18:00:11 #startmeeting Infrastructure (2014-10-23) 18:00:11 Meeting started Thu Oct 23 18:00:11 2014 UTC. The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:11 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:00:11 #meetingname infrastructure 18:00:11 The meeting name has been set to 'infrastructure' 18:00:11 #topic aloha 18:00:11 #chair smooge relrod nirik abadger1999 lmacken dgilmore mdomsch threebean pingou puiterwijk 18:00:11 Current chairs: abadger1999 dgilmore lmacken mdomsch nirik pingou puiterwijk relrod smooge threebean 18:00:14 * pingou 18:00:19 * lmacken 18:00:19 * lanica is here. 18:00:47 * threebean is here 18:01:35 * michel_slm is here 18:02:55 * relrod here 18:03:05 #topic New folks introductions and Apprentice tasks 18:03:17 any new folks like to say hi? or apprentices with questions or comments. 18:05:01 I've seen a couple of people passing by asking for help to get started 18:05:13 pingou: yeah, saw another one the other day too... 18:05:18 but I've not mentionned the meeting, my bad :/ 18:05:38 well, hopefully they will find us. 18:05:43 #topic Freeze reminder 18:05:51 just a reminder we are still frozen... ;) 18:05:59 #topic Applications status / discussion 18:06:04 any applications news this week? 18:06:13 well, anitya got few releases :) 18:06:32 cool :) 18:06:46 most things were in order to find out which project's are failing to retrieve a version 18:06:52 I've been working on a fedmsg-driven atomic ostree composer https://github.com/fedora-infra/fedmsg-atomic-composer 18:06:53 which is now available at https://release-monitoring.org/projects/updates/failed 18:06:58 lmacken++ 18:07:08 lmacken: cool. 18:07:12 * oddshocks pops in 18:08:00 pingou: nice. 18:08:24 the menu "projects" at the top gives you all the filters in place atm 18:08:40 pingou: very nice 18:08:50 using these pages (and the keyboard shortcuts added), we went down from 2500+ projects failing to these 440 :) 18:09:06 good progress. 18:09:09 nirik: and the cron is now correctly set-up to use the lock :) 18:09:15 lmacken: are we planning on using that compose machine to test the atomic consumer + MM1 setup? 18:09:32 oddshocks: potentially, I think threebean said nothing is using it right now 18:09:44 I'm almost ready to start composing trees in stg 18:09:48 seems to make sense to me 18:09:52 yeah 18:10:03 should be fine 18:10:21 +1 for installing MM1 on composer.stg for testing this week 18:10:26 we found a problem with the libtaskotron build currently in production (botched merge left out a bunch of code) but we've fixed the build and it's currently running in dev and stg. may be requesting a freeze break in the next week to update the production clients 18:10:27 then we can blow it away later 18:10:31 * pingou looks forward the MM FAD 18:10:55 tflink: sounds good. 18:11:12 pingou: me too. ;) 18:11:42 lmacken: do I need to request access to that box? 18:12:21 oddshocks: the only way to get in is sshing via root from lockbox at the moment. we probably want to change that for now. 18:12:36 sure. 18:12:39 lmacken: cool. 18:12:47 oddshocks: do you have root on lockbox? :) 18:13:09 pingou: I'm not sure, but I don't think so. I have used it to connect to other machines before though obviously 18:13:24 yeah, we should open up just the staging one. add a fas group.. 18:13:37 * nirik nods. 18:13:40 sysadmin? 18:13:41 works for me 18:14:10 well, sysadmin is a tracking group, not a shell group... so sysadmin-something... 18:14:52 we can sort that out of meeting. 18:14:58 anything else application wise? 18:15:02 sysadmin-ostree or something perhaps, that we could apply to compose01.stg temporarily, and then later use it for any actual ostree/atomic machines we set up? 18:15:04 * oddshocks nods 18:16:04 #topic Sysadmin status / discussion 18:16:17 not too much in sysadmin land right now... since we are in freeze 18:16:47 hopefully we will be done with beta soon. 18:17:45 #topic nagios/alerts recap 18:17:49 the new host1plus has rhel7 on it, I've started ansiblizing it 18:17:54 .tiny https://admin.fedoraproject.org/nagios/cgi-bin//summary.cgi?report=1&displaytype=3&timeperiod=last7days&smon=10&sday=1&syear=2014&shour=0&smin=0&ssec=0&emon=10&eday=23&eyear=2014&ehour=24&emin=0&esec=0&hostgroup=all&servicegroup=all&host=all&alerttypes=3&statetypes=3&hoststates=7&servicestates=120&limit=25 18:17:56 nirik: http://tinyurl.com/l6fc4bl 18:18:05 relrod: nice! :) 18:18:08 relrod: oh yeah, thanks much for working on it. ;) it's pretty much set now? or ? 18:18:20 ideally I would like to add a proxy and a mirrorlists there. 18:18:23 (at least) 18:18:44 nirik: pretty much - need to finish ansibilizing it (was waiting for +1's, but got those yesterday), then can set up some VMs on it. 18:19:00 great! 18:19:31 so, on nagios... I think we wanted to just kill the nuancier messages... 18:19:33 I got email that we need to know what boxes we want warranty extensions and which ones need 24x7x4 hour service on. so that is my rest of my day 18:19:39 oops sorry 18:19:51 so nagios... 18:19:55 smooge: yeah good to check everything and see if we aren't missing any 18:20:42 there's some bodhi slowness lately, but I am really not sure why... nothing much in logs. 18:21:27 yeah, nothing's changed in that code in a while 18:21:41 so, it's probably with the db, or to fas, since that happens during most requests 18:21:43 I guess we could try just rebooting those instances... 18:22:19 there have also been random connection dropouts between taskotron and bodhi 18:23:08 +1 on killing the nuancier datanommer check. 18:23:27 +1 18:23:28 .. wait until after freeze for it? 18:23:49 something like smokeping or page load graphing might help us see if it's some time... 18:23:59 yeah, probibly fine to just wait now 18:24:07 what boxes are we wanting on a smokeping? 18:24:42 lmacken: we have a collectd/web-service role in ansible you might be able to use to graph the time to load the bodhi frontpage. 18:24:44 some bodhi page... see if it's always slow or gets slow at specific times, etc. 18:25:05 lmacken: one of the fixes that was supposed to be in taskotron production was retries for errors on bodhi read-only operations 18:25:12 threebean: that would be real interesting 18:26:00 tflink: that's frustrating 18:26:16 the errors are connection dropouts before the client can finish getting the response body 18:26:25 anyhow, more investigation would be good. 18:27:31 any other sysadmin side things? 18:28:09 #topic Upcoming Tasks/Items 18:28:09 https://apps.fedoraproject.org/calendar/list/infrastructure/ 18:28:21 anything upcoming anyone would like to schedule or note? 18:29:22 we should add the FAD to that calendar ;) 18:29:26 +1 18:29:41 #topic Open Floor 18:29:49 for the FAD, we'll need to plan the hotel room at one point 18:29:54 anyone have anything for open floor? 18:30:03 and please update the wiki page when you know your travel details :) 18:30:04 pingou: oh yeah... also when do we get plane tickets? 18:30:07 not me 18:30:21 nirik: I stickster_afk is the one that will book them 18:30:44 I'll ping Matt to know if he received well my last email 18:30:45 ah right. so we need to gather info to send to him... 18:30:46 oh a couple of things 18:30:54 nirik: likely yes 18:31:33 I will be paying for most of the food at the FAD and some of the hotel rooms 18:31:51 smooge: you're my new best friend :D 18:32:20 as I have to book my flight seperately and it makes filing those expenses possible 18:32:42 cool. Thanks smooge 18:32:48 pingou, what is the webpage for the fad again? 18:33:04 smooge: https://fedoraproject.org/wiki/FAD_MirrorManager2_ansible-migration_2014 18:35:24 smooge: you had another thing? or that was it? 18:39:15 that was it 18:39:23 cool. 18:39:35 anything else from anyone? or shall we call it a meeting? 18:40:15 smooge: https://fedoraproject.org/wiki/FAD_MirrorManager2_ansible-migration_2014#Budget_Realized if you already booked your tickets :) 18:40:56 will be doing so in a short while. I need to get hardware dealt with first 18:41:00 sure 18:41:13 alright. Thanks for coming everyone! 18:41:16 #endmeeting