[07:19:57] I need some new project: [Talknotify-global] and [python] [09:12:00] Hello saper, you there? [09:18:50] Silke_WMDE: yes [09:21:45] hi [09:23:29] I'm having a problem with a "shallow clone". Do you happen to know how I make it a "deep"/normal clone when I can't modify the first cloning? [09:24:05] I found git fetch --depth=n but it doesn't work. [09:26:51] (The big picture is a puppet manifest giving my a shallow clone because otherwise it would timeout. Now it would be much more useful if I could update that clone but I cant.) [09:29:11] I see [09:29:14] let me check [09:31:58] it's 'operations/puppet'? which depth does it do? [09:32:02] 1 [09:32:26] and yes operations/puppet, the mediawiki.pp has depth=1 [09:32:38] good, I'm just testing my theory :) [09:33:18] We discussed it here: There's some theory, too: https://gerrit.wikimedia.org/r/#/c/47424/ [09:34:02] I was hoping I could manually "deepen" it. [09:34:29] yes you should be able to [09:34:38] at least with some low level [09:36:28] ok let me now try to understand [09:36:39] you are creating a labs instance with mediawiki [09:36:44] yes [09:36:52] just to run a wiki for tests/dev/whatever [09:37:02] yes [09:37:07] by default it comes shallow because it's http and it's quick [09:37:13] yes [09:37:34] and fails if I use depth = full [09:37:48] and you want to switch to ssh and have a full history? [09:40:12] I want to be able to run git pull on mediawiki core. When my colleagues merge code to core, I throw away the whole mediawiki and make puppet install a new one. [09:40:47] Can I use ssh from the Labs instances? Does the ssh agent do such things? [09:40:55] yes [09:41:07] you need to have "AgentForwarding yes" enabled [09:41:12] I do [09:41:31] check "ssh-add -l" on your instance, should have your key [09:41:56] Silke_WMDE: will you be happy if you do something else instead of "git pull" to achieve the same effect? [09:42:09] yes I would be happy with any solution [09:44:31] what was the "git fetch --depth X" command you did try? [09:46:20] I tried different things: git fetch --depth=3, also --depth=10 and git fetch --depth=5 origin master [09:46:40] Error is always: fatal: git fetch-pack: expected shallow list [09:47:20] ah [09:47:54] which URL is puppet using for initial clone (or just check "git remote -v") [09:48:17] origin https://gerrit.wikimedia.org/r/p/mediawiki/core.git (fetch + push) [09:48:23] thx [09:49:29] since with SSH it seems to work4me [09:49:46] even "git pull --depth 3" [09:51:20] I'll give ssh a try [09:51:37] (git remote add ....) and the "git pull --depth N newremote) [09:53:37] funny [09:53:50] foo with accounts... [09:53:51] "git pull --depth 1" works 4 me even on http [09:54:14] Silke_WMDE: I am using ~/.ssh/config to setup username host port for "review" [09:54:29] and then use git remote add review ssh://review/.././ [09:55:08] I need to test using some fake repo and update it [10:14:09] Silke_WMDE: can't reproduce it using the test repo [10:14:25] o_O [10:14:27] Silke_WMDE: I'm adding a change and then "git pull --depth 1" over HTTP and it gets updated [10:14:35] maybe because it's fast forward only [10:14:42] can you try my test repo over http? [10:14:52] yes [10:14:59] git clone --depth 1 http://l.saper.info:8888/test-project [10:17:33] I hope it does not hand on clone (there is a small IPv6 problem :) [10:17:36] hang [10:20:14] (here's how it looked like here after pushing 1 change: http://dpaste.org/SJFGS/) [10:20:54] Silke_WMDE: let me know if the clone works so I can push another change [10:24:13] hm. with your repo I get "Already up-to-date." But with mediawiki I get fatal: git fetch-pack: expected shallow list doing the same. [10:25:42] Ok, never mind. This has already taken us too much time. [10:26:22] good [10:26:24] wait [10:26:25] [bz] (8REOPENED - created by: 2T. Gries, priority: 4Normal - 6critical) [Bug 44129] CORE repo broken ? git pull fails: fatal: pack has 14 unresolved deltas - fatal: index-pack failed - https://bugzilla.wikimedia.org/show_bug.cgi?id=44129 [10:27:07] Silke_WMDE_: can you "git pull --depth 1" now ? [10:27:26] I have push a change now [10:27:34] and make "git log" before pulling [10:27:36] and after [10:27:51] I suppose i have to update puppet? [10:28:30] ah, you mean on your test [10:28:49] yes, there it is working [10:29:47] good [10:30:06] so I need to test on mediawiki clone because of its size and stuff [10:30:29] it is probably that JGit used by gerrit does not support something like this, but not sure [10:35:51] Silke_WMDE_: question: were you doing git pull on master or on some other branch? [10:36:00] master [10:36:18] and it only occurs on mw core, not on extensions. [10:36:48] yes, because of its size and objects in packs... [10:37:41] I am redoing the test on my test gerrit but with mediawiki core repo [10:38:07] thank you so mcuh! [10:38:59] At WMDE there are four birthdays today (not mine) and we are having cake right now. [10:39:10] * Silke_WMDE_ offers saper a piece of virtual cake [10:39:15] :) [10:40:07] hehe [10:40:12] saper: btw I think someone also had the theory that everything works fine when you clone via the commandline [10:40:13] * saper is about to make [[faworki]] [10:40:20] but not when puppet does it [10:40:47] did you have problems with 'git pull' on the command line? [10:41:08] good, got a shallow copy of mediawiki core [10:42:39] ^^Maybe I had problems with git pull because the git clone was not done via the commandline [10:42:47] not sure [10:42:56] is seems to be a pretty stupid wrapper [10:44:30] [bz] (8ASSIGNED - created by: 2Antoine "hashar" Musso, priority: 4Normal - 6normal) [Bug 38996] apache2 need manual start - https://bugzilla.wikimedia.org/show_bug.cgi?id=38996 [10:45:50] so it's a brithday paradox day on WMDE today [10:46:57] [bz] (8ASSIGNED - created by: 2Antoine "hashar" Musso, priority: 4Normal - 6enhancement) [Bug 44041] [OPS] adapt varnish mobile puppet class on beta - https://bugzilla.wikimedia.org/show_bug.cgi?id=44041 [10:48:29] happy birthday was made the sentence of the day [10:49:49] ok pushed lots of changes to mediawiki core [10:50:01] (actually updated it first time since october 2012) [10:51:13] that was a clone you did manually or from puppet? [10:52:45] manually [10:53:06] what is git version there? [10:53:25] 1.7.9.5 [10:53:40] newer that what I have [10:54:47] in the discussion on gerrit about this, andrew said "the same clone will succeed on the command line while timing out in git because a commandline clone is interactive which prompts periodic interaction with the server, hence forestalling the timeout." [10:55:43] but the problem is not the timeout on clone, only unabilty to "git pull --depth X"? [10:56:25] yes. what could be the difference between your testrepo with depth 1 and mw core with depth 1? [10:58:15] size of the repo and packing of objects [10:59:22] but just guessing [11:00:58] or the update goes deeper below the objects already fetched? [11:02:08] !log deployment-prep Added anth1y to the project, he is interested in Lucene / Swift stuff :-D [11:02:10] Logged the message, Master [11:02:35] !log deployment-prep Granted anth1y sudo access on the project so he can plays with Lucene [11:02:37] Logged the message, Master [11:02:45] anth1y: I guess you can now create a Lucene instance and play with it :-] [11:03:01] anth1y: the project lack documentation though [11:03:27] I wrote some overview at https://labsconsole.wikimedia.org/wiki/Deployment/Overview [11:03:28] @hashar sweeet i'll get to it tonight [11:03:43] almost all servers are configured using puppet class [11:03:47] and barely have any live hack [11:03:51] though there are a few exceptions [11:04:02] ok [11:04:07] when I want to add a new system to the beta cluster, I usually create a new instance [11:04:20] I'll have to get more familiar with puppet [11:04:22] then once it has booted and works fine, I apply to hit the puppetmaster:self class [11:04:28] https://labsconsole.wikimedia.org/wiki/Help:Self-hosted_puppetmaster [11:04:31] right [11:04:49] that makes puppet to run out of the content from /var/lib/git/operations/puppet (that is documented in https://labsconsole.wikimedia.org/wiki/Help:Self-hosted_puppetmaster ) [11:04:54] so I can do all my puppet hacks there [11:05:03] and manually run puppet with "puppetd -tv" [11:05:15] when I am happy with the local hack, I submit it in Gerrit for review [11:05:17] then ops review it [11:05:28] ok [11:05:30] once that has been reviewed and merged, I create yet another instance [11:05:33] apply the class [11:05:38] and see if that works fine again :-] [11:05:43] ok [11:06:03] sounds good [11:06:04] (sometime I create several instances in between to make sure the patch submitted for review apply fine on a fresh instance) [11:06:09] so for Lucene, [11:06:22] you can create the instance, try to apply the production puppet class on it [11:06:24] and then see what happens [11:06:28] (tip: it WILL fail) [11:06:34] ok [11:06:38] good to know [11:06:43] haha [11:06:45] form there once puppetmaster:self is applied you can hack the puppet class by editing /var/lib/git/operations/puppet [11:06:48] until it works [11:06:57] or you can hack the classes locally, send the patch in Gerrit [11:07:04] then on the instance fetch the content of the patch [11:07:14] Gerrit provides the commands to fetch a patch [11:07:37] good to know [11:07:59] for exampe [11:08:07] I have adapted the production varnish puppet class for beta [11:08:12] the top change is https://gerrit.wikimedia.org/r/#/c/47567/ [11:08:14] Oh, what?! The pope is going to step down. [11:08:15] Silke_WMDE_: maybe puppet should distribute unpacked mediawiki core and update it as well (running php update.php maybe...) [11:08:16] it is not merged yet [11:08:57] anth1y: so on the test instance (deployment-varnish-t3) I just : (cd /var/lib/git/operations/puppet; GIT_SSH=/var/lib/git/ssh git fetch https://gerrit.wikimedia.org/r/operations/puppet refs/changes/67/47567/2 && git checkout FETCH_HEAD -b 46567/2 ) [11:09:12] anth1y: the long git commands comes directly from Gerrit web interface :-] [11:09:18] anth1y: then I run puppet [11:09:27] AHHH ok [11:09:33] anth1y: if puppet fail, I simply amend the change, fetch the new patchset on the instance, run puppet again [11:09:39] anth1y: until I got something working hehe [11:10:00] saper: what do you mean by "unpacked" mw core= [11:10:01] ? [11:10:19] anth1y: you could do the exact same thing on your own computer though. You could setup a virtual machine fetching its changes out of a clone of operations/puppet on your local computer. That probably make it easier [11:10:21] hashar: right, ok this will take me some time to get my head around all this [11:10:28] anth1y: yeah :-] [11:10:32] anth1y: so play with it a bit [11:10:39] anth1y: and do ask in this channel :-] [11:10:54] hashar: Most def [11:11:14] anth1y: OH, when you create the lucene instance please use 'deployment-' as a hostname prefix. Example: deployment-lucene or deployment-searchbox or whatever :-] [11:11:22] (hmm deployment-whatever actually [11:11:51] anth1y: to test your access, you can connect to deployment-bastion and attempt sudo there [11:11:56] hashar: thanks I was wondering about the instance nomenclature [11:12:14] will do [11:12:25] I usually do something like deployment-- [11:12:32] like: deployment-search-01 [11:12:50] beside the 'deployment-' prefix, nothing is enforced [11:13:01] ok [11:13:35] Silke_WMDE_: ok, got it reproduced, normal git (not puppet), mediawiki core, http [11:14:47] ^^does git pull work or not? [11:15:04] @hashar thanks i''m starting to fade [11:15:26] anth1y: yeah get to bed :-] [11:15:36] anth1y: your brain will process all of the above while you sleep [11:15:42] anth1y: though you might want to save it for later reference [11:16:05] Silke_WMDE_: no, it fails [11:16:14] hashar: I am most def [11:16:26] Silke_WMDE_: bug updated [11:18:01] *think* [11:18:34] so, it's not puppet's fault. [11:28:24] the pack does not contain enough information [11:28:33] so I'd say it's also not http vs ssh issue [11:30:20] I need two new project, Talknotify-global and pymediawiki [11:31:46] please help me [11:32:07] me help please [11:34:11] devunt: I have no idea how one can request a new project sorry :/ [11:35:48] devunt: This maybe? https://labsconsole.wikimedia.org/wiki/Special:FormEdit/New_Project_Request [11:36:56] seems to be it [11:37:14] Silke_WMDE_, i requested already 5 days ago, but nothing created :( [11:37:55] oh. :( [11:45:01] devunt: I recommend to ask on labs-l. A few days ago someone had the same problem - that nothing happened with the request. Those who can help in general answer quickly on labs-l. [11:45:44] ah, ok. thankyou [11:45:55] this was the other mail: http://www.mail-archive.com/labs-l@lists.wikimedia.org/msg00783.html [12:22:28] Silke_WMDE_: interesting, it might be that "git pull" does not request all objects it needs [12:24:03] I still don't understand which software is causing it - gerrit? [12:26:30] it's either git client or JGit on the other side [12:26:40] let me try pure git [12:57:25] saper: not sure, if this is the same issue. http://code.google.com/p/gerrit/issues/detail?id=769 [12:57:45] Silke_WMDE_: no [12:57:50] there is no internal error [12:58:02] ok [12:58:18] Silke_WMDE_: simple work around is to add another remote (for example ssh), fetch objects from it and then merge [12:58:41] :D "simple" hihi! :) [12:58:42] this 769 is fixed actually [13:03:20] Silke_WMDE_: I have 2 or 3 theories why it fails, no time to dig more today [13:04:23] understandably so! thanks for having taken so much time for this! [13:05:07] It helped me very mich to know I don't digg aroung in puppet files any more! [13:05:49] you might actually one little option [13:06:47] git clone --depth X --no-single-branch [13:07:01] (maybe puppet can be told to use it) [13:07:18] ok, I'll try that one [13:08:54] but I'm guessing this one [13:08:59] didn't check [13:09:50] ouch sorry no workaround :( [13:32:21] Silke_WMDE_: ... found two JGit bugs, linked to https://bugzilla.wikimedia.org/show_bug.cgi?id=44129 Looks like JGit does not support that fully (but one of them is fixed, maybe just needs to be pulled in) [13:33:24] ok. [13:34:37] thanks for digging! [13:37:22] <^demon> saper: The fixed one is already in, which is why cloning will work. [13:37:31] <^demon> The unfixed one is probably what's causing this. [13:37:35] clone was another one [13:37:50] <^demon> Oh, hrm. [13:38:06] * saper checks again [13:39:20] this was clone I think https://bugs.eclipse.org/bugs/show_bug.cgi?id=301627 [13:40:53] spearce filed 301627 for clone nd 301628 for the rest afair [13:54:49] hah wm-bot is online 46 days... that's more than half of freenode servers I guess :D [14:57:11] !log wikidata-dev wikidata-dev-9: cronjob had stopped to pull core from git because of modified files. Stashed them and pulled manually. [14:57:13] Logged the message, Master [15:03:51] [bz] (8ASSIGNED - created by: 2Antoine "hashar" Musso, priority: 4Normal - 6normal) [Bug 38996] [OPS] apache2 need manual start - https://bugzilla.wikimedia.org/show_bug.cgi?id=38996 [15:15:54] [bz] (8NEW - created by: 2Nemo, priority: 4Low - 6trivial) [Bug 44149] Remove deckle.co.za - https://bugzilla.wikimedia.org/show_bug.cgi?id=44149 [15:44:24] !log integration applying misc::package-builder to integration-jobbuilder [15:44:24] Logged the message, Master [17:14:49] [bz] (8RESOLVED - created by: 2Željko Filipin, priority: 4High - 6major) [Bug 44565] en.wikipedia.beta.wmflabs.org takes 10-20 seconds to load - https://bugzilla.wikimedia.org/show_bug.cgi?id=44565 [18:28:36] Ryan_Lane, are you here? [18:28:47] Platonides: yep [18:28:50] who is handling the tools lab position? [18:29:23] hi sumanah [18:29:29] the managers, I believe [18:29:30] for now [18:29:41] hey there Platonides [18:30:01] "the managers"? :s [18:30:17] !logs [18:30:17] logs http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-labs [18:30:23] ct and folk [18:30:38] oh [18:30:40] Platonides: when you say "handling" what do you mean? [18:30:52] hiring for it? [18:30:57] or who will work with that person? [18:31:14] sumanah, hiring [18:31:27] the people that will be soon receiving a bad report about me :P [18:31:41] Platonides: ? why? [18:31:52] Platonides: CT Woo is the manager who is working on hiring [18:32:17] I have just been talking by phone with Emily [18:32:23] Platonides: and Emily Blanchard and Heather McAndrew are the recruiters who are working on hiring for that position [18:32:28] but I wasn't understanding her well, so it was very frustrating [18:32:51] Platonides: I highly suggest then that you email a followup explaining that [18:33:03] Platonides: what was the barrier? accent? audio quality? [18:33:15] well, she knew about that [18:33:27] Even so. [18:34:18] I need to go to a meeting. Still. Trust me. I am one of WMF's hiring managers. If you are a candidate and you had trouble talking with one of the recruiters, *email and suggest a followup conversation via text chat* [18:35:01] I mentioned irc after making her repeat a question several times [18:35:13] but she didn't want (was not allowed?) [18:35:18] She's allowed. [18:35:19] I'm not sure what was the reason [18:35:33] she had to talk with the people or something [18:35:40] I have to go, sorry [18:43:22] Platonides: we've discussed this. I think CT will be talking with you [19:18:12] Ryan_Lane, it's up to them [19:18:44] I simply came out frustrated [19:21:12] Platonides: what was the issue? audio quality? accent? [19:21:40] would it have been easier if it was a hangout, so that you could see the other person? [19:23:29] I'm not sure which was the exact issue [19:24:11] there was some echo, and I suspect she was getting too quick when doing a sentence [19:24:35] ah. I bet she was on speakerphone in a conference room [19:24:47] so I then lost a few keywords and was confused [19:25:04] some of our conference rooms can only do speakerphone and they have a lot of echo [19:25:12] I asked her to repeat several times [19:25:22] she was very patient, but we didn't really advance the conversation [19:25:28] * Ryan_Lane nods [19:25:49] it's also possible that I didn't know some of the words when spoken [19:25:55] I'm not a native speaker, after all [19:26:16] right [20:44:12] Quick question if anyone's around [20:44:38] So Puppet is handling an instance's MySQL config; how do I ask puppet to make mySQL listen for connections on * instead of 127.0.0.1? [20:47:20] FastLizard4: I guess that is in the /etc/my.cnf [20:47:27] there might not be a class for that though [20:47:29] hashar: I tried that [20:47:39] But every few minutes, Puppet will come by and revert my changes [20:48:45] yeah that is how puppet works [20:48:58] it ensure the file on your server match what is defined in the manifest [20:49:48] FastLizard4: which of our puppet classes are you using? [20:50:07] generic::mysql::server supports the "bind_address" parameter [20:50:13] which default to "127.0.0.1" [20:50:32] and is then expand in /etc/mysql/my.cnf for the bind_address mysql configuration statement [20:50:36] hashar: base, ldap::client::wmf-test-cluster, exim::simple-mail-sender, sudo::labs_project, role::labs-mysql-server [20:50:41] ahh [20:50:43] the role [20:51:07] which indeed does not map bind_address :-] [20:51:18] do you have a copy of operations/puppet.git ? [20:52:04] role::labs-mysql-server provides some sane default but ultimately invoke generic::mysql::server [20:52:26] hashar: I don't think I do [20:52:27] if you want to further tweak the mysql installation, I think you should apply generic::mysql::server [20:52:37] and add the class parameters [20:53:06] though I am not sure how one can pass parameters to a class using labsconsole :( [20:53:14] hashar: How would I go about doing that? I don't see generic::mysql::server in labsconsole [20:53:44] you will have to add the puppet class to your project using https://labsconsole.wikimedia.org/wiki/Special:NovaPuppetGroup [20:54:05] but then I fail to see how to pass parameters to ti [20:54:11] so that is not working well [20:55:56] hashar: I think I have an idea [20:59:13] got a conf call, will be back in half an hour [20:59:22] maybe andrewbogott can help you too [21:00:05] (andrewbogott FastLizard4 tries to override a mysql setting defined by role::labs-mysql-server , it calls generic::mysql::server with some hardcoded setting, not sure how one can pass parameters to a class using labsconsole ) [21:00:55] FastLizard, let me catch up on the backscroll. [21:01:03] andrewbogott: okay :) [21:07:35] FastLizard4: Do you relish the opportunity to do some puppet development, or would you like me to change it for you? [21:07:55] andrewbogott: I can't say I relish it, but I don't mind learning :) [21:14:39] FastLizard4: Ok, sorry, multiply distracted here. Do you have/have you ever set up a puppetmaster:self instance? [21:14:50] andrewbogott: Nope :P [21:17:03] FastLizard4: But you have written > 0 lines of code in your life? [21:17:12] andrewbogott: Yes [21:17:12] :P [21:18:33] FastLizard4: OK, so to avoid messing up the system that you actually care about, probably best for you to spin up a new instance just to make this one change. There's a guide about making a puppet-development system here: https://labsconsole.wikimedia.org/wiki/Help:Self-hosted_puppetmaster [21:18:42] And I'm happy to explain any of the subdetails there if it's unclear. [21:18:54] Okay [21:18:56] * FastLizard4 opens link [21:19:01] I think once you have a testing system running making the actual code change will be close to obvious. [21:55:16] andrewbogott: From what I understand of the documentation you sent me, since I found "bind_address = <%= bind_address %> [21:55:47] andrewbogott: ...in templates/generic_my.cnf.erb, doesn't that mean I can just define bind_address in labsconsole? [21:56:55] FastLizard4: Tragically not, because that variable is explicitly set in the class that uses the template. [21:57:07] andrewbogott: Ahh, I see [21:57:19] I'm pretty sure that it's passed as a parameter all the way up and that we just need to add the last bit, in the role. [21:57:35] Which we can do either by opening it up to be definied in labsconsole or creating a new special-purpose role that defines it the way you want [21:57:46] Depending on whether you think it's generally useful to have it be user-defined or not. [21:57:56] Well, I'd say it'd be useful to have it user-defined [21:58:12] In this case, our setup is that we have two instances, one serving the web app, the other serving as the database backend [21:58:28] With the default Puppet configuration, this wouldn't work of course, since the database server would only listen to itself [21:58:35] Sure, ok. That's not super hard. You'll want ot put in a scrap of code to make sure it defaults to 127.0.0.1 if the user doesn't specify. [21:58:43] Hmm [21:58:46] Lemme switch over to Git on my computer :P [21:59:08] Since it looks like I might just end up submitting this to Gerrit after all :P [21:59:41] Yeah… probably will be six or eight lines in one file; unless it turns out that the parameter needs to pass through another layer [22:03:34] when a bot dir is created (on data/bla...), do i have to change the dir owner? [22:04:38] probably [22:04:49] most important question - noodles or rice for tea [22:13:32] Damianz: i'm not familiar with this dir system, i don't want to *breack the glass* :P [22:14:06] well define bot dir and I assume you're talking about /data/project/ [22:15:26] yes. i've created and changed the owner [22:16:39] bots should run as a system account so yeah, change the owner :) [22:16:58] * Damianz goes to murder noodles [22:29:22] andrewbogott: Okay, I think I got it, but I want to test it before I submit a patchset to Gerrit; how would I go about doing that? [22:29:53] FastLizard4: The virtue of using puppetmaster::self is that you now have a local puppet repo that puppet will use for puppet runs. [22:30:02] andrewbogott: Ahh, okay [22:30:09] So, copy your change into the /etc/puppet repo [22:30:25] And then select the class you're testing (role:labs:mysql or whatever) in the web console for that VM [22:30:35] And 'sudo puppetd -tv' and see what happens [22:32:12] brb [22:33:16] andrewbogott: How goes the salting? [22:35:08] andrewbogott: thanks to have stepped in :-] [22:38:09] andrewbogott, you can create projects, right? [22:38:59] o.O There is no hg? :s is it possible to install, or we must stick to git? [22:39:14] Use hg if you want [22:39:25] Just never, ever use cvs or svn ;) [22:39:57] lool, i update pywp by svn :P [22:40:06] Which instance do you need it on? [22:40:23] i'm on bots-4 [22:40:43] if there's hg on other instance, i can move to there [22:42:44] Just installed it on 4, no biggie [22:43:16] We should make svn, git, hg standard... hmm *notes to add to template* [22:43:44] Ryan_Lane: Any time sometime this decade to go over bots stuff re puppetizing or should I annoy the fuck out the new guy when you hire him? [22:43:56] heh [22:44:02] well, I'm doing gerrit today [22:44:05] so maybe not today [22:44:07] but this week [22:44:12] maybe wednesday? [22:44:21] <^demon> Damianz: git is part of the base packages, all servers should have it now. [22:44:58] sounds good - I'll mentally note to leave work on time and idle around here :D [22:45:01] ^demon: :D [22:45:04] <3 git [22:45:06] svn should die [22:45:20] Though we use cvs at work still *shudder*... moving to git slowly [22:45:21] <^demon> You're telling me :p [22:51:11] Damianz: Salting is going haltingly but I'm not blocked by anything in particular [22:51:37] :) [22:51:44] not that I'm really much of a help anyway lololol [22:51:58] Damianz: I don't know that you should regard it as a canonical example, but labs-morebots is puppetized and I don't hate how it's working. [22:52:20] JasonDC, what's up? [22:52:48] http://lists.wikimedia.org/pipermail/labs-l/2013-January/000751.html [22:53:00] The thing that worries me is ideally each bot should be a module - re-usable outside labs, but due to puppet duplicate class issues etc that's impossible so we're allmost going against the goals by doing it [22:53:04] =\ [22:54:43] Damianz: Yes, true. For the 'real' solution we should probably wait for person-to-be-named-later to implement a grid solution. I don't know how that will relate to puppet though [22:55:42] Puppet would still configure the servers, the 'grid' would just be like PaaS for bots, which can be open... how requirements for bots etc tie in is the complex bit [22:56:10] * Damianz is seriously considering using salt states at work because of this exact issue when dealing with cross teams modules [22:56:49] JasonDC, sorry for delay, I'll create that now. You want the project name to just be 'Opensim'? [22:56:58] Ryan_Lane would stab me but I'd just make everyone use a virtual env :D [22:56:58] that or Wiki3d [22:57:13] actually, opensim cause it will prob be renamed at some later point [22:57:26]