18:00:09 #startmeeting Infrastructure (2014-03-27) 18:00:09 Meeting started Thu Mar 27 18:00:09 2014 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:09 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:09 #meetingname infrastructure 18:00:09 The meeting name has been set to 'infrastructure' 18:00:09 #topic welcome y'all 18:00:09 #chair smooge relrod nirik abadger1999 lmacken dgilmore mdomsch threebean pingou puiterwijk 18:00:09 Current chairs: abadger1999 dgilmore lmacken mdomsch nirik pingou puiterwijk relrod smooge threebean 18:00:16 here 18:00:20 who all is around for a exciting infra meeting? 18:00:25 * danofsatx-work is nuking lunch 18:00:26 * tflink is around 18:00:31 * lmacken 18:00:33 * docent also 18:00:34 * pingou 18:00:40 * adimania too 18:00:42 * relrod here 18:00:52 here 18:00:56 * abadger1999 in other meeting 18:01:32 #topic New folks introductions and Apprentice tasks 18:01:42 * randomuser is around 18:01:44 any new folks like to introduce themselves today? 18:01:47 * roshi is around 18:01:50 or apprentices with questions or comments? 18:02:23 I poked around with some fedorahosted stuff over the weekend, please yell at me if you find anything broken 18:02:49 EOF 18:03:08 randomuser: cool. I see there's one new hosted project add pending... if you want to do it. ;) 18:03:35 Instead of doing apprentice work I'm poking around Docker for the last month (Fedora Docekrfiles; together with Scott Collier) 18:03:58 docent: cool. docker is pretty nifty. 18:04:17 I was at the gitlab-rpm meeting last week and there is possibility that we'll build some gitlab containers in orde rto test some gitlab spins within fedora 18:04:29 nice. 18:04:40 ok, that's it about me :) 18:05:21 #topic Applications status / discussion 18:05:30 any applications news this week or upcoming? 18:05:40 pkgdb2 is getting slowly closer 18:05:51 abadger1999: worked on the cron job to sync ACL pkgdb2 <-> git 18:06:05 there is still the bugzilla and postfix ones to do 18:06:22 ok 18:06:24 a new bodhi release went out this week to fix el5 repodata, should hopefully fix bug #1080373 after the push that is happening right now finishes. 18:06:31 #info pkgdb2 moving forward 18:06:46 #info new bodhi release to fix el5 repos, should be fixed later today 18:08:46 I'm going to look at making a sundries01/02/01.stg set... this would be to move: geoip, freemedia, mirrormanager-admin, and possibly busmon to. 18:08:56 ie, move those off the app servers. 18:09:12 since they are all small services that don't really need their own instances. 18:09:13 what about the cron jobs? 18:09:14 * relrod is working on the jenkins <-> fedmsg plugin again today. I hope to have it done by this weekend sometime though school has been giving me a lot of work these past few weeks. :( 18:09:39 relrod: school first, but cool news :) 18:10:01 * threebean wanders in late 18:10:01 pingou: good question. We could do cron on those too... unless the crons need access to shared storage or anything? 18:10:16 should identify all the crons we have on bapp02 now. 18:10:19 nirik: not for the ones I know 18:10:28 but I don't know them all :) 18:10:52 we will still need a mirrormanager-backend01 for replacing some bapp02 services.... (s3 uploads, crawling mirrors, updating mirrorlists, etc) 18:11:12 and I think there's some pkgdb1 stuff running there? 18:11:29 just some crons perhaps 18:11:56 * mirek-hm is late, but here 18:12:26 oh, another thing I wanted to bring up app wise... 18:12:49 have we started looking at what we need to do for rhel7 migrations? I assume packging would need to be worked for our apps... 18:13:07 * threebean nods 18:13:17 yeah, I noticed there was now a 7 branch in EPEL 18:13:20 some bits are ready, others are not. I don't think we've looked systematically at it yet. 18:13:34 we should probably setup a rhel7 dev instance to test with 18:13:58 * pingou notes we do have an el7 builder on jenkins 18:14:15 #info copr is now using 20 builders, which is nearly as twice than previously 18:14:15 we also have el7b cloud image... 18:15:03 and you can start creating packages in copr, which have rhel7 chroot :) 18:15:16 mirek-hm: so that issue this morning was just that it wasn't terminiating old builders? 18:15:54 nirik: yes, I still dunno why some are not terminated, but I moved the script from cron.daily to cron.hourly 18:16:11 I mean script which terminate those left overs 18:16:11 * nirik would like to start moving to el7 as soon as it's out... but of course there's going to be lots of work needed. 18:16:30 do we have any el5 installs lying around still? 18:16:47 mirek-hm: yeah. Wonder if you could add some debugging to the regular playbooks it uses... it terminates them in there after use right? 18:17:03 yes 18:17:04 lmacken: nope, long gone. ;) app07 lingered on, but it was nuked a while back 18:17:21 nirik: ah okay. I wanted to test that updateinfo sha256 issue, but I guess we'll see what happens :) 18:17:58 lmacken: I have one here... http://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers 18:18:08 thats the one I was testing with the other day 18:18:14 oh, excellent 18:19:10 #info start thinking about el7 porting/bringup 18:19:10 cool, on it with root. will test :) 18:19:39 ok, any other application news? 18:20:20 #topic Sysadmin status / discussion 18:20:30 ok, on the sysadmin side... 18:20:47 #info migrating netapp lun to another filer, so then they can move our nfs volumes. 18:21:01 smooge: any roundup of news on new machines? 18:21:29 systems are being wired in. that si about it 18:21:38 noting from osu or ibiblio on theirs 18:21:51 There is a rhel6 kernel update, so likely I will look at updating *stg* tomorrow and schedule an outage for monday to do the rest. 18:22:01 yay 18:22:05 smooge: we should actively ping them again... (osu and ibiblio) 18:22:12 * pingou also has to reboot 18:22:55 lets see... what else. 18:23:44 We will be ordering some new retrace hardware soon and actually our cloud storage got moved to Q1, so we will be looking for it soon too. 18:24:06 smooge: we should talk to retrace folks and see needs. I can send an email on that. 18:24:18 and we should look around at storage solutions and see what we can get. 18:24:19 I sent one yesterday 18:24:24 waiting for reply 18:24:25 great :) 18:24:44 I sent one ot the retrace people that is 18:25:00 I thought I ccd you but I guess I didn't 18:25:03 smooge: cool. Can you send osu/ibiblio emails too? or want me to? 18:25:19 I will do so (and make sure I cc you :)) 18:25:48 on storage we need something that speaks iscsi (thats what the cloud can consume easily), ideally 2 units so we can fail back and forth, but 1 if it's HA enough I guess. 18:26:12 #info added fed-cloud03 as another cloud compute node this week. 18:27:08 ok, I think thats mostly it on the sysadmin side off hand... 18:27:14 #topic Upcoming Tasks/Items 18:27:14 https://apps.fedoraproject.org/calendar/list/infrastructure/ 18:27:21 anything upcoming folks would like to note or schedule? 18:27:27 I'd like to mention a few things: 18:27:42 * bodhi2/taskotron FAD. 18:27:52 * threebean nods 18:27:57 threebean: has been doing great work moving this forward. 18:28:24 so, we just got FPL approval this morning 18:28:29 oh nice. 18:28:48 so the next step is figuring out, I think, how to actually purchase travel tickets and hotel rooms 18:29:06 did we decide for sure on a date? 18:29:21 early june? 18:29:40 ah, kinda. I sent out an email to all the potential participants suggesting june 1st through june 4th 18:29:50 and I only got back positive confirmations so far. 18:29:54 \รณ/ 18:30:18 cool. 18:30:19 I think that counts as a decision? :P 18:30:27 sounds fine to me. 18:30:29 agreed :) 18:30:33 cool :) 18:31:34 * relrod didn't get an email but would love to go (and those dates should be fine for me) 18:31:57 oh yeah! 18:32:02 relrod: sign up here https://fedoraproject.org/wiki/FAD_Bodhi2_Taskotron_2014 18:32:06 we should possibly mail the infra list at least about it too? 18:32:11 might be others that are interested? 18:32:29 threebean: thanks 18:32:35 cool.. its just that this will impact the scope of the budget that the fpl just approved today. 18:32:46 :( sorry, I didn't know about that page 18:32:56 true. ;( well, we can always cut off today and then re-ask for any additions ? 18:33:27 huh.. I probably should've emailed the list with it sooner. sorry.. first time ;/ 18:33:42 also, now that we know dates, people can fill in actual costs? 18:34:04 yeah, I actually looked up plane fare for the dates and put it in the budget at the bottom already. 18:34:08 ah, nice. 18:34:13 Ha! I will stay 450 miles away the entire time! 18:34:27 smooge: drive on up? ;) 18:34:42 only one car to the household so not really :) 18:35:14 ok, anything else on this fad? :) 18:35:33 where, again? 18:35:38 denver, co 18:35:41 * nirik nods. 18:35:50 did we decide on dates for sure? 18:36:01 dgilmore: everyone said june 1-4th worked. 18:36:08 dgilmore: I got back only "yes"s for june1-4th, yeah. 18:36:56 threebean: nirik: cool, ill look at flights 18:37:41 So, I was wondering about another fad later in the year... 'finish ansible migration'. Would there be interest in that from folks? Would people prefer that be seperate later in the year, or perhaps we could see about tying it to flock and have people arrive early or stay late? (but then flock is longer/more busy) 18:38:03 I would be interested in that. 18:38:13 * jsmith is half-interested 18:38:29 cool. I dont know if everyone interested would also be going to flock? or if seperate would be better? 18:38:30 nirik: i think seperate from flock 18:39:03 separate, and yes interested :) 18:39:03 * pingou is also interested but expensive to fligh 18:39:12 I would be interested in that. 18:39:22 flight expense may be an issue for me as well. 18:39:32 ok. I'll float the idea on the list too, just wanted to toss it out there.... possibly september/oct? 18:39:41 adimania: where are you located again? 18:39:48 India. 18:40:03 ah, ok. Are you going to make it to flock? (in prague) 18:40:15 It might be an issue for me with school though, depending on when it was and how long. :/ 18:40:21 I hope so. 18:40:28 nirik: a couple of days either side of flock might be doable 18:40:49 dgilmore: thats what I was wondering, but as noted it makes flock longer, more time away, etc. 18:41:00 anyhow, will float the idea on the list. 18:41:16 and finally, talking about flock: have folks submitted talks? 18:41:32 Are there any infra wide things we want to submit or work on? workshops, etc? 18:41:34 I have submitted a talk and a workshop. 18:41:38 nirik: I plan on being in .cz for at least 2 weeks anyway 18:41:41 adimania: great! 18:41:47 talk on ansible and workshop on docker. 18:42:02 * dgilmore needs to submit talks last i looked there was cert issues and i didnt look into it closer 18:42:10 I submited copr talk 18:42:12 * tflink may submit a proposal for a talk on phabricator 18:42:14 adimania: ah ha. Yours is the infrastructure ansible one? or the jenkins ansible one? I was thinking of doing an ansible one, but you beat me to it. 18:42:41 dgilmore: deadline is soon. 18:42:47 apr 3rd I think. 18:42:50 nirik: certs sorted out? 18:42:54 so it's really far in advance. 18:42:58 dgilmore, that is because *.rhcloud.com cert is installed there. 18:43:04 smooge / lmacken: ^ 18:43:10 nirik: a bit too far in advance :( 18:43:11 dgilmore, certs aren't going to be settled for a bit 18:43:24 but one of the link goes to somethign.rhcloud.com so then cert is ok. 18:43:29 pretty shoddy to happen 18:43:40 https://flock-lmacken.rhcloud.com/ 18:43:49 pingou: not where i ended up 18:44:01 nirik, mine is infrastructure ansible one but I would be more than happy to collaborate with you on this. :) 18:44:04 nirik, lmacken the flocktofedora.org is not owned by Red Hat. So it can't get a cert for it. Until the domain is transferred the owner (tom) has to get one. 18:44:05 a signed cert really only tells you that someone paid for it. 18:44:07 tflink: oh neat. 18:44:13 yes, one link is registration.flock.com and that have wrong cert 18:44:17 or had 18:44:48 adimania: ah, ok. Yeah, happy to help 18:46:10 I was thinking of submitting some more community type talks... like 'how to get your way in community projects' or something. will come up with some before the deadline I'm sure. 18:46:35 do we want the infra present & future talk? 18:46:42 as in, what we are working on? 18:46:59 we could. have each person do a part of it? 18:47:22 +1 18:47:22 if they want, sure that'll make only better :) 18:47:33 I could submit that if you like. 18:47:59 you can put me co-owner if you want 18:48:49 how do you do that? multiple fas names in that field? 18:49:07 that what I did for the talk with Stanislav 18:49:11 ok 18:49:26 #topic Open Floor 18:49:29 * relrod is happy to speak at that too. Could talk about Fedora Mobile or something :) 18:49:36 any items for open floor? 18:49:41 I won't be there next week, I'll be at LGM w/ ryan :) 18:49:53 (and a number of other Fedora contributors) 18:50:21 open floor item: easyfix tickets for fi-apprentices? 18:50:40 danofsatx-work: sysadmin or dev? 18:50:46 I've been looking for something to cut my teeth on, but the last tickets were submitted in Feb 18:51:03 danofsatx-work: keep meaning to try and file some. ;( will try harder 18:51:05 sysadmin....only code I can write is scripts to to sysadmin work ;) 18:52:17 other than that, nirik knows I want to help with the cloud migration to havana 18:52:18 sometimes it's hard to come up with good ones, but I will try. ;) 18:52:31 documentation :).. 18:53:25 I am going to work some more on mediawiki ansible migration over the weekend. I hope I'll be able to finish it this weekend. 18:54:11 adimania: great. 18:55:05 ok, if nothing else will close out in a minute. 18:55:54 alright. Thanks for coming everyone! 18:55:57 #endmeeting