[15:17:09] ok, so qchris [15:17:17] puppet does not do recursive directory permissions [15:17:53] 1sec... let me open the puppet file [15:18:47] hmm [15:18:57] i think if you want the dir contents to be group owned by a different user [15:18:58] so it manifests/misc/statistics.pp:936 [15:19:00] you'll have to run an exec [15:19:06] ja [15:19:12] There is: group => 'www-data', [15:19:25] and in git::clone usage doc: [15:19:25] # $+group+:: Group owner of $directory, default: 'root' [15:19:41] annnnd [15:19:52] and the dir is owned by stats or root [15:19:55] deponding on the host [15:20:06] "dir is owned" should read [15:20:07] oh just the main dir, not the contents? [15:20:16] "group owner is" [15:20:21] yeah hm, the git clone just runs the exec with the group [15:20:22] hm [15:20:33] yes, the dirs inderneath /a/geowiki [15:20:39] (just the first level) [15:20:42] dunno what exec group really does [15:20:43] group [15:20:43] The group to run the command as. This seems to work quite haphazardly on different platforms – it is a platform issue not a Ruby or Puppet one, since the same variety exists when running commands as different users in the shell. [15:22:01] Ah. ... so I cannot rely on "group" setting the group ownership ... Neither for "file", nor for "git::clone"? [15:22:14] Ok. [15:22:30] I'll try to find some other way then. [15:22:34] yeah this git clone doesn't work rigiht [15:22:34] hmmm [15:22:36] i'm fixing [15:25:10] hmmm [15:27:42] Ahhhh, i dunno qchris, hm. [15:27:51] Ok. [15:27:53] i was going to change git::clone so that the $directory permissions were set properly [15:28:13] but this might break things where others use file { $directory: … } in there manifests somewhere [15:28:16] duplicate file defintions [15:28:17] sooooo [15:28:18] hm [15:28:28] you can do it yourself after the git clone [15:28:40] Ok. [15:28:44] file { $blablapath: [15:28:44] ensure => 'directory', [15:28:44] require => Exec["git_clone_$blabla"], [15:28:44] owner => $owner, [15:28:44] group => $group, [15:28:45] mode => $mode, [15:28:45] } [15:28:47] or sorry [15:29:04] file { $blablapath: [15:29:04] ensure => 'directory', [15:29:04] require => Git::Clone['blabla'], [15:29:04] owner => $owner, [15:29:04] group => $group, [15:29:04] mode => $mode, [15:29:05] } [15:29:15] Ok. I'll try that. [15:29:18] Thanks. [15:44:39] ottomata / qchris: paravoid says to use git deploy / sartoris [15:44:43] this makes sense [15:44:50] ja [15:44:52] cool [15:44:52] because he wants everyone to be trying the same standard system [15:44:56] for the two repos? [15:45:01] two? [15:45:11] a frozen dep repo? [15:45:39] uh, I'm not sure how sartoris works yet [15:45:42] i've gotta read up [15:46:02] ja [15:47:24] ok, ottomata, but the very least we'd need is a place to deploy it [15:47:27] so is that stat1001? [15:47:47] and the client we're deploying from would be tin? [15:48:15] yup and yup [16:30:04] hi [16:32:34] Good morning :) [16:34:14] Hi! [16:34:42] hey qchris , halfak [16:34:50] hey milimetric [16:34:59] :) [16:35:07] morning [17:11:12] :) [18:31:48] qchris_away: geowiki private perms look better now I think [18:47:53] ottomata: Thanks! [21:03:42] milimetric: can you hear us? [21:03:45] (hangout) [23:01:48] (PS11) Milimetric: [ready for review] Async cohort upload [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/93081 [23:02:13] (PS12) Milimetric: [ready for review] Async cohort upload [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/93081