15:03:00 #startmeeting 15:03:00 Meeting started Mon Mar 25 15:03:00 2013 UTC. The chair is ewoud. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:03:00 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:03:08 * dcaro is here 15:03:08 #chair Rydekull dneary 15:03:08 Current chairs: Rydekull dneary ewoud 15:03:13 #chair dcaro 15:03:13 Current chairs: Rydekull dcaro dneary ewoud 15:04:03 The agenda from http://www.ovirt.org/Infrastructure_team_meetings#2013-03-25 15:04:09 Introductions 15:04:09 Hosting 15:04:09 rackspace01 installed? 15:04:09 Puppet 15:04:09 Summarize the ML thread 15:04:12 Jenkins 15:04:14 Other business? 15:04:17 Trac review 15:04:56 I think we should also discuss eedri's progress on alterway02, but he's not online 15:05:10 anything else? 15:06:24 it seems not 15:06:27 #topic introductions 15:06:52 anyone who would like to introduce themselves? 15:07:10 * Rydekull doesnt think so 15:07:30 #topic Hosting 15:07:41 relaying info from quaid (who's on a plane right now) 15:08:03 he's having issues while installing fedora 17 and 18 15:08:39 anaconda crashes and he doesn't know if it's the iDRAC, fedora or his browser/java but working on it with rackspace support 15:09:07 Rydekull suggested to use a kickstart instead of manually installing 15:09:12 so work in progress 15:09:33 eedri mailed he installed ovirt on alterway02 15:09:54 http://lists.ovirt.org/pipermail/infra/2013-March/002486.html 15:10:17 we should decide on a host naming scheme so please respond to it 15:11:04 feels like a bit of a monologue here 15:11:13 Hehe, im here, all the time 15:11:19 Just got nothing to comment :-) 15:11:26 I can comment that, if You want 15:11:36 that reminds me, next week is easter so we should discuss that as well 15:12:01 anything else to add on the hosting front? 15:12:22 not as far as I know, i've been ill during last week 15:12:38 been a bit busy myself as well 15:12:55 right, moving on to the next point on the agenda 15:12:58 #topic puppet 15:13:15 I haven't seen discussion on the ML thread so little to summarize 15:13:25 +1 15:13:46 also been too busy to work on my proof of concept puppet repo, but I hope to work on it later today 15:14:01 dcaro: anything to add? 15:15:36 moving on 15:15:39 #topic jenkins 15:15:55 I haven't heard problems about the new jenkins master so I'm assuming it works well 15:16:09 ewoud: no, not yet 15:16:15 dcaro: ok 15:16:43 I feel... we need to keep track and organize access to Jenkins Slaves 15:16:50 document them in some way 15:16:57 ewoud: I saw a strange error with a specific patch of vdsm, it was not reproduced but made the job fail 15:17:16 Rydekull: that's a good point 15:17:42 dcaro: is that an issue of a specific slave? all slaves? master? 15:17:44 ewoud: you fixed /releases/3.2/rpm/EL/6/? 15:17:48 outi: yes 15:17:56 the same fix is needed for /releases/beta/rpm/EL/6/ 15:17:59 ewoud: it happened on both f18 slaves 15:18:09 outi: oh wait, maybe I only fixed the beta 15:18:18 dcaro: ok 15:18:51 outi: it appears the beta is a symlink to 3.2 so both should be fixed 15:19:07 Rydekull: I think that the goal should be to have it in puppet 15:19:08 outi: try a yum clean if it doesn't get picked up 15:19:24 dcaro: not sure how that's related 15:19:37 dcaro: but what about one time access to debug a build 15:19:46 dcaro: As with any machine, or service under infrastructure teams control, we should have it documented 15:19:59 ewoud: oh, yes. sorry. browser refresh fixed it ;) 15:20:00 we've seen that happen as well with for example selinux, do we handle that in puppet as well? 15:20:38 dcaro: you were talking about my magic patch? 15:20:48 apuimedo: yep :) 15:20:52 :P 15:21:02 Rydekull: I'm a proponent of code-as-documentation so my preference would be to handle it in puppet (unless it's procedures) 15:21:02 I still don't know what could it be 15:21:24 Rydekull: ewoud: exactly 15:21:52 You all speak like developers :-) 15:21:55 :-P 15:21:59 Rydekull: I am a developer 15:22:01 I know 15:22:03 Im not 15:22:23 I feel like we should have a sprint some time to actually work this stuff, but not sure how to organize it 15:22:36 14[[07Intel Workshop May 201314]]4 10 02http://www.ovirt.org/index.php?diff=8107&oldid=7879&rcid=8309 5* 03Theron 5* (+146) 10/* Accommodations */  15:23:18 maybe we should start by documenting on the wiki what we have and then move that to puppet code 15:23:36 agree 15:23:48 I'm all for automating stuff. But we need to know what to automate 15:24:13 #agreed we need to start documenting stuff on the wiki before we can automate it with puppet 15:24:22 ewoud: about selinux, puppet can manage it too... but never used it so I don't know how easy is to be used 15:24:24 I think that's the conclusion then 15:25:03 dcaro: I was refering to a specific example where dan walsh was granted access because a build was failing due to selinux and it needed debugging because it wasn't reproducable 15:25:09 not sure about the details 15:26:08 Rydekull: any good idea how to get started on documentation? 15:26:23 Who knows the Jenkin Slaves? I know nothing about them 15:26:30 I'm thinking about making a list of machines we have and a list of users that have access 15:27:52 Rydekull: all I know is http://jenkins.ovirt.org/computer/ 15:28:26 looking at the clock difference I suspect the master may not be in sync 15:29:03 I'll get started on making that list 15:29:03 I helped migrating them to the new master 15:29:18 then you can add to it 15:29:32 We ought to have a system for how we accept a new slave, under which conditions, no? 15:29:34 #action ewoud start a list of machines infra has 15:30:21 Rydekull: are you thinking of letting anyone 'donate' new slaves to the master? 15:30:36 ewoud: well, im already tasked with creating a list of machines and services, so I can finish the backup-ticket in the future :-) 15:30:52 Rydekull: ok 15:30:55 #undo 15:30:55 Removing item from minutes: 15:31:08 dcaro: Well, no, im thinking about publicly definining who we allow slaves and under what conditions 15:31:15 Rydekull: let me know if you need any help 15:31:19 dcaro: what we require in return and what information we want 15:31:23 ewoud: sir yes, sir 15:31:52 Rydekull: as someone who donates a slave I can confirm there aren't any real guidelines on it 15:32:35 Maybe we can create a vm template or something similar so it's easy to donate slaves 15:32:44 ewoud: exactly :-) 15:32:52 dcaro: that might be one option 15:33:04 dcaro: I think he's more refering to the legal requirements around it 15:33:15 Everything around it 15:33:45 on a technical level I'd prefer if any donated machine runs puppet so we can ensure it's properly installed 15:33:49 We need to know, to be able to guide other. and by publishing our guidelines, others have easier to adopt 15:34:39 Rydekull: something like https://ring.nlnog.net/ is what you're thinking of? 15:35:48 haha, funny, yes, that is a excellent example 15:35:49 (not saying we need to copy that list, but the idea of a simple list of requirements) 15:37:21 they even publish their puppet classes on https://github.com/NLNOG/ring-puppet 15:39:10 I think we can move on to the next point on the agenda 15:39:24 * quaid arrives at the airport 15:39:37 * ewoud should learn to use #info more 15:39:47 I realize the minutes will be quite small now 15:39:58 We live and we learn 15:40:26 * quaid reading the buffer 15:41:11 #action Rydekull add the jenkins slaves to a list ont he wiki 15:41:19 going to go back to hosting for the minutes 15:41:21 #topic hosting 15:42:10 #info quaid is having issues installing fedora on rackspace; may be the iDRAC, fedora or his browser/java but is working on it with rackspace support 15:42:53 #info eedri installed ovirt on alterway02; we now need to start installing VMs and think of a proper naming scheme 15:43:14 #topic other business 15:43:47 as I said earlier, next week is easter so I'm wondering how many are available next monday 15:44:35 * dcaro won't be available 15:44:54 I will most likely be available 15:45:07 I'm not sure yet but most likely I'm available as well 15:45:31 quaid: dneary ? 15:45:43 ewoud, ? 15:45:55 Next Monday is a holiday in France 15:46:00 I will be offline 15:46:03 that's why I'm asking 15:46:33 I could take next Monday off, sure 15:47:17 #idea organizing a sprint - put up a whenisgood calendar, look for crossover spots where a few people can be together at a time, with bleed over to the next block of people 15:47:59 so what do we do next week? 15:48:00 I also could attend 15:48:06 our holiday is Friday this week 15:48:17 (I think) 15:48:21 quaid: do you have any experience with setting up such a sprint? 15:48:36 I do actually 15:48:44 let's give it to me :) 15:48:57 * quaid used to do docs for Fedora etc. 15:49:04 #action quaid to setup docs sprint 15:50:58 do we have more or shall we go to the trac review? 15:51:09 no more from me thx 15:51:25 for Trac review, I stopped mid-way last week, we could start from that point ? 15:51:34 #topic trac review 15:51:36 let's 15:52:23 * quaid looking in minutes for where that was 15:53:05 38 was the last one 15:53:22 https://fedorahosted.org/ovirt/report/1 15:53:23 in that list 15:53:41 oops, closed that one 15:54:03 21 was last before that 15:55:24 btw, I also noticed http://releases.ovirt.org still shows the old wordpress site without CSS so I think we should fix that as well 15:55:52 holy moley 15:56:08 sounds like a new ticket :) want to file it? 15:56:31 euh, resources.ovirt.org that is 15:56:31 * quaid makes one for upgrading the theme that might solve other tickets 15:57:09 * quaid did the mental transpose, how funny 15:59:49 * ewoud created https://fedorahosted.org/ovirt/ticket/40 15:59:55 I'll take a look 16:01:09 https://fedorahosted.org/ovirt/ticket/39 16:02:41 ticket https://fedorahosted.org/ovirt/ticket/10 - Rename arch@ mailing list to devel@ 16:02:51 quaid: any idea about a migration plan? 16:03:58 I'm assuming the mediawiki issues depend on #39 now 16:04:18 right most likely 16:04:20 14[[07Features/Otopi Infra Migration14]]4 !10 02http://www.ovirt.org/index.php?diff=8108&oldid=8090&rcid=8310 5* 03SandroBonazzola 5* (+39) 10/* engine-setup */  16:04:22 well, it's openshift 16:04:32 https://fedorahosted.org/ovirt/ticket/29 Add new admins to OpenShift 16:05:24 #idea migration by just trying out the upgrade during a window, then rollback if it doesn't work & make a more stepped plan 16:06:00 I have little experience with mailman so I'm of little use 16:06:36 oh, that migration 16:06:46 this was about mediawiki 16:07:07 quaid: could you give me access to openshift so I can help if there are issues? 16:07:24 quaid: and upgrading the theme during a window sounds fine 16:07:44 sure 16:07:51 do you want to use the same sshkey as you have otherwise 16:08:42 yes 16:08:54 you may be added already checking now 16:09:46 in that case I have to look for the login 16:11:08 I'm getting both, you need the long string to login yep 16:11:25 * quaid will paste here for the log in ... one ... CSS .. moment 16:11:38 I found 847edb45aea84198838f915be6faa066@wiki-ovirt.rhcloud.com in the logs 16:11:51 ssh://847edb45aea84198838f915be6faa066@wiki-ovirt.rhcloud.com/~/git/wiki.git/ 16:12:12 ok, none for you yet, let me add 16:12:46 ok, you should be able to checkout now ... 16:13:14 correct 16:15:34 quaid: think we can close #29 now and open new tickets when we need to add more? 16:16:54 ok 16:17:02 * quaid heading to gate now 16:17:06 ttyl 16:17:06 quaid: ok 16:17:15 then I think we can end the meeting now 16:17:23 thanks all 16:19:00 thanks! 16:19:11 #endmeeting