[02:28:29] New patchset: Rfaulk; "rm. Request notification handler to maintain job queue." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73721 [02:28:29] New patchset: Rfaulk; "rm. refs to request notification handler from api handlers entry point." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73722 [02:28:30] New patchset: Rfaulk; "mod. response handler to utilize file broker context." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73723 [02:28:30] New patchset: Rfaulk; "rm. obselete refs." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73724 [02:28:51] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73721 [02:29:04] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73722 [02:29:22] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73723 [02:29:33] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73724 [03:35:30] New patchset: Rfaulk; "rm. base class initialization - class is effectively virtual." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73727 [03:35:30] New patchset: Rfaulk; "rm. request notification deps." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73728 [03:35:31] New patchset: Rfaulk; "add. handle instances for FileBroker where file not exists on read." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73729 [03:35:31] New patchset: Rfaulk; "add. timouts between broker item polls." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73730 [03:35:52] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73727 [03:36:03] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73728 [03:36:17] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73729 [03:36:29] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73730 [11:15:04] New patchset: QChris; "Switch PageViewEvalFunc's returned schema to tuple" [analytics/kraken] (master) - https://gerrit.wikimedia.org/r/73746 [11:15:33] New patchset: QChris; "Switch Zero's returned schema to tuple" [analytics/kraken] (master) - https://gerrit.wikimedia.org/r/73747 [13:55:06] mornin! [13:56:41] Morning ottomata :-) [14:10:42] qchris, we gonna set up replication to github wikimedia/kraken? [14:11:20] https://gerrit.wikimedia.org/r/#/c/73735/ [14:11:27] ottomata: ^ yes, I'd love to :-) [14:12:25] awesome [14:12:46] yeah, i'm waiting on my zookeeper replication commit too, just want to hear from ^demon that its cool [14:12:52] is he back this week? [14:13:04] I am not sure. [14:13:18] Either this or the next week IIRC. [14:14:32] aye [14:14:33] k [14:14:53] Mhmm ... ottomata is the zookeeper patch already in gerrit? I could only find the jmxtrans one. [14:14:58] (Or is that the same thing?) [14:15:21] (Looks different to me) [14:15:26] oh [14:15:29] sorry it was jmxtrans [14:15:31] you are right [14:15:39] i had almost done zookeeper but decided to to jmxtrans [14:15:47] i have a few more as well [14:16:00] Some that we can do without ^demon? [14:16:17] Or are they all like jmxtrans? [14:16:27] man are you asking me to remember things from a week ago?! [14:16:33] :-))) [14:16:41] No, don't bother. Hehe. [14:16:44] oh right [14:16:48] i remember now [14:17:36] ok, i have 3 new ones: [14:17:41] jmxtrans [14:17:41] zookeeper [14:17:41] kafka [14:18:05] New review: Diederik; "Ok." [analytics/kraken] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/73746 [14:18:06] Change merged: Diederik; [analytics/kraken] (master) - https://gerrit.wikimedia.org/r/73746 [14:18:11] both zookeeper and kafka are brand new repos [14:18:19] so, we ahve to delete the ones on github currently [14:18:39] so, i guess they are the same boat as jmxtrans [14:18:53] ottomata: Yes, I think so too. [14:18:57] but, i'm pretty sure that there is some git hook somewhere [14:19:05] that creates the github repos whenever the gerrit repos are created [14:19:09] and that has nothing to do with replication pushes [14:19:10] so [14:19:16] New review: Diederik; "Ok." [analytics/kraken] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/73747 [14:19:16] Change merged: Diederik; [analytics/kraken] (master) - https://gerrit.wikimedia.org/r/73747 [14:19:18] we can probalby just create the empty repos on each git hub location [14:19:19] and it would work [14:19:34] The latter approach should work. [14:19:42] But I lack permission to do that. [14:19:43] Can you? [14:20:18] yeah [14:20:20] oh [14:20:21] hm [14:20:23] on github? [14:20:24] hm [14:20:28] i *think* so [14:20:46] Yes, on github. [14:20:58] If you can add the repos, we could start replicating right away. [14:21:01] hmmm [14:21:03] right [14:21:03] hm [14:21:05] lemme see [14:21:07] (For me that was the only show stopper) [14:21:49] yup! [14:21:49] https://github.com/wikimedia/puppet-jmxtrans [14:22:18] so, i will merge the jmxtrans one and run puppet on manganese [14:22:20] right? [14:22:49] Looks good to me. [14:26:34] After the puppet run you'll have to reload the replication plugin for replication to grab the new config. [14:30:44] yeah uhhh, how do I do that? :) [14:31:15] ottomata: ssh -p 29418 gerrit.wikimedia.org gerrit plugin reload replication [14:31:31] I guess ops have the permission to run that. [14:33:25] hmm, oook [14:38:10] So upon the next commit on the repo, it should sync. [14:38:22] If you want to force start replication [14:38:36] ssh -p 29418 gerrit.wikimedia.org replication start [14:38:52] ssh -p 29418 gerrit.wikimedia.org replication start --all [14:40:39] ok thanks, i'm pushing the zookeeper one too and then will try that [14:40:53] Ok. [14:54:51] mmmmmmmmmoooooornnnning [14:55:05] yo qchris! [14:55:18] Morning drdee! [14:55:19] thanks for setting up the gerrit repo, i merged your two patch sets, they looked good [14:55:32] I saw the merges. Thanks for them. [14:55:32] although not sure why it didn't work in the beginning [14:55:41] it was returning a schema AFAICT [14:56:14] Yes, it was returning a schame, but modelled slightly different. [14:56:30] Maybe current svn does stricter checking? (No idea) [14:56:38] svn/ [14:56:39] ? [14:57:25] Either way, do some places automatically pull from our git repos (so we should take care nothing got broken)? [14:58:06] no, we have a jenkins setup that pulls from github but is not fully functional because of some missing dependencies [14:58:14] qchris, nothing in production that I know of automatically pulls from there [14:58:16] Ok. [14:58:33] ther eare some cloned repos on the analytics nodes, but nothing official [14:58:43] and those should remote to gerrit eventually when we puppetize them [14:58:46] (if that's how we do it) [14:58:49] but that will be a while [14:58:51] so no worries [14:58:55] So it's safe to break the git repo :-)) [14:59:08] btw. I see that replicaiton kicked in [14:59:18] great [14:59:48] https://github.com/wikimedia/puppet-jmxtrans [14:59:48] ^ has 8 commits. [15:00:01] So it works. [15:00:36] yeehaw [15:00:37] yup! [15:00:41] and this [15:00:42] https://github.com/wikimedia/puppet-zookeeper [15:00:47] shoudl I merge the kraken one? [15:01:04] I guess I'll have to rebase that one :-) [15:01:20] Afterwards ... that would be great. [15:06:25] ottomata: I just rebased https://gerrit.wikimedia.org/r/#/c/73735 [15:06:35] So if you are in a merging mood � :-) [15:17:22] yo milimetric [15:22:39] hoookey dokey qchris, that should be replicating now [15:22:52] you should probably edit the repo discription on github so that it mentions that gerrit is the master [15:25:37] drdee, logrotate is killing kafka producers [15:26:00] why? [15:26:06] when udp2log is sighuped [15:26:14] it kills all its filters [15:26:15] but [15:26:20] java doesn't shut down fast enough for it [15:26:24] and it tries to restart the filter [15:26:36] but the jmx port isn't available because the old java proc is sitll there [15:26:41] ahhhhhhh [15:26:41] so it doesn't start up [15:26:41] but [15:26:46] because of all these shell wrappers [15:26:59] udp2log doesn't notice that the end of the pipe kafka producer isn't running [15:27:04] got it [15:27:07] so it doesn't try to restart it [15:27:39] solution could be: ori's UdpKafka, or if we want to avoid so many changes [15:27:47] i could hardcode the java exec command in the udp2log config file [15:27:50] rather than the shell wrappers [15:27:59] that way udp2log pipes directlyin into java [15:28:01] i think i will try that [15:28:18] hmmm [15:28:22] i think that will work [15:28:40] i'm not certain, since we do have procs before the java proc: grep, udp-filter, etc. [15:29:55] k [15:35:38] nope, same deal [15:36:42] ok, drdee, it only works [15:36:49] if udp2log controls the java process itself [15:36:54] it won't work if there are any pre processing pipes [15:37:25] i could make the udp2log logrotate script sighup, sleep 1 second,a nd then sighup again [15:37:25] that works [15:39:02] hm [15:50:02] a hack on a hack on a hack ;) [15:54:39] drdee: what's this about hacks? [16:15:44] erosen: ask ottomata [16:15:56] hewo? [16:34:30] ottomata: Setting the project description for https://github.com/wikimedia/kraken is easier said then done :-) [16:34:38] ottomata: Once again I am lacking permission. [16:34:46] ottomata: Whom cauld I ask to get permission? [16:34:52] maybe me [16:35:00] * qchris asks drdee :-) [16:35:01] what is your github username? [16:35:09] drdee: somechris [16:35:16] k [16:39:07] try again [16:39:54] Now it works. Thanks drdee [17:01:01] qchris; average; scrum [17:04:42] [travis-ci] develop/cbc561d (#140 by milimetric): The build has errored. http://travis-ci.org/wikimedia/limn/builds/9104040 [17:04:51] New patchset: Rfaulk; "fix. logic to account for empty targets." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73777 [17:04:51] New patchset: Rfaulk; "mod. fix logging for popping request." [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73778 [17:05:13] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73777 [17:05:35] Change merged: Rfaulk; [analytics/user-metrics] (repair_runtime) - https://gerrit.wikimedia.org/r/73778 [17:22:37] hey drdee. puppet-cdh4 maps to operations/puppet/cdh4 right [17:23:42] i belief it does, ask ottomata [17:24:10] YuviPanda: operations/puppet/cdh4 replicates to both puppet-cdh4 and operations-puppet-cdh4 [17:24:40] ah, right. there's the pull request in puppet-cdh4 that i'm configuring to bring back [17:24:42] let me fix [17:24:58] k danke [17:35:38] ok moving on to better internet and lunch [17:35:39] back in a few [18:05:42] thanks YuviPanda! [18:05:42] ottomata: https://gerrit.wikimedia.org/r/#/c/73791/ [18:05:44] :) [18:05:56] ottomata: sorry about the bugginess, just had to add custom name mappign features [18:06:15] ottomata: do keep the comments, etc in github for now, though. that stuff is still a little buggy. [18:06:38] ottomata: and you can tell them to re-open other PRs, should work [18:06:40] ok [18:06:46] ok [18:07:00] so I have 3 other repos I'd like to add custom mapping to [18:07:03] (one isn't ready yet) [18:07:07] s'ok? [18:07:35] ottomata: So yea, sorry about that bastion setup ticket [18:07:41] ottomata: sure [18:07:43] its fixed now and i pinged qchris_away via PM with details [18:07:46] ottomata: let me know and i can add them. one liner [18:08:16] np, thanks RobH [18:08:19] ottomata: you can also add them directly to https://github.com/yuvipanda/SuchABot/blob/master/suchabot/repos.yaml and send a pull request :) [18:08:20] great, should I do the same for stefan? [18:08:23] is a yaml file [18:08:28] https://gerrit.wikimedia.org/r/#/c/69313/ [18:08:38] https://rt.wikimedia.org/Ticket/Display.html?id=5406 [18:08:46] great, danke YuviPanda [18:08:51] :) [18:08:57] so, since the pull request has been merged in gerrit [18:09:02] what do I do for it in github? [18:09:03] just close it? [18:09:10] ottomata: yeah [18:09:10] ok [18:09:10] ottomata: it'll get auto-closing soon [18:09:16] ottomata: it was working for awhile, and then i broke something. [18:10:25] danke! [19:23:36] ottomata: also you might want to change the repo description for https://github.com/wikimedia/puppet-cdh4. it is currently not too helpful [19:42:28] milimetric: got a sec for the hangoout? [19:43:49] YuviPanda: done thanks [19:43:54] hey drdee [19:43:57] i'm not feeling well [19:44:01] taking a break and seeing how it goes [19:44:07] i think i have some stomach thing [19:44:13] aight, that sucks [19:44:13] really hoping it's not the giardia that my cat has [19:44:16] ottomata: sweet. remember to do that for other ones you want to accept pull requests [19:44:24] qchris, around? [19:44:24] milimetric: take care! [19:45:03] kitty feaver? [19:45:08] ottomata, you got a sec, i need to check / confirm whether the new team calendar is working, should take 2 minutes [19:45:08] :D [19:45:17] i am in the hagout [19:45:29] ja [19:57:57] drdee: Now qchris is around. What's up? [19:58:10] nm [19:58:49] Oh. I thought you pinged me before? [19:58:55] Oh, nevermind. [20:04:55] ottomata: did you build .class and .jar using automake ? [20:05:03] * average is trying to do that now [20:05:20] (with not much success so far) [20:06:44] bwerrrrrrrr [20:06:46] what in dclass? [20:06:48] me? [20:06:56] i have done nothing :) [20:06:59] in general.. [20:07:00] ok [20:09:28] surprisingly it just worked, automake is alien [20:10:10] unlike GNU Make, when I use automake I can never tell if something(even easy) is going to work or not [20:10:12] ha, awesome [20:34:01] the gerrit bot si down ? [20:34:03] *is [20:34:39] average: which one? [20:35:44] qchris: gerrit-wm [20:36:10] managed to make the .jar too with automake [20:36:13] There is an open bug that it swallows some messages. [20:36:22] this is the changeset for now, but still needs some things before it can be merged https://gerrit.wikimedia.org/r/73860 [20:37:50] average: about gerrit-wm ... the bug was only for abandoning changes and 2+ inline comments. [20:38:48] Automake to build jars ... sweet :-) [20:40:00] ottomata: bast1001 works now, just as most of analytics10[0-2][0-9].eqiad.wmnet \o/ [20:40:09] ottomata: But stat1002 does not :-( [20:40:38] ottomata: stat1002 gives "Permission denied (publickey)." [20:40:52] ottomata: Is that something we can fix easily, or do I need to file that somewhere? [20:41:44] ahhh grrrrr, did we have an rt for stat1002?! [20:41:46] i made one, rigth? [20:41:47] grr [20:42:00] eeg, i think we didn't [20:42:16] Not sure. We only discussed in a video conf once ... [20:42:28] ah foo [20:42:42] bar [20:42:50] foobar [20:43:05] qchris, sorry [20:43:07] i really thought I did this! [20:43:08] aggghhh [20:43:12] now we have to wait 3 days :( :( :( [20:43:27] Oh. Well ... :-) [20:43:42] I guess then I cannot install it until tomorrow :-D [20:44:48] ottomata: As I lack access to rt, could you please add the ticket for me? [20:44:59] qchris: stat1002.eqiad.wmnet right ? [20:44:59] ottomata: Do I have to get approval from anyone as well? [20:45:03] yes [20:45:04] toby [20:45:07] average: yes [20:45:17] ottomata: Ok, then I'll write an email to toby as well. [20:45:27] k, I CCed him on the RT ticket [20:45:40] ottomata: Ok. Thanks. [20:45:50] well i don't think we should wait another 3 days [20:45:53] that's crazy [20:45:59] OOf average, drdee, anyway we can do the dclass demo now [20:46:01] we need to get christian goign [20:46:11] I rememberd I have to be somewhere at 5.30 [20:46:11] yeah drdee [20:46:17] average: are you ready to demo? [20:46:18] I can ask in ops, not sure who can approve that though [20:46:40] well they approved the analytics machines so this machine belongs to that group [20:46:55] not really, i mean i guess so [20:46:56] does it? [20:47:03] i thikn the stat nodes are a separate group [20:47:46] the same conceptual group ;) [20:47:51] they are all machines that we use [20:48:41] drdee: I'm ready to demo yes [20:48:46] drdee: hangout link please ? [20:48:58] batcave? [20:48:58] https://plus.google.com/hangouts/_/2da993a9acec7936399e9d78d13bf7ec0c0afdbc?authuser=1 [20:49:05] ottomata: that's for me ? [20:49:11] yup [20:49:28] yup [21:04:01] ok outies for now, lataaas [21:43:33] hey drdee [21:43:39] i'm around again [21:43:41] aight [21:43:45] did you still wanna hangout [21:43:59] nahhh it was a quick question about the team calendar [21:44:03] it works now [23:33:00] New patchset: Milimetric; "flake fixes" [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73901 [23:33:35] erosen: I closed db_sessions all over the place and that fixed the "gateway timeout" problem I was having on the deployment [23:33:41] nice [23:33:43] wanna hang out quick so I can pass off some deployment info [23:33:46] yaaa [23:33:50] sorry I've been spotty today [23:34:01] Change merged: Milimetric; [analytics/wikimetrics] (master) - https://gerrit.wikimedia.org/r/73901 [23:34:08] that [23:34:10] 's okay