine. anyone have any ideas? | ||
| [2010/01/13 17:40:09] <vorlon2261> Which OS? | ||
| [2010/01/13 17:40:20] <BarnacleBob> ubuntu 8.05 | ||
| [2010/01/13 17:40:22] <BarnacleBob> 8.04 | ||
| [2010/01/13 17:40:24] <vorlon2261> Some versions of cron don't like newlines in the crontab... | ||
| [2010/01/13 17:40:32] <BarnacleBob> thats the error output | ||
| [2010/01/13 17:40:34] <BarnacleBob> not the cron | ||
| [2010/01/13 17:40:34] <BarnacleBob> sorry | ||
| [2010/01/13 17:40:39] <BarnacleBob> just didn't want to paste extra lines | ||
| [2010/01/13 17:40:39] <vorlon2261> Does the existing crontab have any blank lines? | ||
| [2010/01/13 17:40:45] @ tmd left channel #puppet () | ||
| [2010/01/13 17:40:48] <BarnacleBob> the existing crontab is empty | ||
| [2010/01/13 17:41:51] <BarnacleBob> i wish there was a way to see the crontab puppet is writing | ||
| [2010/01/13 17:42:04] <BarnacleBob> i have no idea how to do that tho | ||
| [2010/01/13 17:42:17] <vorlon2261> It should be creating a temp file somewhere on the client... | ||
| [2010/01/13 17:42:26] <vorlon2261> Problem is that it won't exist for very long... | ||
| [2010/01/13 17:42:52] <BarnacleBob> yeah | ||
| [2010/01/13 17:43:31] @ Quit: dbristow: "Leaving" | ||
| [2010/01/13 17:44:37] <BarnacleBob> there is something wrong with this | ||
| [2010/01/13 17:45:28] <vorlon2261> Sounds like it... It sounds like its putting in extra newlines, which cron doesn't like... | ||
| [2010/01/13 17:45:47] <vorlon2261> I'm still on 0.25.1, so can't do any tests to check for you... | ||
| [2010/01/13 17:50:03] @ Quit: themurph: Read error: 60 (Operation timed out) | ||
| [2010/01/13 17:51:41] @ Bevo joined channel #puppet | ||
| [2010/01/13 17:51:51] <whack> BarnacleBob: fwiw, I don't use puppet's default cron resource | ||
| [2010/01/13 17:52:13] <whack> I have a custom define that installs cron jobs, one per file, to /etc/cron.d | ||
| [2010/01/13 17:52:38] <whack> though I'm still on 0.25.1 so I can't test your problem :( | ||
| [2010/01/13 17:53:08] <hggh> whack: #2428 | ||
| [2010/01/13 17:53:09] <gepetto> hggh: whack: #2428 is http://projects.reductivelabs.com/issues/show/2428 "Puppet - Bug #2428: cron should support non-user targets - ReductiveLabs.com" | ||
| [2010/01/13 17:53:19] <Bevo> Anyone have any idea why an exported resource wont appear on a node by using a tag File <<| tag == 'exportedfile' |>> but will work fine with File <<| title == '/path/to/file' |>> | ||
| [2010/01/13 17:53:22] <BarnacleBob> i hacked in something to print out my crontab | ||
| [2010/01/13 17:53:29] <BarnacleBob> well print out what its going to write | ||
| [2010/01/13 17:53:33] <BarnacleBob> and it writes it perfectly fine | ||
| [2010/01/13 17:54:19] <BarnacleBob> oh | ||
| [2010/01/13 17:54:24] <BarnacleBob> its not a valid crontab | ||
| [2010/01/13 17:54:27] <BarnacleBob> its printing one too many fields | ||
| [2010/01/13 17:54:33] <BarnacleBob> * 12 * * * * /nail/sys/bin/puppetrun 60 | ||
| [2010/01/13 17:54:36] <BarnacleBob> there is only 5 right | ||
| [2010/01/13 17:54:40] <BarnacleBob> minutes first | ||
| [2010/01/13 17:54:48] <BarnacleBob> something is broken on this machine | ||
| [2010/01/13 17:54:58] <Bevo> is this for a per user crontab? | ||
| [2010/01/13 17:55:03] <BarnacleBob> yeah | ||
| [2010/01/13 17:55:03] <Bevo> or /etc/crontab? | ||
| [2010/01/13 17:55:10] <BarnacleBob> root cron | ||
| [2010/01/13 17:55:32] <Bevo> so as root; crontab -e ? | ||
| [2010/01/13 17:55:46] <Bevo> if its /etc/crontab it needs a user field, | ||
| [2010/01/13 17:55:55] <BarnacleBob> yeah its crontab -e | ||
| [2010/01/13 17:56:01] <BarnacleBob> so 5 fields | ||
| [2010/01/13 17:56:05] <BarnacleBob> puppet is adding an extra | ||
| [2010/01/13 17:56:15] <Bevo> whats the config you have in puppet? | ||
| [2010/01/13 17:56:23] @ M- joined channel #puppet | ||
| [2010/01/13 17:56:56] <BarnacleBob> http://pastie.org/777177 | ||
| [2010/01/13 17:57:07] <BarnacleBob> it only doesn't work on this one machine of 600+ | ||
| [2010/01/13 17:57:40] <whack> hggh: I'll always need a custom define for cron, though, since I wrap all jobs with a helper script and take other parameters | ||
| [2010/01/13 17:57:42] <Bevo> same distributions? | ||
| [2010/01/13 17:57:57] <BarnacleBob> for the most part same distros | ||
| [2010/01/13 17:58:04] <BarnacleBob> i think ~400 on the same distro as this | ||
| [2010/01/13 17:58:09] <BarnacleBob> rest are earlier ubuntu | ||
| [2010/01/13 17:58:30] <Bevo> BarnacleBob, is there other crons in that machine crontab file | ||
| [2010/01/13 17:59:42] <BarnacleBob> nope | ||
| [2010/01/13 17:59:45] <BarnacleBob> hrm | ||
| [2010/01/13 18:00:00] <BarnacleBob> i'm finding more problems with puppet also | ||
| [2010/01/13 18:00:00] <Bevo> does it change by removing the environment argument | ||
| [2010/01/13 18:00:28] <BarnacleBob> i don't have puppet on the machine anymore at the moment | ||
| [2010/01/13 18:00:32] <BarnacleBob> i tried to reinstall it :) | ||
| [2010/01/13 18:00:42] @ bug joined channel #puppet | ||
| [2010/01/13 18:00:53] <Bevo> Anyone have any idea why an exported resource wont appear on a node by using a tag File <<| tag == 'exportedfile' |>> but will work fine with File <<| title == '/path/to/file' |>> | ||
| [2010/01/13 18:01:16] <Volcane> did u tag the file? | ||
| [2010/01/13 18:01:39] @ Quit: mattwynne: Remote closed the connection | ||
| [2010/01/13 18:01:54] <Bevo> Volcane, @@file { "/etc/cluster_nodes/$hostname": content => "${ipaddress}\n", tag => ["exportedClusterNodes"] } | ||
| [2010/01/13 18:02:45] <hggh> why do you make an array as tag? | ||
| [2010/01/13 18:03:05] @ jaredrhine joined channel #puppet | ||
| [2010/01/13 18:03:17] <Bevo> hggh, i've tried both ways | ||
| [2010/01/13 18:04:11] <Bevo> hggh, should have fixed that up in my paste originally i had it as tag => "exportedClusterNodes" | ||
| [2010/01/13 18:07:10] <Volcane> try lowercasing it | ||
| [2010/01/13 18:08:09] <BarnacleBob> blargh | ||
| [2010/01/13 18:08:13] <BarnacleBob> still getting this cron issue | ||
| [2010/01/13 18:09:42] <Bevo> BarnacleBob, try adding another cron for that hostname and see if its repeatable | ||
| [2010/01/13 18:10:09] <Bevo> BarnacleBob, try adding another cron for that hostname and see if its repeatable | ||
| [2010/01/13 18:10:11] <BarnacleBob> i just manually inserted the cron | ||
| [2010/01/13 18:10:14] <Bevo> oops | ||
| [2010/01/13 18:10:17] @ lak joined channel #puppet | ||
| [2010/01/13 18:10:17] <BarnacleBob> now puppet doesn't try to add it again | ||
| [2010/01/13 18:10:38] <BarnacleBob> the machine has bad ram. whatever is wrong its gonna get a reinstall soon any way | ||
| [2010/01/13 18:11:06] <Bevo> Volcane, that worked! why cant tags have upper case? | ||
| [2010/01/13 18:11:25] <Volcane> Bevo: apparently not, and now you know why you should never show us pseudo or hypothetical cde | ||
| [2010/01/13 18:11:35] <Volcane> Bevo: since your fake sample of the problem code were all lower case | ||
| [2010/01/13 18:11:41] <Volcane> Bevo: resulting in wasting our time and yours | ||
| [2010/01/13 18:12:10] @ Quit: lak: Client Quit | ||
| [2010/01/13 18:12:25] <himanshu> how can I ensure service is not running | ||
| [2010/01/13 18:12:33] <BarnacleBob> ensure=>stopped | ||
| [2010/01/13 18:12:34] @ bindr joined channel #puppet | ||
| [2010/01/13 18:12:50] <Bevo> Volcane, true appologies *slaps myself* | ||
| [2010/01/13 18:13:01] <himanshu> so what is some resource try to notify it? | ||
| [2010/01/13 18:13:12] <Bevo> Volcane, im curious as to why you cant use upper case though in a tag? | ||
| [2010/01/13 18:13:19] <BarnacleBob> if something notifies it it will ensure its still stopped | ||
| [2010/01/13 18:13:32] <himanshu> ohk | ||
| [2010/01/13 18:14:01] <Volcane> Bevo: not sure, docuentation doesnt say you cant | ||
| [2010/01/13 18:14:06] <himanshu> suppose if I start this service manually and say puppet client again sync with master, will it stop this service then? | ||
| [2010/01/13 18:14:11] <Volcane> Bevo: just seemed plausable | ||
| [2010/01/13 18:14:32] @ lak joined channel #puppet | ||
| [2010/01/13 18:14:46] <Bevo> Volcane, didnt even think of it because i looked in the database and its stored exactly the same (ie with the upper case) | ||
| [2010/01/13 18:14:51] <neek> himanshu: yes. | ||
| [2010/01/13 18:14:57] <himanshu> oh | ||
| [2010/01/13 18:15:03] <Volcane> Bevo: might be worth filing a bug perhaps | ||
| [2010/01/13 18:15:09] <neek> stopped will make it stop. running will make it run. | ||
| [2010/01/13 18:15:27] <neek> sometimes running will make it run every time because service xxx status doesnt always behave right.. | ||
| [2010/01/13 18:15:57] <Volcane> neek: by default it doesnt do service foo status, it greps the ps list | ||
| [2010/01/13 18:16:00] <Bevo> Volcane, will do | ||
| [2010/01/13 18:16:18] <himanshu> what is default behaviour if I wont specify ensure keyword in service | ||
| [2010/01/13 18:16:52] <himanshu> actually, I want to manage start, stop of service outside the puppet configs | ||
| [2010/01/13 18:17:09] <himanshu> and when I install package for first time, service should be stopped | ||
| [2010/01/13 18:18:37] @ Quit: pheezy: Remote closed the connection | ||
| [2010/01/13 18:19:31] @ Quit: ctso: "Leaving." | ||
| [2010/01/13 18:20:20] <neek> volcane: interesting. what would it ps for? | ||
| [2010/01/13 18:20:26] <Volcane> neek: the service name | ||
| [2010/01/13 18:20:35] <Volcane> service{"foo":...} greps for foo | ||
| [2010/01/13 18:20:51] <Volcane> not all operating systems provides status, so hasstatus => true is needed | ||
| [2010/01/13 18:20:52] <neek> that sounds kinda crummy considering how frequently service names and proc names dont match... | ||
| [2010/01/13 18:21:09] <BarnacleBob> it is crummy | ||
| [2010/01/13 18:21:14] <BarnacleBob> but if you run redhat based os | ||
| [2010/01/13 18:21:17] @ Quit: crdant: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 18:21:21] <BarnacleBob> you can almost always just do hasstatus=>true | ||
| [2010/01/13 18:21:25] <BarnacleBob> you could even default it | ||
| [2010/01/13 18:21:25] <Volcane> though with debian kind of reaching the 1999 equivelant of other distros it would be a nice good to change the default behaviour :) | ||
| [2010/01/13 18:22:00] <BarnacleBob> i wouldn't consider ubuntu in the same decade behind as debian, but it still has 0 status scripts | ||
| [2010/01/13 18:22:00] <Volcane> s/a nice good to/good to | ||
| [2010/01/13 18:22:10] <Volcane> ah, sux | ||
| [2010/01/13 18:22:25] <BarnacleBob> i think just auto setting the default based on os would be nice | ||
| [2010/01/13 18:22:28] @ Quit: WALoeIII: "Bai." | ||
| [2010/01/13 18:22:38] <neek> I would expect servce {} to follow the convention of the operating system (so rhel uses service status, freebsd does something entirely different, etc) | ||
| [2010/01/13 18:22:53] <Volcane> even redhat has broken init scripts | ||
| [2010/01/13 18:22:57] <BarnacleBob> yeah | ||
| [2010/01/13 18:22:58] <neek> oh, pleny | ||
| [2010/01/13 18:23:02] <BarnacleBob> but at least lots of them do have a status | ||
| [2010/01/13 18:23:04] <Volcane> so it would mostly fail | ||
| [2010/01/13 18:23:05] <neek> but defaultedly... | ||
| [2010/01/13 18:23:18] <neek> from what I see it mostly fails already. :) | ||
| [2010/01/13 18:23:19] <Volcane> and we'd just be having the same conversation | ||
| [2010/01/13 18:23:21] @ lilmatt joined channel #puppet | ||
| [2010/01/13 18:24:32] * neek returns to the land of hacky execs, borken requires, and missing statuses. | ||
| [2010/01/13 18:24:55] <BarnacleBob> just evaluate each service itself | ||
| [2010/01/13 18:25:04] <BarnacleBob> if it has a status that works | ||
| [2010/01/13 18:25:06] <BarnacleBob> use it | ||
| [2010/01/13 18:25:14] @ WALoeIII joined channel #puppet | ||
| [2010/01/13 18:28:22] <neek> now if someone could make nfs mounts not try to remount... ;) | ||
| [2010/01/13 18:29:21] <kjetilho> neek: huh? NFS mounts don't remount | ||
| [2010/01/13 18:29:29] <kjetilho> unless you're using an automounter | ||
| [2010/01/13 18:29:49] <neek> http://projects.reductivelabs.com/issues/2731 | ||
| [2010/01/13 18:30:25] <kjetilho> ah, that | ||
| [2010/01/13 18:30:33] <kjetilho> sorry | ||
| [2010/01/13 18:30:42] <neek> :) | ||
| [2010/01/13 18:31:32] @ Quit: vinci: Read error: 60 (Operation timed out) | ||
| [2010/01/13 18:32:06] <kjetilho> I was challenged by Markus to provide a patch since I was only complaining his wasn't any good :) | ||
| [2010/01/13 18:32:16] <kjetilho> not for the exit code of mount, though | ||
| [2010/01/13 18:32:28] <kjetilho> the topic of the bug drifted... | ||
| [2010/01/13 18:32:31] <neek> heh | ||
| [2010/01/13 18:33:03] @ Quit: kaptk2: "Leaving." | ||
| [2010/01/13 18:34:53] @ Robbie_ joined channel #puppet | ||
| [2010/01/13 18:35:56] <mqr> kjetilho: So, do you have one? | ||
| [2010/01/13 18:36:34] @ vinci joined channel #puppet | ||
| [2010/01/13 18:36:35] <kjetilho> mqr: problem is I'm more of a C and Perl hacker :/ | ||
| [2010/01/13 18:36:53] <kjetilho> but it's basically just syscalls, so it should be pretty similar | ||
| [2010/01/13 18:37:32] <mqr> I think what's suckering a lot of people in (myself included) is that the ruby methods have names identical to stdlib routines. | ||
| [2010/01/13 18:38:13] <mqr> And they even call those routines (generally, but not always). | ||
| [2010/01/13 18:38:20] <kjetilho> hmm | ||
| [2010/01/13 18:38:23] <mqr> But they also do other stuff. | ||
| [2010/01/13 18:38:30] <mqr> And there in lies the rub. | ||
| [2010/01/13 18:38:41] <kjetilho> would be good to fix it, a related problem is Puppet leaving defunct processes behind (specifically "mount") | ||
| [2010/01/13 18:38:59] <mqr> ruby can sort of look like C, or perl, etc. but it's got its own semantics. | ||
| [2010/01/13 18:39:00] <kjetilho> not sure if that one is fixed in 0.25.x series, we're still running 0.24.8 here | ||
| [2010/01/13 18:39:20] @ crdant joined channel #puppet | ||
| [2010/01/13 18:40:01] <mqr> I _think_ that is fixed in the current 0.25.x, but I'm an empiricist; I'll change my believes in the face of evidice. | ||
| [2010/01/13 18:40:10] <kjetilho> hehe | ||
| [2010/01/13 18:40:25] <mqr> s/belives/beliefs/ | ||
| [2010/01/13 18:41:16] <kjetilho> I think your read(1), timeout, waitpid patch is vulnerable -- only waitpid when there's a timeout. but that's just from memory. | ||
| [2010/01/13 18:41:38] <mqr> It wasn't our first attempt. | ||
| [2010/01/13 18:41:45] <kjetilho> nor the last :) | ||
| [2010/01/13 18:41:50] <mqr> Yep. | ||
| [2010/01/13 18:42:19] <kjetilho> I might get time to look into it tomorrow. past midnight here now. | ||
| [2010/01/13 18:42:34] <mqr> I'm hoping to get an automated test rig set up and beef up the tests even further to go at it systematically. | ||
| [2010/01/13 18:42:38] <kjetilho> thank you for your efforts, and good night :) | ||
| [2010/01/13 18:42:47] <mqr> Ditto. Sleep well. | ||
| [2010/01/13 18:44:00] @ Quit: Robbie_: Read error: 60 (Operation timed out) | ||
| [2010/01/13 18:49:24] @ Quit: lak: | ||
| [2010/01/13 18:49:45] @ Quit: jcape: Read error: 110 (Connection timed out) | ||
| [2010/01/13 18:51:08] <himanshu> should I use recurse => true option to create recursive directory? | ||
| [2010/01/13 18:51:27] <himanshu> I want to simpulate mkdir -p option using puppet | ||
| [2010/01/13 18:52:13] <Volcane> there's no such option | ||
| [2010/01/13 18:52:35] <himanshu> recurse -> true can be used for creating recursive directory? | ||
| [2010/01/13 18:52:48] <Volcane> not equiv to mkdir -p | ||
| [2010/01/13 18:52:59] <Volcane> it *copies* recursively | ||
| [2010/01/13 18:53:17] <himanshu> so what should I for mkdir -p | ||
| [2010/01/13 18:53:37] <Volcane> file{["/one", "/one/two"]: ensure => directory} | ||
| [2010/01/13 18:53:52] <himanshu> ohk thanks | ||
| [2010/01/13 18:55:27] <himanshu> will the puppet respect the order this way? | ||
| [2010/01/13 18:55:56] <Volcane> yes, children auto requires parents | ||
| [2010/01/13 18:56:07] <himanshu> thanks | ||
| [2010/01/13 18:56:49] @ crashdummyMCH joined channel #puppet | ||
| [2010/01/13 18:58:44] <crashdummyMCH> hi all, quick question I have a manifest/nodes directory and within that directory I have a bunch of directories named by the node name. within the /manifest/nodes/<nodename> I have the init.pp file. I'm trying to figure out how to import all of these using site.pp | ||
| [2010/01/13 18:59:08] <crashdummyMCH> I've tried import "nodes/**" to no avail | ||
| [2010/01/13 18:59:18] <himanshu> what is the best practice to deploy file which is environment specific | ||
| [2010/01/13 18:59:45] <himanshu> say for production I have verisonX of file and for development environment I have versionY of file | ||
| [2010/01/13 18:59:55] <Volcane> himanshu: source => ["puppet:///path/to/file.${environment}", "puppet:///path/to/file"] | ||
| [2010/01/13 19:00:30] <himanshu> oh ${environment} variable is internal to puppet? | ||
| [2010/01/13 19:00:38] <Volcane> wiki:UsingEnvironments | ||
| [2010/01/13 19:00:43] <Volcane> erm | ||
| [2010/01/13 19:00:45] <himanshu> ohk thanks | ||
| [2010/01/13 19:00:52] <whack> http://reductivelabs.com/trac/puppet/wiki/UsingMultipleEnvironments | ||
| [2010/01/13 19:00:53] <Volcane> wiki:UsingMultipleEnvironments | ||
| [2010/01/13 19:00:54] <gepetto> Volcane: wiki: wiki:UsingMultipleEnvironments is http://reductivelabs.com/trac/puppet/wiki/UsingMultipleEnvironments | ||
| [2010/01/13 19:00:55] <whack> that one, probably | ||
| [2010/01/13 19:04:58] <crashdummyMCH> how do you recursively import all pp files from a directory? | ||
| [2010/01/13 19:06:05] @ Demosthenes joined channel #puppet | ||
| [2010/01/13 19:13:38] @ Quit: crdant: Read error: 110 (Connection timed out) | ||
| [2010/01/13 19:17:20] @ vorlon2261 left channel #puppet () | ||
| [2010/01/13 19:17:22] @ Quit: gebi: Read error: 113 (No route to host) | ||
| [2010/01/13 19:18:37] @ Quit: verwilst_: "Ex-Chat" | ||
| [2010/01/13 19:22:26] @ Quit: bug: | ||
| [2010/01/13 19:24:05] <Bevo> Does anyone know if you can collect say the hostname from multiple hosts in an exported resource and push it all into a single file? | ||
| [2010/01/13 19:25:03] @ Quit: bobbyz: Read error: 60 (Operation timed out) | ||
| [2010/01/13 19:25:41] @ Quit: jab_doa: "Verlassend" | ||
| [2010/01/13 19:25:43] <chadh> Bevo: I think you export "host" resources and collect those, rather than the facts | ||
| [2010/01/13 19:27:36] @ Quit: crashdummyMCH: "Leaving." | ||
| [2010/01/13 19:36:59] <Bevo> chadh, yep this works with: @@host { "$hostname": ip => $ipaddress, tag => "taggedname", target => "/path/to/file" } Host <<| tag == 'taggedname' |>> | ||
| [2010/01/13 19:38:57] @ lak joined channel #puppet | ||
| [2010/01/13 19:39:26] @ charles1 joined channel #puppet | ||
| [2010/01/13 19:40:03] @ charles1 left channel #puppet () | ||
| [2010/01/13 19:40:11] <Bevo> has anyone had problems with puppetqd just stop storing configs quietly? | ||
| [2010/01/13 19:42:40] <ashp> i just rescued a redhat box in the worst way possible | ||
| [2010/01/13 19:42:50] @ GhostOfRickJames joined channel #puppet | ||
| [2010/01/13 19:42:51] <ashp> i had to ftp libelf.so.1 and libgcc_s to it | ||
| [2010/01/13 19:42:55] <ashp> to be able to run rpm again | ||
| [2010/01/13 19:42:59] <ashp> as it completely destroyed itself | ||
| [2010/01/13 19:43:15] <ashp> i'm now force installing all the upgrade rpm's for 5.4 over the broken 5.2 installation | ||
| [2010/01/13 19:43:21] <ashp> if this even boots again i will be shocked and stunned | ||
| [2010/01/13 19:43:29] <Bevo> lol nice | ||
| [2010/01/13 19:44:02] <ashp> something broke and rpm corrupted its databases (as the java garbage it runs ran it out of memory) | ||
| [2010/01/13 19:44:15] <ashp> i think i'm going to recommend we reinstall this as soon as possible :) | ||
| [2010/01/13 19:44:23] <ashp> luckily it should just be a reinstall then drop the java back on top | ||
| [2010/01/13 19:51:52] <chadh> ashp: yeah, I've had some rpm-based boxes run out of disk space in the middle of an update. not cool | ||
| [2010/01/13 19:52:02] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 19:52:15] @ n0ts joined channel #puppet | ||
| [2010/01/13 19:53:00] @ peiriannydd joined channel #puppet | ||
| [2010/01/13 19:53:00] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 19:53:31] @ n0ts joined channel #puppet | ||
| [2010/01/13 19:53:31] @ Quit: n0ts: Client Quit | ||
| [2010/01/13 19:53:52] @ n0ts joined channel #puppet | ||
| [2010/01/13 19:54:05] <ashp> holy shit that worked | ||
| [2010/01/13 19:54:10] <ashp> i am absolutely in awe that worked | ||
| [2010/01/13 19:54:17] <ashp> i did rpm -Uvh * on the yum cache | ||
| [2010/01/13 19:54:21] <ashp> it was like 200 packages | ||
| [2010/01/13 19:54:36] <ashp> after ftp'ing the libraries that rpm needed from a 5.4 box to it afte rit died | ||
| [2010/01/13 19:54:38] <peiriannydd> ashp: That was a great line to drop into the room to | ||
| [2010/01/13 19:55:47] <chadh> ashp: keeping a tarball of the rpm binary and necessary libraries is probably a good idea. I know when I did Gentoo, we use to have a rescue tgz for portage | ||
| [2010/01/13 19:56:34] <ashp> chadh: yeah, after this I should | ||
| [2010/01/13 19:56:52] <ashp> i am going to make my boss agree that i can rebuild this tomorrow because this is a disaster waiting to strike again | ||
| [2010/01/13 19:56:57] <chadh> ashp: the rpm db is probably not accurate, though | ||
| [2010/01/13 19:57:08] <ashp> i've never seen this before: | ||
| [2010/01/13 19:57:08] <ashp> warning: Value of 'preferred_serialization_format' (pson) is invalid for report, using default (b64_zlib_yaml) | ||
| [2010/01/13 19:57:14] <ashp> but this is on my crazy server so.. | ||
| [2010/01/13 19:57:53] @ Quit: giskard: Remote closed the connection | ||
| [2010/01/13 19:58:12] <BarnacleBob> %(*&$#(%* | ||
| [2010/01/13 19:58:20] <BarnacleBob> i upgrade puppet to fix one minor bug i was up against | ||
| [2010/01/13 19:58:26] <BarnacleBob> and i get another more major one in its place | ||
| [2010/01/13 19:58:42] <ashp> that's my favourite kind of upgrade! It's what I just did: a minor upgrade that destroyed the box | ||
| [2010/01/13 19:58:46] <ashp> what did yours break? | ||
| [2010/01/13 19:58:55] <BarnacleBob> cron's wont install any more | ||
| [2010/01/13 19:59:01] <ashp> oh, that's a big one :/ | ||
| [2010/01/13 20:00:20] <peiriannydd> BarnacleBob: What version did you upgrade to? | ||
| [2010/01/13 20:00:25] <BarnacleBob> 25.3 | ||
| [2010/01/13 20:00:54] <peiriannydd> Interesting....I saw some noise about crons on the mailing list, but I thought that was fixed. | ||
| [2010/01/13 20:02:35] <jamesturnbull> BarnacleBob: you getting some error output? | ||
| [2010/01/13 20:02:43] <jamesturnbull> BarnacleBob: how'd you upgrade also btw? | ||
| [2010/01/13 20:03:50] <z00dax> so, is the 0.25.3 client able to work with a 0.24.8 server ? | ||
| [2010/01/13 20:04:38] @ Quit: yarihm: "This computer has gone to sleep" | ||
| [2010/01/13 20:04:49] <peiriannydd> z00dax: No | ||
| [2010/01/13 20:04:57] <peiriannydd> Server upgrade, then client | ||
| [2010/01/13 20:05:06] <peiriannydd> Otherwise interesting things can happen | ||
| [2010/01/13 20:05:44] <z00dax> humm... ok, back to 0.24 on the client then | ||
| [2010/01/13 20:05:49] @ Quit: brd: "leaving" | ||
| [2010/01/13 20:05:52] <z00dax> ta | ||
| [2010/01/13 20:15:31] @ Quit: bindr: "Ex-Chat" | ||
| [2010/01/13 20:18:34] @ Quit: svend: Read error: 110 (Connection timed out) | ||
| [2010/01/13 20:19:13] @ Quit: ezmobius: Read error: 110 (Connection timed out) | ||
| [2010/01/13 20:21:06] @ Quit: clemens_: Read error: 110 (Connection timed out) | ||
| [2010/01/13 20:28:03] <bitmonk> peiriannydd: 'interesting things' nice | ||
| [2010/01/13 20:29:13] <peiriannydd> bitmonk: Well, theoretically nothing will happen, but generally it will just tell you you're using an incompatible version and puke | ||
| [2010/01/13 20:31:11] <bitmonk> sure i just like when things are explained like that.. | ||
| [2010/01/13 20:31:32] <peiriannydd> Glad to brighten your day | ||
| [2010/01/13 20:31:34] <bitmonk> ;d | ||
| [2010/01/13 20:33:24] @ Quit: danielbln: "Leaving." | ||
| [2010/01/13 20:34:27] @ Quit: teyo: | ||
| [2010/01/13 20:37:28] @ crdant joined channel #puppet | ||
| [2010/01/13 20:39:44] @ peiriannydd left channel #puppet () | ||
| [2010/01/13 20:40:30] <ohadlevy> lutter: still around? | ||
| [2010/01/13 20:44:22] <explody> I have teleported 2 billion goats | ||
| [2010/01/13 20:47:11] @ Quit: lpmusic: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 20:49:48] <lutter> ohadlevy: yes .. still | ||
| [2010/01/13 21:01:27] <BarnacleBob> jamesturnbull, sorry was afk. i upgraded by using uupdate on the 25.1 source deb from debian. then doing an apt-get upgrade | ||
| [2010/01/13 21:01:34] <BarnacleBob> jamesturnbull, i do get error output. let me get it for your | ||
| [2010/01/13 21:03:47] @ Quit: Demosthenes: Read error: 110 (Connection timed out) | ||
| [2010/01/13 21:03:48] @ Quit: GhostOfRickJames: "Leaving." | ||
| [2010/01/13 21:04:39] @ bodepd joined channel #puppet | ||
| [2010/01/13 21:05:36] <BarnacleBob> jamesturnbull, the crontab line has one too many * fields: http://pastie.org/777418 | ||
| [2010/01/13 21:05:49] <BarnacleBob> got to run | ||
| [2010/01/13 21:05:58] <BarnacleBob> i'll be asking more about it tomorrow | ||
| [2010/01/13 21:06:03] @ Quit: BarnacleBob: "This computer has gone to sleep" | ||
| [2010/01/13 21:07:10] @ Quit: cwebber: | ||
| [2010/01/13 21:10:01] <ohadlevy> lutter: i have a question about the domain fact - am I correct to assume that you wrote it? | ||
| [2010/01/13 21:10:20] <lutter> ohadlevy: yeah, IIRC | ||
| [2010/01/13 21:10:51] @ GhostOfRickJames joined channel #puppet | ||
| [2010/01/13 21:10:54] <lutter> ohadlevy: there was much gnashing of teeth in what order to try the various tools like dnsdomainname, domainname etc.) | ||
| [2010/01/13 21:11:01] <ohadlevy> lutter: can I ask why do you use the domainname command? | ||
| [2010/01/13 21:11:13] <ohadlevy> I mean, it will return the NIS/LDAP domain, | ||
| [2010/01/13 21:11:51] <ohadlevy> I have a few issues with solaris machines, where I end up writting a simple dnsdomainname script to return the right value | ||
| [2010/01/13 21:12:34] <lutter> heh | ||
| [2010/01/13 21:12:42] <ohadlevy> and in my opinion, the domainname should not be used at all (unless dns domain name doesnt exist at all) but going through the code, it tries dnsdomain then domainname and then resolv.conf - I would use domainname as a last option | ||
| [2010/01/13 21:12:43] @ Quit: crdant: "Leaving." | ||
| [2010/01/13 21:12:52] <lutter> I need to dig up some emails, but it might have just slipped in there with no good reason | ||
| [2010/01/13 21:13:16] <ohadlevy> ok - I just wanted to check with you, I'll be happy to send a patch | ||
| [2010/01/13 21:13:54] <jamesturnbull> ohadlevy: there is a ticket for fixing that fact I think | ||
| [2010/01/13 21:13:58] <ohadlevy> on Solaris, the domainname is usually different the dnsdomain name, this is because solaris use it to resolv some of the maps locations in the ldap tree | ||
| [2010/01/13 21:14:12] @ Quit: nmi: "Leaving" | ||
| [2010/01/13 21:14:14] <jamesturnbull> ohadlevy: so patches towards that == awesome | ||
| [2010/01/13 21:15:48] @ brd joined channel #puppet | ||
| [2010/01/13 21:16:46] <lutter> ohadlevy: if I got it right, the changes were in response to http://projects.reductivelabs.com/issues/show/1654 | ||
| [2010/01/13 21:16:58] <lutter> ohadlevy: and it seems they were for a very similar issue | ||
| [2010/01/13 21:17:29] <ohadlevy> jamesturnbull: there is also #2085 and #2533 | ||
| [2010/01/13 21:17:29] <gepetto> ohadlevy: jamesturnbull: #2085 is http://projects.reductivelabs.com/issues/show/2085 "Facter - Bug #2085: facter determins fqdn in strange and suboptimal ways - ReductiveLabs.com" | ||
| [2010/01/13 21:17:29] <gepetto> ohadlevy: jamesturnbull: #2533 is http://projects.reductivelabs.com/issues/show/2533 "Facter - Feature #2533: Facter's domain.rb resolves DNS domain wrong - ReductiveLabs.com" | ||
| [2010/01/13 21:18:27] <lutter> ohadlevy: looking back, it seems that checking dnsomainname, domainname, and hostname were there from the beginning - what I did was solely to impose a fixed order in which they are checked. Maybe dropping domainname would be the right thing all around | ||
| [2010/01/13 21:18:42] <ohadlevy> so the real question is domain==dns domain | ||
| [2010/01/13 21:18:58] <ohadlevy> maybe simply add nisdomain fact if a value exists from domainname | ||
| [2010/01/13 21:19:04] @ Quit: fluxdude: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 21:19:10] <explody> ohadlevy: not that you would specifically remember this, but I don't suppose this guy ever shared what "sorted out" his SSL problem did he? http://www.mailinglistarchive.com/puppet-users@googlegroups.com/msg00925.html | ||
| [2010/01/13 21:19:14] <jamesturnbull> nasrat: ^^^ ping | ||
| [2010/01/13 21:19:17] <ohadlevy> because at the end of the day, the real problem is that fqdn=name+domain | ||
| [2010/01/13 21:19:24] @ fluxdude joined channel #puppet | ||
| [2010/01/13 21:20:31] <ohadlevy> explody: you have the same problem? | ||
| [2010/01/13 21:20:45] <explody> yeah, just on one AIX box | ||
| [2010/01/13 21:20:47] <lutter> ohadlevy: I would assume that everybody assumes domain == dns domai | ||
| [2010/01/13 21:20:52] @ jbu joined channel #puppet | ||
| [2010/01/13 21:20:58] @ jbu is now known as tjoe | ||
| [2010/01/13 21:21:19] <chadh> lutter: windows domain? yp domain? | ||
| [2010/01/13 21:21:37] <chadh> probably best to disambiguify | ||
| [2010/01/13 21:21:50] <explody> ohadlevy: I suspect this is not a puppet issue at all really, I think it's something to do with the ruby build. | ||
| [2010/01/13 21:21:54] * chadh adds new word to his dictionary | ||
| [2010/01/13 21:22:37] <ohadlevy> lutter: in that case, we can drop the domainname and search resolv.conf and dnsdomainname - maybe add another lookup for the hostname itself (if it contains the short or long name) | ||
| [2010/01/13 21:23:19] <ohadlevy> chadh: pastie your errors :) | ||
| [2010/01/13 21:23:55] <chadh> ohadlevy: not me, unless you want to help me generate new rrds from my reports ;-) | ||
| [2010/01/13 21:24:24] <ohadlevy> chadh: oops :) - btw: new version of foreman generats graphs without rrds | ||
| [2010/01/13 21:24:36] <ohadlevy> explody: pastie please | ||
| [2010/01/13 21:24:41] @ Quit: GhostOfRickJames: "Leaving." | ||
| [2010/01/13 21:24:55] <explody> ohadlevy: the error? | ||
| [2010/01/13 21:25:03] <lutter> ohadlevy: agreed | ||
| [2010/01/13 21:25:13] <chadh> ohadlevy: does it use raw puppet reports as the source? | ||
| [2010/01/13 21:25:19] <ohadlevy> chadh: yes | ||
| [2010/01/13 21:25:26] <chadh> I have 18 months of raw puppet reports that I need to process | ||
| [2010/01/13 21:25:53] <chadh> puppet seems to only do the last day's worth | ||
| [2010/01/13 21:26:04] <ohadlevy> lutter,jamesturnball: I'll send a patch sometime soon :) | ||
| [2010/01/13 21:26:30] @ Quit: jes5: "Leaving." | ||
| [2010/01/13 21:26:36] <ohadlevy> chadh: do you really need 18 months of puppet reports? :) | ||
| [2010/01/13 21:26:38] @ jes5 joined channel #puppet | ||
| [2010/01/13 21:27:01] @ bill joined channel #puppet | ||
| [2010/01/13 21:27:23] <chadh> ohadlevy: probably not, but I have them :) I don't even know how to do the last six months | ||
| [2010/01/13 21:27:37] @ Quit: bill: Client Quit | ||
| [2010/01/13 21:27:59] <chadh> actually, do the rrds limit how far they go back, or do they just keep accumulating? | ||
| [2010/01/13 21:28:07] <chadh> (since they are incremental) | ||
| [2010/01/13 21:28:08] <explody> ohadlevy: http://pastie.org/777436 this is what happens immediately after --clean / generate new req / --sign / re-run puppetd. Server and client are both 0.25.3 | ||
| [2010/01/13 21:28:11] <ohadlevy> chadh: did you look into foreman for puppet reporting? | ||
| [2010/01/13 21:28:33] <chadh> ohadlevy: not yet for that installation. I have a fresh installation where it is going first | ||
| [2010/01/13 21:28:55] <ohadlevy> explody: does the daemon stay in mem? | ||
| [2010/01/13 21:29:28] <explody> ohadlevy: it never even touches the daemon, this happens entirely on the client side | ||
| [2010/01/13 21:30:01] <ohadlevy> explody: I mean if the client side runs as a daemon, someone mentioned that when he resigned his clients, he needed to restart puppetd | ||
| [2010/01/13 21:30:25] <explody> oh oh, no it's not daemonized, they're manual one-off runs | ||
| [2010/01/13 21:30:50] <ohadlevy> and i see that you use a different servername - are you sure that its certificate include that servername? | ||
| [2010/01/13 21:31:04] @ jcape joined channel #puppet | ||
| [2010/01/13 21:31:05] <explody> and I lied, there are packets flowing between client and server, it just doesn't trigger any logging on the server side | ||
| [2010/01/13 21:31:58] <ohadlevy> and, if you look at the log, it does say something like using cache certificate .... if you want to be sure, stop all puppetd and delete ssldir from your client and then try again | ||
| [2010/01/13 21:32:13] <explody> tried that already, same thing | ||
| [2010/01/13 21:32:42] @ Quit: alexine_dsouza: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 21:32:46] <explody> this is only one out of about 6 test boxes on a wide variety of OSes, all the others work | ||
| [2010/01/13 21:32:52] <ohadlevy> chadh: cool, might want to look at the dashboard functionality within foreman, and maybe the summary email would be also intersting | ||
| [2010/01/13 21:33:03] <ohadlevy> explody: clock are in sync? | ||
| [2010/01/13 21:33:08] <explody> yep | ||
| [2010/01/13 21:34:03] @ Quit: lak: | ||
| [2010/01/13 21:34:27] <explody> ohadlevy: I really appreciate the help, but let's just let it go for now, I really suspect the underlying environment on this box; ruby or ssl or something... the only stuff different between it and the boxes that work are the OS ver (6.1 vs 5.x) and the fact it has it's own ruby build | ||
| [2010/01/13 21:34:51] <chadh> ohadlevy: oh yeah, I can't wait. It is driving me nuts that I haven't gotten to it yet. I am currently working through what version of ruby, passenger, etc. to use on the new install | ||
| [2010/01/13 21:36:20] <chadh> ohadlevy: for this cluster install, I think I can take advantage of most of foreman's features, except for the stuff I already get from cobbler. | ||
| [2010/01/13 21:36:43] <ohadlevy> chadh: hopefully foreman can replace cobbler ;) | ||
| [2010/01/13 21:37:13] <chadh> ohadlevy: I am not against that, although it does its just really well. Does foreman manage repos? | ||
| [2010/01/13 21:37:19] <chadh> s/just/job/ | ||
| [2010/01/13 21:37:34] @ teyo joined channel #puppet | ||
| [2010/01/13 21:38:28] <ohadlevy> chadh: partly, it wont sync it for you, but you can specify different repos per os etc, there was not that much of a demand for that, becuase most people just have a simple script to clone a repo from the internet | ||
| [2010/01/13 21:40:15] <chadh> ohadlevy: I was starting from scratch (we use rhn on all our other stuff), and I needed repo management, dns, dhcp, etc., so cobbler fit the bill really well. I haven't really gotten to the kickstart stuff yet, though, and I will try to leverage foreman as much as possible for provisioning/external nodes stuff | ||
| [2010/01/13 21:41:03] <ohadlevy> chadh: great, if you have any feedback I'll be happy to hear about it :) foreman will also handle the puppetca part of things etc | ||
| [2010/01/13 21:41:44] <chadh> ohadlevy: how so? Oh, an interface for signing clients? | ||
| [2010/01/13 21:42:08] <chadh> I just use autosign for now, particularly on the cluster network | ||
| [2010/01/13 21:42:47] <ohadlevy> chadh: once a host has been created in foreman, it would be allowed to request a kickstart (or preseed or a jumpstart), this will add it to the autosign file, and once completed its first puppet run, it would be removed from the autosign automaticilly | ||
| [2010/01/13 21:43:23] <chadh> ohadlevy: are you working on authN/authZ for the interface? For our main network, an interface for some users to be able to do some foreman tasks (add to autosign list, configure puppet classes, etc.) | ||
| [2010/01/13 21:43:30] <chadh> would be great | ||
| [2010/01/13 21:43:54] <ohadlevy> you mean role base authorization ? | ||
| [2010/01/13 21:43:58] <chadh> ohadlevy: yeah | ||
| [2010/01/13 21:44:06] <ohadlevy> yeah, its on the todo list | ||
| [2010/01/13 21:44:16] <ohadlevy> brb | ||
| [2010/01/13 21:46:43] <ohadlevy> chadh: it will have something like who can edit which hosts, or a group of hosts, who can crud partition table layouts etc | ||
| [2010/01/13 21:47:51] <chadh> ohadlevy: that would be great. We "empower" a lot of our researchers to load their own machines | ||
| [2010/01/13 21:49:12] @ GhostOfRickJames joined channel #puppet | ||
| [2010/01/13 21:49:56] @ Quit: WALoeIII: Read error: 113 (No route to host) | ||
| [2010/01/13 21:53:46] @ Quit: teyo: | ||
| [2010/01/13 21:58:07] * ohadlevy is enjoying reading the comments to the puppet vs chef blog | ||
| [2010/01/13 22:00:32] <jamesturnbull> ohadlevy: yeah funny | ||
| [2010/01/13 22:06:26] <chadh> wow, I read them earlier today, but there weren't nearly as many | ||
| [2010/01/13 22:09:19] @ WALoeIII joined channel #puppet | ||
| [2010/01/13 22:09:24] @ cwebber joined channel #puppet | ||
| [2010/01/13 22:11:20] @ Quit: whaley: Remote closed the connection | ||
| [2010/01/13 22:12:11] @ PaulWay[w] joined channel #puppet | ||
| [2010/01/13 22:12:41] <madduck> jamesturnbull: is there anything more you need for #2968 from me? | ||
| [2010/01/13 22:12:42] <gepetto> madduck: jamesturnbull: #2968 is http://projects.reductivelabs.com/issues/show/2968 "Puppet - Bug #2968: Fails to work with ipv6 resolver - ReductiveLabs.com" | ||
| [2010/01/13 22:12:54] <madduck> jamesturnbull: i can also just demonstrate to you next week if you prefer | ||
| [2010/01/13 22:15:04] @ Quit: jes5: "Leaving." | ||
| [2010/01/13 22:17:24] @ Quit: cwebber: | ||
| [2010/01/13 22:24:58] @ pheezy joined channel #puppet | ||
| [2010/01/13 22:26:08] @ cwebber joined channel #puppet | ||
| [2010/01/13 22:27:56] @ Quit: ambroff: Read error: 110 (Connection timed out) | ||
| [2010/01/13 22:28:22] @ Quit: cwebber: Client Quit | ||
| [2010/01/13 22:35:19] <jamesturnbull> gepetto: seen nigelk | ||
| [2010/01/13 22:35:19] <gepetto> jamesturnbull: nigelk was last seen 21 hours, 29 minutes and 22 seconds ago, quitting IRC (Read error: 110 (Connection timed out)) | ||
| [2010/01/13 22:36:40] <bodepd> did you guys read this stuff. apparently puppet is only suitable for running at dentists offices and lawyers, I never realized that they run so much linux infrastructure there :) | ||
| [2010/01/13 22:36:52] @ Quit: pheezy: Remote closed the connection | ||
| [2010/01/13 22:36:56] <jamesturnbull> bodepd: yep read it | ||
| [2010/01/13 22:37:00] <bodepd> lawyers = law offices | ||
| [2010/01/13 22:37:03] <agaffney> bodepd: wtf? | ||
| [2010/01/13 22:37:04] <bodepd> lmao | ||
| [2010/01/13 22:37:36] <bodepd> I am quoting some of the venom spew that was inspired by bitfields blog posting | ||
| [2010/01/13 22:37:55] @ OpenMedia joined channel #puppet | ||
| [2010/01/13 22:38:17] <agaffney> bodepd: URL? | ||
| [2010/01/13 22:38:32] <bodepd> http://bitfieldconsulting.com/puppet-vs-chef | ||
| [2010/01/13 22:38:53] <bodepd> there is a mix of insightfulness, venom and ironic halariousness | ||
| [2010/01/13 22:41:01] <jamesturnbull> bodepd: Teyo's comment btw - really good analysis | ||
| [2010/01/13 22:41:20] <bodepd> yeah, I hear this stuff in the office all the time :) | ||
| [2010/01/13 22:41:37] <chadh> jamesturnbull: yep, I'm saving that now. I have been looking for some good puppet vs other stuff arguments | ||
| [2010/01/13 22:42:08] <bodepd> I have been working on a doc that outlines the advantages of a declarative model. | ||
| [2010/01/13 22:42:18] <chadh> The alternatives around around are declarative (sequence of patches), and Teyo's rebuttal really gets at one of the strengths of puppet | ||
| [2010/01/13 22:42:20] <OpenMedia> How well does puppet deal with version differences between client/svr? | ||
| [2010/01/13 22:42:33] <bodepd> not well enough | ||
| [2010/01/13 22:42:39] <chadh> ack. s/declarative/procedural/ | ||
| [2010/01/13 22:42:43] <bodepd> client can never be ahead of the server | ||
| [2010/01/13 22:43:11] <bodepd> there may be some trouble between .24.8 client and .25.x server (james?) | ||
| [2010/01/13 22:43:43] <OpenMedia> .25.1 svr and .25.3 client = issues | ||
| [2010/01/13 22:44:08] <bodepd> yes, clients cant be ahead of the server. | ||
| [2010/01/13 22:47:34] @ ezmobius joined channel #puppet | ||
| [2010/01/13 22:49:46] @ jason^ joined channel #puppet | ||
| [2010/01/13 22:51:28] <OpenMedia> bodepd: Cheers | ||
| [2010/01/13 22:51:51] <OpenMedia> Dang. Should have cached a local RPM for .25.1 while I was playing around in the testbed. | ||
| [2010/01/13 22:52:28] @ Quit: shadoi: "Leaving." | ||
| [2010/01/13 22:55:53] <jamesturnbull> bodepd: I've got 0.24.8 clients and a 0.25.3 server without issue | ||
| [2010/01/13 22:56:17] <jamesturnbull> bodepd: but the other way is FAIL | ||
| [2010/01/13 22:56:47] <bodepd> yeah, I was thinking that there were issues with either the restful change or the switch to pson. | ||
| [2010/01/13 22:56:51] <bodepd> thanks, I wasnt sure | ||
| [2010/01/13 22:57:14] @ Quit: nDuff: Read error: 110 (Connection timed out) | ||
| [2010/01/13 22:57:45] @ nDuff joined channel #puppet | ||
| [2010/01/13 23:00:36] @ Quit: rmiller4pi8: Read error: 110 (Connection timed out) | ||
| [2010/01/13 23:02:54] @ ambroff joined channel #puppet | ||
| [2010/01/13 23:03:29] @ Quit: ezmobius: Remote closed the connection | ||
| [2010/01/13 23:05:13] @ Quit: bodepd: | ||
| [2010/01/13 23:09:06] @ lak joined channel #puppet | ||
| [2010/01/13 23:15:14] @ nakano is now known as nakano_ | ||
| [2010/01/13 23:17:06] <OpenMedia> is there a sensible way to have a service reload rather than restart when config files update? | ||
| [2010/01/13 23:19:30] @ rmiller4pi8 joined channel #puppet | ||
| [2010/01/13 23:20:03] <Djelibeybi> OpenMedia: set the "restart" command to be /etc/init.d/service reload instead? :) | ||
| [2010/01/13 23:21:20] @ Quit: jason^: Read error: 104 (Connection reset by peer) | ||
| [2010/01/13 23:22:39] @ jason^ joined channel #puppet | ||
| [2010/01/13 23:25:14] @ jes5 joined channel #puppet | ||
| [2010/01/13 23:26:26] @ Quit: lak: | ||
| [2010/01/13 23:27:38] @ Mode +o jamesturnbull by ChanServ | ||
| [2010/01/13 23:27:45] @ jamesturnbull set topic "Puppet 0.25.4rc1 released - http://bit.ly/4EQL1r | Log bugs and feature requests at http://tinyurl.com/46ryk7" | ||
| [2010/01/13 23:27:54] @ Mode -o jamesturnbull by ChanServ | ||
| [2010/01/13 23:29:03] <gepetto> ::trac:: Downloading Puppet edited @ http://reductivelabs.com/trac/puppet/wiki/DownloadingPuppet?version=105 | ||
| [2010/01/13 23:40:33] @ OpenMedia left channel #puppet () | ||
| [2010/01/13 23:44:58] @ Quit: Bass10: Read error: 110 (Connection timed out) | ||
| [2010/01/13 23:57:22] @ jmeeuwen_ joined channel #puppet | ||
| [2010/01/13 23:57:29] @ Quit: jmeeuwen: Read error: 54 (Connection reset by peer) | ||
| [2010/01/13 23:57:36] @ Quit: jcape: | ||
| [2010/01/13 23:57:48] @ rmiller4pi81 joined channel #puppet |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!