| *** msabramo has joined #openstack-infra | 00:03 | |
| *** devkulkarni has quit IRC | 00:04 | |
| *** amcrn has quit IRC | 00:07 | |
| *** anode has quit IRC | 00:07 | |
| *** kmartin has quit IRC | 00:11 | |
| *** asselin has quit IRC | 00:11 | |
| *** kmartin has joined #openstack-infra | 00:11 | |
| *** asselin has joined #openstack-infra | 00:11 | |
| *** maxbit has joined #openstack-infra | 00:14 | |
| *** asselin has quit IRC | 00:14 | |
| *** asselin has joined #openstack-infra | 00:15 | |
| *** gokrokve has quit IRC | 00:15 | |
| *** annegent_ has joined #openstack-infra | 00:15 | |
| *** rcarrillocruz has quit IRC | 00:16 | |
| *** msabramo has quit IRC | 00:16 | |
| *** rkukura has joined #openstack-infra | 00:18 | |
| *** chuckC has quit IRC | 00:19 | |
| *** annegent_ has quit IRC | 00:21 | |
| *** rcarrillocruz has joined #openstack-infra | 00:21 | |
| *** msabramo has joined #openstack-infra | 00:22 | |
| clarkb | jhesketh: bare-precise images are currently being build in hpcloud | 00:26 |
|---|---|---|
| clarkb | jhesketh: I will do rax shortly too | 00:26 |
| jhesketh | thanks clarkb :-) | 00:27 |
| clarkb | jhesketh: that will update the swift upload script on the node the experimental job runs on | 00:27 |
| clarkb | and hopefully we can make another go at it | 00:27 |
| jhesketh | yep | 00:27 |
| fungi | stab at release notes/announcement for git-review 1.24: https://etherpad.openstack.org/p/git-review-1.24 | 00:30 |
| SlickNik | shashankhegde/ Sukhdev / anteaya: So one thing I noticed when compiling / building this is that the gerrit trigger plugin already has changes that prevent this patch from being cleanly applied on master. | 00:31 |
| clarkb | fungi: looks good. | 00:31 |
| ianw | lifeless: any chance you could approve this -> https://review.openstack.org/#/c/103451/ (d-i-b: Remove fixfiles from rpm-distro finalize) | 00:31 |
| *** zz_dimtruck is now known as dimtruck | 00:31 | |
| SlickNik | shashankhegde/ Sukhdev / anteaya: for example, the gerrit events module has been moved to a separate repository. | 00:31 |
| SlickNik | shashankhegde/ Sukhdev / anteaya: So the patch as it stands might need some TLC before we can merge it upstream | 00:32 |
| clarkb | SlickNik: we don't use gerrit trigger plugin upstream | 00:32 |
| fungi | i'll leave it to collect edits/suggestions, and then tag 1.24 tomorrow | 00:32 |
| * clarkb feels like he is missing context about this | 00:32 | |
| pleia2 | fungi: lgtm | 00:32 |
| fungi | clarkb: pleia2: thanks! | 00:32 |
| anteaya | clarkb: the gerrit trigger plugin doesn't allow rechecks | 00:33 |
| *** packet has quit IRC | 00:33 | |
| *** asselin has quit IRC | 00:33 | |
| anteaya | Sukhdev and shashank hacked a patch so it will | 00:33 |
| clarkb | oh I see | 00:34 |
| clarkb | upstream here is not openstack but gerrit trigger plugin | 00:34 |
| SlickNik | clarkb: Sorry I should have clarified. When I said upstream, I meant merge the patch we have into the gerrit-trigger-plugin codebase. | 00:34 |
| *** xianghui has joined #openstack-infra | 00:34 | |
| anteaya | reddwarf ci doestn' suport rechecks | 00:34 |
| anteaya | clarkb: correct | 00:34 |
| anteaya | and reddwarf ci will still be alive another month, we have to lead by example here | 00:34 |
| SlickNik | anteaya: I'm hoping to kill reddwarf CI before that (I can't wait), but I hear you on the lead by example part. :) | 00:35 |
| anteaya | SlickNik: you are doing great | 00:35 |
| anteaya | keep it up | 00:36 |
| anteaya | and when reddwarf ci dies, we will have a party | 00:36 |
| jesusaurus | it looks like https://review.openstack.org/#/c/104340/ exposes a fun little bug in gerrit. note the two (apparently different) Peter Balland's | 00:36 |
| anteaya | I think I need a new gerrit group since we can't delete accounts, perhaps deprecated-third-party? | 00:36 |
| clarkb | jesusaurus: its a 'bug' peter setup two accounts | 00:37 |
| clarkb | jesusaurus: hard to prevent | 00:37 |
| *** otherwiseguy has joined #openstack-infra | 00:37 | |
| jesusaurus | well, its a bug that both accounts get listed as reviewers when one reviews | 00:37 |
| fungi | someone may have added the other as a requested reviewer | 00:37 |
| anteaya | jesusaurus: if he uses two seperate email accounts | 00:37 |
| anteaya | we can't prevent it | 00:37 |
| jesusaurus | oh, yeah, i suppose one could have been requested | 00:38 |
| jesusaurus | i was thinking that gerrit was doing a lookup on the name | 00:38 |
| fungi | and then he probably responded by reviewing with the other | 00:38 |
| *** markwash has joined #openstack-infra | 00:38 | |
| *** maxbit has quit IRC | 00:38 | |
| *** bdpayne has quit IRC | 00:40 | |
| openstackgerrit | A change was merged to openstack-infra/config: Enable gating for python-congressclient https://review.openstack.org/104340 | 00:42 |
| jesusaurus | is it possible to negate the ref match in a zuul pipeline trigger? as in 'any ref that doesn't match refs/master/foo'? | 00:43 |
| clarkb | jesusaurus: yes, use negative lookahead regex | 00:44 |
| jesusaurus | ah, are those fully pcre? | 00:44 |
| clarkb | should be plenty of examples in our job list that exclude branches that way | 00:44 |
| clarkb | jesusaurus: they are python re lib | 00:44 |
| clarkb | so anything python re will do they can do | 00:44 |
| anteaya | fungi: any thoughts on a gerrit group for retired third party ci accounts? | 00:44 |
| fungi | yeah, we already have tons of zuul project overrides to run on "not this branch but all others" | 00:45 |
| fungi | anteaya: are they ever likely to come out of retirement? if no, is there some other reason to have them grouped? | 00:45 |
| jesusaurus | im talking about gerrit triggers. i want all changes with topic foo to go into their own pipeline | 00:46 |
| anteaya | can we delete them? I thought we couldn't | 00:46 |
| anteaya | and I have zero idea if they will ever come out of retirement or not | 00:46 |
| clarkb | jesusaurus: ya I think that is a regex match as well | 00:46 |
| jesusaurus | awesome | 00:46 |
| clarkb | finally got all the bare-precise builds in progress | 00:48 |
| clarkb | 03 and 04 are still not done with their jobs so now I am going to run home | 00:49 |
| anteaya | fungi: I have commented on your etherpad | 00:49 |
| fungi | thanks anteaya | 00:49 |
| anteaya | np | 00:49 |
| anteaya | mikal: thanks for weighing in on the ml regarding response times | 00:51 |
| *** hdd_ has joined #openstack-infra | 00:51 | |
| *** gokrokve has joined #openstack-infra | 00:55 | |
| IAX | Hi guys, I have a doubt, I'm using the exim module, but its parameters lacks a lot of settings that I need to configure it with sendgrid. | 00:56 |
| anteaya | IAX hi, what is your context? | 00:56 |
| anteaya | what are you doing with the exim module? | 00:56 |
| *** unicell has joined #openstack-infra | 00:56 | |
| IAX | anteaya, I'm configuring to send email to external email address, for this purpose I need to configure it with sendgrid server | 00:57 |
| IAX | the documentation that i'm following is here: https://sendgrid.com/docs/Integrate/Mail_Servers/exim.html | 00:58 |
| IAX | actually the plan is to use this MTA for gerrit | 00:59 |
| anteaya | IAX why are you asking in this channel? | 01:00 |
| IAX | because the module is located in infra. | 01:00 |
| IAX | am I wrong? | 01:00 |
| *** liusheng-leave is now known as liusheng | 01:00 | |
| mordred | IAX: hi! | 01:01 |
| anteaya | I don't know | 01:01 |
| anteaya | ah looks like mordred will help you | 01:01 |
| IAX | hi mordred! | 01:01 |
| mordred | IAX: home in 5 ... will help in just a sec | 01:01 |
| *** arosen has left #openstack-infra | 01:02 | |
| IAX | i'm not sure about my options, the only that comes to my mind is to extend the module to take more input variables and fill the template accordingly | 01:03 |
| *** maxbit has joined #openstack-infra | 01:05 | |
| mordred | IAX: that's probably what we'll need to do - but let me go read the thing real quick | 01:06 |
| jhesketh | clarkb: we should probably merge 101101 if you're removing envinject from real puppets? | 01:07 |
| jhesketh | *jenkins' | 01:07 |
| *** chuckC has joined #openstack-infra | 01:07 | |
| clarkb | ya | 01:08 |
| mattoliverau | IAX: I don't think the exim module handles the update-exim.conf.conf file. What exactly are you trying to do? You could extend it, or in your puppet configuration override the exim4.config.erb or add some extra files to conf.d | 01:08 |
| clarkb | thats the change puppet side? | 01:08 |
| mordred | IAX: ok- I don't think think we need to use al of those | 01:08 |
| jhesketh | clarkb: yep | 01:08 |
| *** yaguang has joined #openstack-infra | 01:09 | |
| *** Ryan_Lane has quit IRC | 01:09 | |
| jhesketh | clarkb: progress: https://jenkins07.openstack.org/job/experimental-swift-logs-config-pep8/10/console | 01:10 |
| mordred | IAX: to start, you want to pass route_list = "* smtp.sendgrid.net::587 byname" | 01:11 |
| mordred | gah | 01:11 |
| mordred | IAX: route_list = "* smtp.sendgrid.net::587 byname" | 01:11 |
| mordred | gah | 01:11 |
| * mordred is failing at cmoputerse | 01:11 | |
| mordred | IAX: you want to pass "* smtp.sendgrid.net::587 byname" to the exim module's smarthost param | 01:12 |
| *** chianingwang has joined #openstack-infra | 01:12 | |
| *** nosnos has joined #openstack-infra | 01:13 | |
| mordred | IAX: one sec for the other bits | 01:13 |
| IAX | mordred, reviewing.... | 01:15 |
| IAX | mordred, but /etc/exim4/passwd.client is not handled in the module | 01:18 |
| IAX | or i'm missing something? | 01:18 |
| clarkb | jhesketh: I am assuming that gives you enough to go off of? | 01:18 |
| * clarkb continues jenkins restarts and plugin removal | 01:19 | |
| jhesketh | yep, looks like the packaged python-magic isn't what I expected | 01:19 |
| *** nosnos has quit IRC | 01:20 | |
| fungi | IAX: the module would likely need to be extended to manage passwd.client for you, and you'd probably want to use something like hiera injection to avoid putting your smarthost account password in your puppet manifest directly anyway | 01:20 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Add support for remote auth smarthosts https://review.openstack.org/104404 | 01:21 |
| *** baohua has joined #openstack-infra | 01:21 | |
| mordred | IAX: there. that should handle it for you ^^ | 01:21 |
| mordred | IAX: (sorry, figured it would just be quicker to write you a patch - hope that's ok) | 01:22 |
| *** Sukhdev has quit IRC | 01:22 | |
| mordred | IAX: obviously that's completely untested ... | 01:22 |
| mordred | IAX: but the key takeaway is that the config section that's relevant is the one under teh cPanel section | 01:23 |
| clarkb | 03 and 04 restarting now | 01:23 |
| mordred | which is TERRIBLE documentation | 01:23 |
| *** weshay has quit IRC | 01:23 | |
| IAX | mordred, cool ! I'll make the test | 01:24 |
| mordred | jeblair: ^^ if you get a second, I'd love a second set of exim eyes on that - I think I did it right | 01:25 |
| clarkb | wouldn't it be nice if hpcloud made ^ not necessary? | 01:25 |
| mordred | clarkb: YES | 01:25 |
| mordred | clarkb: I brought it up with product management again today | 01:25 |
| clarkb | mordred: yay | 01:25 |
| clarkb | 01 and 02 are in shutdown mode now. Iam going t oeat dinner while I wait for them | 01:26 |
| *** baohua has quit IRC | 01:27 | |
| *** atiwari has quit IRC | 01:28 | |
| *** gargola has quit IRC | 01:29 | |
| *** hdd_ has quit IRC | 01:29 | |
| *** gargola has joined #openstack-infra | 01:29 | |
| *** gargola has quit IRC | 01:30 | |
| *** gargola has joined #openstack-infra | 01:31 | |
| *** gokrokve has quit IRC | 01:35 | |
| *** maxbit has quit IRC | 01:35 | |
| *** ArxCruz has joined #openstack-infra | 01:37 | |
| *** marcoemorais has quit IRC | 01:39 | |
| *** marun is now known as marun_afk | 01:39 | |
| *** crc32 has quit IRC | 01:40 | |
| *** otter768 has joined #openstack-infra | 01:41 | |
| *** wenlock has joined #openstack-infra | 01:43 | |
| *** gokrokve has joined #openstack-infra | 01:43 | |
| *** markwash has quit IRC | 01:44 | |
| mordred | nibalizer: you ever figure out anything for the in-tree hiera stuff? | 01:47 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/config: Fix magic in zuul_swift_upload https://review.openstack.org/104411 | 01:47 |
| jhesketh | clarkb: ^ that should fix the latest problem we've seen | 01:47 |
| *** dimtruck is now known as zz_dimtruck | 01:49 | |
| clarkb | jhesketh: so we have installed the lib? | 01:49 |
| clarkb | one of them/ | 01:49 |
| jhesketh | clarkb: yep, I had it installed from apt/yum in preparation for the scripts to work | 01:50 |
| jhesketh | not realising the packaged version was different to the pypi one I was testing on | 01:50 |
| clarkb | gotcha | 01:50 |
| *** otter768 has quit IRC | 01:51 | |
| clarkb | I went ahead and approved it | 01:51 |
| clarkb | and can trigger more image builds in a bit | 01:51 |
| jhesketh | cool thanks :-) | 01:52 |
| *** rkukura has quit IRC | 01:52 | |
| *** rkukura has joined #openstack-infra | 01:52 | |
| *** rkukura has quit IRC | 01:53 | |
| *** otter768 has joined #openstack-infra | 01:53 | |
| openstackgerrit | A change was merged to openstack-infra/config: Fix magic in zuul_swift_upload https://review.openstack.org/104411 | 01:55 |
| clarkb | it didn't occur to me how awesome that commit message is until just now | 01:56 |
| clarkb | jhesketh: good job | 01:56 |
| clarkb | :) | 01:56 |
| *** ArxCruz has quit IRC | 01:56 | |
| jhesketh | I know :-) | 01:56 |
| jhesketh | I was tempted to put in "literally" | 01:57 |
| jhesketh | in hindsight, I should have | 01:57 |
| clarkb | jhesketh: but ya when I fire up laptop again to do jenkins01 and jenkins02 I can do image builds too | 01:57 |
| anteaya | jhesketh: I like it how it is | 01:59 |
| mordred | clarkb: how are we mirror-wise. decent for now? | 01:59 |
| clarkb | mordred: I think so? all the slave images are updated for that at this point | 01:59 |
| clarkb | mordred: and bandersnatch updated the mirror at somepoint this morning | 02:00 |
| mordred | cool. I'm going to work ona patch for bandersnatch to do things at some point | 02:00 |
| *** SumitNaiksatam has joined #openstack-infra | 02:00 | |
| clarkb | mordred: dstufft indicated he didn't want bandersnatch to do things | 02:00 |
| clarkb | unless you mean something on our side parsing the logs | 02:00 |
| dstufft | yea I'd really rathe rnot have every bandersnatch deployment issuing purge requests | 02:01 |
| clarkb | apparently r1chardj0nes commented on that change | 02:01 |
| dstufft | I like that purge is unauthenticated but if that happens i'll probably need to write some VCL to make it authenticated | 02:01 |
| clarkb | and I don't really grok what he means but whatever it merged /me moves on | 02:01 |
| anteaya | mordred: any sense when you can write the thing you were talking about regarding open source? | 02:01 |
| anteaya | mordred: I'd really like a url I can point to | 02:01 |
| *** blamar has joined #openstack-infra | 02:01 | |
| dstufft | (but openstack itself can purge stuff becuase when Iactually fix it I can go bother y'all to stop doing it) | 02:01 |
| mordred | dstufft: oh. ok | 02:03 |
| dstufft | mordred: the fact that purges fail is a bug in PyPI basically | 02:03 |
| dstufft | that's how I see it anyways | 02:03 |
| mordred | dstufft: then I'll just write a thing to catch them in our runs :) | 02:03 |
| dstufft | yea that's fine | 02:03 |
| mordred | dstufft: and turn it off when you tell me to | 02:03 |
| mordred | btw - I submitted a patch to bandersnatch and if totes failed pyflakes because I SUCK | 02:03 |
| dstufft | mordred: if you do that, if you log some metrics about how often it happens that'd be cool | 02:04 |
| dstufft | I don't have any idea how often it happens | 02:04 |
| dstufft | if it's happening a lot i'll prioritize fixing it | 02:04 |
| mordred | dstufft: I can TOTALLY do that | 02:04 |
| mordred | we have a graphite and everything | 02:04 |
| *** guangjun has joined #openstack-infra | 02:05 | |
| dstufft | there's a lot of broken things and not a lot of data about how often those broken things are actually affecting people | 02:05 |
| dstufft | so I just kind of guess on what my priorities sohuld be | 02:05 |
| *** bdpayne has joined #openstack-infra | 02:05 | |
| dstufft | mostly based on how often people are yelling at me about something :D | 02:05 |
| clarkb | jhesketh: did you want to address richard's comment before image rebuilds happen? | 02:05 |
| mordred | well, I believe your priorities shoudl be "fix things #openstack-infra carps about, then other stuff" | 02:06 |
| clarkb | jhesketh: I don't think it is that important | 02:06 |
| jhesketh | clarkb: yeah I guess it'll only take a second... please hold | 02:06 |
| clarkb | ok | 02:06 |
| clarkb | jhesketh: I left a comment in response | 02:06 |
| clarkb | jhesketh: I am actually nto a fan of try except else | 02:06 |
| clarkb | but whatever | 02:07 |
| clarkb | (else in funny places just never reads right, sure python lets you for else, try else, else all the things, but arg readability) | 02:07 |
| dstufft | clarkb: you cray | 02:07 |
| dstufft | I <# else | 02:07 |
| dstufft | <3 | 02:07 |
| dstufft | ;) | 02:07 |
| clarkb | dstufft: it doesn't help readability imo | 02:07 |
| clarkb | dstufft: especially for people not familiar with python | 02:08 |
| clarkb | for else wtf <- seems to be a common thing | 02:08 |
| *** moted has quit IRC | 02:08 | |
| *** MarkAtwood has quit IRC | 02:08 | |
| dstufft | clarkb: eh, it's nicer than the alternative, which is setting a flag and then testing the flag | 02:08 |
| *** Johnson_Yang has joined #openstack-infra | 02:08 | |
| mordred | I like try except else - but is for else really a thing? | 02:08 |
| clarkb | mordred: yes for else is a thing | 02:09 |
| mordred | O M G | 02:09 |
| mordred | wow | 02:09 |
| mordred | that's so ruby | 02:09 |
| clarkb | see | 02:09 |
| clarkb | this is the problem with else in funny places | 02:09 |
| *** Johnson_Yang has quit IRC | 02:09 | |
| clarkb | dstufft: or you could write less crazy flow control | 02:09 |
| mordred | anteaya: sorry - I got swamped this week | 02:09 |
| clarkb | like try except maybe I should bail out here, then do stuff if not | 02:09 |
| *** KarenZhu has joined #openstack-infra | 02:09 | |
| dstufft | clarkb: sometimes you gotta do a thing | 02:09 |
| mordred | clarkb: you know the best thing about teh bandersnatch mirror? | 02:09 |
| dstufft | I don't use for else every day or anything | 02:10 |
| mordred | clarkb: now I don't have to refactor pypi-mirror and make it suck less | 02:10 |
| clarkb | mordred: :) | 02:10 |
| anteaya | mordred: you're always swamped | 02:10 |
| anteaya | but I will keep hoping | 02:10 |
| clarkb | dstufft: ya and I'm not saying python shouldn't have it either | 02:10 |
| mordred | clarkb: we might want to RFA it, btw | 02:10 |
| clarkb | dstufft: but mordreds response basically makes my point for me which is nice | 02:10 |
| mordred | clarkb: I think the HP internal people are using it still because bandersnatch doesn't work right with proxies or something | 02:10 |
| mordred | OR | 02:10 |
| clarkb | yes the proxies | 02:10 |
| mordred | maybe we should just fix bandersnatch to work with proxies | 02:10 |
| dstufft | ^^ | 02:10 |
| clarkb | but they should learn to tsocks | 02:10 |
| dstufft | bandersnatch uses requests I think | 02:11 |
| dstufft | it should already work with proxies... | 02:11 |
| dstufft | oh xmlrpc | 02:11 |
| dstufft | no wait | 02:11 |
| mordred | yup | 02:11 |
| dstufft | it's using xmlrpc2 | 02:11 |
| dstufft | which uses requests | 02:11 |
| wenlock | proxies make me sad these days | 02:11 |
| mordred | maybe it works with them NEW | 02:11 |
| mordred | NOW | 02:11 |
| clarkb | wenlock: me too. this is why I don't use them | 02:11 |
| *** gokrokve has quit IRC | 02:11 | |
| mordred | it didnt' work 6 months ago | 02:11 |
| mordred | wenlock: ++ | 02:11 |
| mordred | proxies are stupid | 02:11 |
| *** gokrokve has joined #openstack-infra | 02:11 | |
| dstufft | I have an awesome proxies are stupid bug | 02:11 |
| clarkb | but really if a proxy is causing problems. tsocks is almost always a sanish answer | 02:11 |
| wenlock | plus its fake sense of network security when you ip everything public anyway | 02:12 |
| dstufft | requests (or maybe urllib3) stops respecting proxies after 51 requests | 02:12 |
| dstufft | but only on some sites | 02:12 |
| dstufft | https://github.com/pypa/pip/issues/1805 | 02:12 |
| wenlock | i mean routable ip's | 02:12 |
| clarkb | I mean you can even tether off your phone using adb and connectbot with tsocks in a mostly crazy but useable way | 02:12 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/config: Don't catch too much in zuul_swift_upload https://review.openstack.org/104416 | 02:12 |
| dstufft | that doesn't compare to this bug though http://stackoverflow.com/questions/24522467/python-program-hangs-forever-when-called-from-subprocess | 02:12 |
| jhesketh | clarkb: ^ | 02:12 |
| clarkb | jhesketh: cool | 02:13 |
| clarkb | jhesketh: I am assuming you know the magic to poke richard to rereview? | 02:13 |
| wenlock | ive been fighting jenkins today to make it proxy happy | 02:13 |
| jhesketh | yep, already pinged him | 02:13 |
| clarkb | cool I will wait before approving this time | 02:13 |
| wenlock | using openid plugin | 02:13 |
| jhesketh | he's smart enough to not be in this channel | 02:13 |
| clarkb | wenlock: you should use my solution | 02:14 |
| clarkb | wenlock: just don't use a proxy | 02:14 |
| clarkb | >_> | 02:14 |
| wenlock | haha | 02:14 |
| *** homeless has quit IRC | 02:14 | |
| dstufft | clarkb: time to make a movement to break proxy support in all software | 02:14 |
| dstufft | imo | 02:14 |
| wenlock | clarkb i might have to do that if i can't get a solid working config | 02:14 |
| jesusaurus | dstufft: clarkb: ++ lets do this | 02:14 |
| mordred | wenlock: also, go ahead and ditch jenkins and move to jhesketh's turbo-hipster | 02:16 |
| wenlock | carefull someone might try to do that with firmware, then we'll really be in trouble | 02:16 |
| wenlock | mordred is that a real thing ? like 1-4 week thing? | 02:16 |
| mordred | no | 02:16 |
| mordred | it's CLOSE | 02:16 |
| wenlock | nice | 02:16 |
| mordred | but it is not a 1-4 week thing for you | 02:17 |
| wenlock | we're going to be upgrading our components soon... would be nice to role that in too | 02:17 |
| mordred | unless jhesketh disagrees | 02:17 |
| wenlock | might just call it a new blueprint for us | 02:17 |
| jhesketh | so I'm not sure how turbo-hipster would go running devstack jobs... but at the moment it works for our needs | 02:18 |
| clarkb | ok I am going to do 01 and 02 then jenkins.o.o real soon now | 02:18 |
| jhesketh | I wouldn't recommend it to anybody who doesn't want to do hand holding of the system | 02:18 |
| jhesketh | clarkb: Richard has +1'd it | 02:19 |
| clarkb | jhesketh: I think you can self approve | 02:19 |
| mordred | jhesketh: next step for me in my ansible puppet patches | 02:20 |
| mordred | jhesketh: is to make a turbo-hipster thing to run the ansible script | 02:20 |
| mordred | jhesketh: so I may have some questions for you | 02:20 |
| jesusaurus | mordred: how are you liking ansible? how does it compare to salt? | 02:20 |
| clarkb | 02 is coming back now | 02:21 |
| mordred | jesusaurus: I like it a lot | 02:21 |
| mordred | jesusaurus: the reason I like it more than salt is that I can personally use it to do things on infra without infra having chosen to use it | 02:21 |
| mordred | jesusaurus: which makes it WAY nicer to experiment with | 02:21 |
| jesusaurus | oh nice | 02:21 |
| mordred | yah. I mean, I've done real activities on our infra with ansible from my laptop and infra does not yet use ansible | 02:22 |
| jesusaurus | but that seems like it also opens up the doors for everything being magic | 02:22 |
| mordred | well, it's just ssh | 02:22 |
| mordred | jesusaurus: http://paste.openstack.org/show/85366 | 02:22 |
| jesusaurus | well, i mean, if i want to know where/how something is happening, i may not even know what host the command is coming from | 02:23 |
| mordred | jesusaurus: I did that a little while ago to do a project rename | 02:23 |
| openstackgerrit | A change was merged to openstack-infra/config: Don't catch too much in zuul_swift_upload https://review.openstack.org/104416 | 02:23 |
| mordred | jesusaurus: well, in general I'm starting to commit playbooks to config | 02:23 |
| clarkb | and 01 is coming back now | 02:23 |
| clarkb | jenkins.o.o is in shutdown mode now | 02:24 |
| jesusaurus | mordred: i don't see how you get orchestration out of that. it just looks like 'do all these things at once' without any ordering to them | 02:25 |
| mordred | jesusaurus: it's ordered... | 02:25 |
| jhesketh | clarkb: okay ready for a rebuild whenever you are | 02:26 |
| clarkb | jenkins.o.o is coming back now | 02:26 |
| clarkb | jhesketh: I will trigger them shortly | 02:26 |
| jesusaurus | mordred: can you repeat hosts? i thought we were trying to solve the 'do X on host 1, Y on host 2, then Z on host 1' problem | 02:26 |
| mordred | jesusaurus: what that file says is "on git*, in parallel, remove that thing, then on review.o.o, remove and move thigns, then on *slave.o.o, in parallel, remove workspaces" | 02:26 |
| clarkb | ok jobs running in jenkins.o.o | 02:27 |
| clarkb | jesusaurus: generally we don't have to do two things on host 1 | 02:28 |
| clarkb | jesusaurus: its typically host1 then host 2 then host3 type situations | 02:28 |
| mordred | clarkb: look at the file above | 02:28 |
| jesusaurus | oh, okay | 02:29 |
| clarkb | mordred: the file above? | 02:29 |
| mordred | clarkb: http://paste.openstack.org/show/85366 | 02:29 |
| clarkb | ya thats host1 then host2 then host3 | 02:29 |
| mordred | clarkb: yah, but there are 3 tasks on host 2 | 02:29 |
| mordred | and host 1 and host 3 are actually sets of hosts | 02:29 |
| clarkb | sure | 02:29 |
| mordred | but yeah | 02:29 |
| clarkb | I was just contrasting with jesusaurus statement | 02:30 |
| mordred | gotcha | 02:30 |
| mordred | jesusaurus: you CAN do then Z on host 1 though if you need to | 02:30 |
| jesusaurus | yeah, thats what i was wondering | 02:30 |
| clarkb | jhesketh: bare-precise rebuilds going now | 02:30 |
| jhesketh | shiny | 02:30 |
| clarkb | we are going to use all the bandwidths today | 02:30 |
| clarkb | mordred: the install pip step fails at a relatively high rate btw | 02:32 |
| clarkb | I have seen 2 fails today on that step | 02:32 |
| mordred | clarkb: install pip step? | 02:33 |
| clarkb | for image builds | 02:33 |
| clarkb | we get connect reset errors | 02:34 |
| mordred | gotcha. hrm. thats odd | 02:34 |
| mordred | I blame dstufft | 02:34 |
| dstufft | is it installing from PyPI | 02:34 |
| dstufft | or from pypi.o.o | 02:34 |
| clarkb | fro mthe get pip script | 02:34 |
| clarkb | and a wheels eems to be involved somehow | 02:34 |
| clarkb | because the install step for the wheels is what fails | 02:35 |
| dstufft | the get-pip.py script supports everything pip does, so that doesn't tell me if it's pypi.p.o or pypi.o.o | 02:35 |
| mordred | dstufft: it shoudl be pypi.p.o | 02:35 |
| dstufft | ok | 02:35 |
| mordred | clarkb: I think what we might want to do there is figure out what to tell dib to cache and populate and then do get-pip --no-download | 02:36 |
| dstufft | I was just verifying, there is a known connection reset bug we're debugging | 02:36 |
| dstufft | If you can get tcpdumps of connection reset errors that would be really great | 02:36 |
| clarkb | http://paste.openstack.org/show/85367/ | 02:36 |
| dstufft | th eonly thing I have access too that can reliabliy reproduce the connection reset error is Travis, and trying to get tcpdumps in that isn't working very well :( | 02:37 |
| *** ArxCruz has joined #openstack-infra | 02:37 | |
| *** ArxCruz has quit IRC | 02:37 | |
| clarkb | I mean it could be a local issue too | 02:37 |
| dstufft | Connection Reset by Peer is being reported by a number of people in their CI systems | 02:38 |
| openstackgerrit | Julia Kreger proposed a change to openstack-infra/gitdm: Add Julia Kreger https://review.openstack.org/104421 | 02:38 |
| mordred | dstufft: we have a large farm of thigns ... you want me to set up something to massively attack your url and email you tcpdumps? :) | 02:38 |
| mordred | TheJulia: woot | 02:38 |
| dstufft | mordred: sure! Fastly isn't sure what's causing it, and my travis tcpdumps are cut short for some reason and don't include the RST packet | 02:39 |
| dstufft | and I can't manually repro it | 02:39 |
| clarkb | just need an until ! python get-pip.py | 02:40 |
| anteaya | dstufft: where are you seeing people reporting this from their CI systems? | 02:40 |
| morganfainberg | mordred, clarkb, jeblair, should i/do i need to mark the keystonemiddleware package as wheels? | 02:40 |
| clarkb | morganfainberg: should be nothing | 02:40 |
| clarkb | morganfainberg: wheels are free | 02:40 |
| *** bdpayne has quit IRC | 02:41 | |
| dstufft | anteaya: travis-ci was one, and ryanpetrello was having issues on openstack CI (his jobs hit pypi.p.o not the openstack mirror), and uh, people were telling me it was affecting their inhouse stuff too | 02:41 |
| morganfainberg | clarkb, ok so add it to setup.cfg or not worry? | 02:41 |
| clarkb | morganfainberg: I don't think you need to do anything | 02:41 |
| morganfainberg | clarkb, we're within 1 patch merging of being ready to release | 02:41 |
| mordred | morganfainberg: you CAN mark is with universal if you want | 02:41 |
| morganfainberg | clarkb, ok cool. just trying to keep it as easy on everyone as possible :) | 02:41 |
| clarkb | oh right | 02:41 |
| *** bdpayne has joined #openstack-infra | 02:41 | |
| clarkb | universal means it all the things | 02:41 |
| clarkb | but i have a feeling it isn't universal | 02:41 |
| clarkb | because keystone | 02:41 |
| *** zz_dimtruck is now known as dimtruck | 02:42 | |
| mordred | clarkb: it means "dont' build py2 and py3 specific wheels" | 02:42 |
| morganfainberg | clarkb, keystoneclient is "universal" | 02:42 |
| anteaya | dstufft: okay thanks | 02:42 |
| mordred | morganfainberg: then you should add the universal line | 02:42 |
| morganfainberg | clarkb, this code is no different than ksc's | 02:42 |
| clarkb | morganfainberg: but middleware is from keystone I thought | 02:42 |
| clarkb | oh it came from the client? | 02:42 |
| morganfainberg | clarkb, yep | 02:42 |
| clarkb | then yes universal | 02:42 |
| dstufft | if setup.py doesn't output different bits on different pythons or platforms or what have you | 02:42 |
| dstufft | then it is universal | 02:42 |
| clarkb | dstufft: even if it doesn't work on python3? | 02:43 |
| anteaya | dstufft: I don't know travis-ci | 02:43 |
| clarkb | anteaya: travis-ci is pretty cool | 02:43 |
| mordred | clarkb: yah. because tarballs may not work on ptyhon3 and we upload them anyway | 02:43 |
| anteaya | dstufft: it is an account on our gerrit? | 02:43 |
| dstufft | anteaya: no it's a CI system | 02:43 |
| dstufft | travis-ci.org | 02:43 |
| anteaya | okay thanks | 02:43 |
| dstufft | lots of OSS projects use it | 02:43 |
| clarkb | anteaya: its basically ci for the masses | 02:43 |
| mordred | dstufft: it's a thing you can use to do CI on your github | 02:43 |
| anteaya | ah | 02:43 |
| mordred | at least, it is until we have a zuul github hook | 02:44 |
| morganfainberg | clarkb, mordred, i'll mark it as universal, if we can get it in this release great, if not we'll have a 2nd release coming (with a bunch of refactoring) in the near-term this cycle. initial release is so we can get adoption and make sure we don't break anything when projects use it | 02:44 |
| mordred | :) | 02:44 |
| mordred | morganfainberg: kk. it's not urgent. infra works either way | 02:44 |
| morganfainberg | mordred, ++ cool | 02:44 |
| clarkb | if I didn't already run a really awesome ci system I would probably use travis :) | 02:44 |
| dstufft | clarkb: yes, the wheel stuff isn't about declaring what versions of Python you support, it's about selecting which wheel file has the correct bits for what would happen if you ran ``setup.py on that Python`` | 02:44 |
| clarkb | dstufft: gotcha | 02:45 |
| dstufft | we have another thing in metadata 2.0 which will declare supported versions | 02:45 |
| clarkb | and since 2to3 is a thing you have to distinguish | 02:45 |
| dstufft | yea | 02:45 |
| dstufft | plus people can do crazy stuff in their setup.py | 02:45 |
| anteaya | would travis-ci be something third party groups could use? | 02:45 |
| clarkb | right | 02:45 |
| anteaya | the ones that have such a hard time keeping a jenkins running? | 02:45 |
| clarkb | anteaya: maybe? they would have to convince travis to talk to our gerrit | 02:45 |
| wenlock | clarkb so we took the ec2.rb rm route you mentioned to me yesterday, this seems to be working on our end...it fixed our demo area for sure | 02:45 |
| clarkb | and I don't think travis ci does that today | 02:45 |
| clarkb | wenlock: ok | 02:46 |
| anteaya | I wonder if that is worth putting any effort into | 02:46 |
| clarkb | wenlock: I thnik we knew it would most work. we just felt dirty doing it | 02:46 |
| clarkb | anteaya: well its closed source iirc | 02:46 |
| clarkb | anteaya: so you can't do anything about it | 02:46 |
| * clarkb double checks that may not be true | 02:46 | |
| wenlock | clarkb yeah, i can see that, so we made it a hiera configuration option you can flip on and off | 02:46 |
| dstufft | clarkb: if you made it so setup.py failed on Python 3, then you wouldn't publish a wheel though (like if sys.version_info.major == 3: raise Exception("Don't do this thing") | 02:46 |
| dstufft | travis-ci is OSS | 02:46 |
| dstufft | "OSS" | 02:47 |
| mordred | anteaya: the big key would be convincing them to care | 02:47 |
| wenlock | clarkb for how we provision the puppetmaster | 02:47 |
| dstufft | they have an OSS license | 02:47 |
| dstufft | and you can see the code | 02:47 |
| mordred | oh really? | 02:47 |
| mordred | neat | 02:47 |
| dstufft | but it's near impossible to run a copy on your own | 02:47 |
| mordred | I did not know that | 02:47 |
| dstufft | there is zero instructions on how to do it | 02:47 |
| mordred | ah ... the launchpad thing | 02:47 |
| dstufft | and it's fairly complex | 02:47 |
| clarkb | oh hey it is open | 02:47 |
| clarkb | https://github.com/travis-ci/travis-api is what would need updating I think | 02:47 |
| clarkb | but maybe not | 02:48 |
| anteaya | clarkb: ah closed source doesn't help | 02:48 |
| anteaya | :( | 02:48 |
| clarkb | hrm I don't see any licenses on that | 02:48 |
| openstackgerrit | A change was merged to openstack-infra/gitdm: Add Julia Kreger https://review.openstack.org/104421 | 02:48 |
| dstufft | I would talk to them first before trying to write some code, because I have a feeling they probably consider it out of scope | 02:48 |
| clarkb | dstufft: is it open source or just on github? | 02:48 |
| anteaya | mordred: yeah, well if they are closed source they probably don't care | 02:48 |
| clarkb | because I don't see a clear indication of what the license is | 02:48 |
| clarkb | anteaya: well it may be open, trying to figure out what their code is licensed under | 02:49 |
| dstufft | clarkb: hm, I thought there was a license | 02:49 |
| dstufft | I can ask them | 02:49 |
| dstufft | they are friends | 02:49 |
| clarkb | oh MIT | 02:49 |
| clarkb | at least for -core | 02:49 |
| clarkb | -api wasn't clear | 02:49 |
| clarkb | so ya open source cool | 02:49 |
| dstufft | ah cool | 02:50 |
| anteaya | clarkb: ah travis-ci is open source? | 02:50 |
| anteaya | woooo | 02:50 |
| clarkb | anteaya: well some parts at least :) | 02:50 |
| dstufft | travis-ci is one of those things, where it's really good for the ~80% use case | 02:50 |
| clarkb | others seem to have the typical problem of no clear license even though I can read the code | 02:50 |
| clarkb | dstufft: just like github | 02:50 |
| dstufft | yea | 02:50 |
| dstufft | pip uses them, and I constantly consider moving off because we'd like OSX and Windows support (they techincally have OSX support, but the Python support is broken) plus things like "python without threads" and stuff like that | 02:51 |
| *** MarkAtwood has joined #openstack-infra | 02:51 | |
| clarkb | python without threads? | 02:51 |
| dstufft | but their workflow makes me happy and not having to maintain infra makes me happy | 02:51 |
| *** otherwiseguy has quit IRC | 02:51 | |
| dstufft | clarkb: you can compile python without threading enabled | 02:52 |
| dstufft | pip supports being run in this configuration | 02:52 |
| dstufft | we don't actually test it though | 02:52 |
| clarkb | oh I see | 02:52 |
| dstufft | theortically pip wants to support running with everything but zlib disabled (lzma, threading, ssl, etc) | 02:52 |
| dstufft | but no tests, and the only way to do that on travis is compile python inline, or make your own packages and install them inline in the test run | 02:53 |
| anteaya | well if it isn't clearly all open source, I don't think I can recommend it | 02:53 |
| anteaya | if someone chooses to use it, well I will have to wait to hear what devs think and then assess | 02:53 |
| clarkb | anteaya: I think it "is" its just a common issue where people neglect to make it explicit so it technically may not be | 02:53 |
| clarkb | anteaya: the problem is large enough that github now suggests licenses for you when you create new repos | 02:54 |
| dstufft | [22:49:53] <dstufft> josh-k: is there a license for the travis-ci code? | 02:54 |
| dstufft | [22:52:09] <@josh-k> MIT I believe | 02:54 |
| dstufft | [22:53:21] <dstufft> josh-k: ah ok, didn't see a license here https://github.com/travis-ci/travis-api | 02:54 |
| dstufft | josh-k is the uh | 02:54 |
| dstufft | CEO or something | 02:54 |
| dstufft | of travis-ci | 02:54 |
| clarkb | ya | 02:54 |
| anteaya | dstufft: ah okay thanks | 02:54 |
| dstufft | so msot likely if something is missing a license it was an oversight | 02:55 |
| clarkb | ya | 02:55 |
| clarkb | its a common thing | 02:55 |
| anteaya | dstufft: would josh-k care enough to chat with me about our third party ci situation to see if travis-ci is something that would fit? | 02:55 |
| *** otherwiseguy has joined #openstack-infra | 02:55 | |
| clarkb | not intentionally done | 02:55 |
| dstufft | anteaya: possibly! He's in #travis on freenode | 02:55 |
| clarkb | anteaya: thats a can of worms. my guess is way out of scope for them | 02:56 |
| clarkb | anteaya: also doesn't help third most third parties as they could test on us if they can test on travis | 02:56 |
| anteaya | ah | 02:56 |
| anteaya | fooey | 02:56 |
| clarkb | the reason they don't test on us is they have some proprietary thing or hardware or whatever else | 02:56 |
| anteaya | right | 02:56 |
| anteaya | sigh | 02:56 |
| clarkb | and those things pose the same problems for travis as they do for us | 02:56 |
| anteaya | can they not just leave that behind | 02:56 |
| anteaya | and test on us? | 02:56 |
| anteaya | that would be so much easier | 02:56 |
| anteaya | for them too | 02:56 |
| anteaya | wondering if we should have a ci conf | 02:57 |
| anteaya | us, travis-ci, other things | 02:57 |
| *** bdpayne has quit IRC | 02:57 | |
| clarkb | these things exist | 02:58 |
| anteaya | then the third party folks could show up | 02:58 |
| anteaya | they do? | 02:58 |
| anteaya | what are they called? | 02:58 |
| clarkb | I think a lot of people show up at jenkins conf even if not using jenkins | 02:58 |
| clarkb | there are also sqa conferences | 02:58 |
| anteaya | I didn't even know sqa was a thing | 02:59 |
| anteaya | I dislike acronyms | 02:59 |
| clarkb | TLAs FTW | 02:59 |
| * clarkb hides | 02:59 | |
| anteaya | isn't jenkins conf a jave conf | 02:59 |
| wenlock | is third party ci just about running a ci system that connect and vote on o.o? | 02:59 |
| anteaya | yes | 03:00 |
| wenlock | like behind a proxy | 03:00 |
| anteaya | if there is a reason to do so | 03:00 |
| anteaya | preferably not | 03:00 |
| clarkb | wenlock: yes. the idea beind it is to allow people with special requirements to deal with that stuff on their end | 03:00 |
| wenlock | they should use redstone, wink wink | 03:00 |
| anteaya | proxies and firewalls make it unhappy | 03:00 |
| mordred | wenlock: funny story - I was just telling anteaya about forj | 03:00 |
| clarkb | because I don't want to run specia lswitch hardware and deal with licensing | 03:00 |
| mordred | wenlock: however, you should make a gerrit-less blueprint | 03:00 |
| clarkb | anteaya: no jenkins conf is focused on jenkins | 03:00 |
| wenlock | we have | 03:00 |
| wenlock | tuleap and modus are gerritless | 03:00 |
| clarkb | anteaya: but there is leakage into the CI community | 03:01 |
| mordred | wenlock: do they have code hosting? | 03:01 |
| wenlock | perforce wants to make one too... ive been trying to get them to do zuul | 03:01 |
| wenlock | no, its all host yourself | 03:01 |
| wenlock | no $$ to host it | 03:01 |
| mordred | wenlock: I mean, one that is redstone is just points to openstack's gerrit | 03:01 |
| clarkb | people use perforce? | 03:01 |
| clarkb | thats silly | 03:01 |
| wenlock | mordred ahhh that is intreasting | 03:01 |
| mordred | wenlock: we have 10s of groups all trying to run third-party ci pointed at our gerrit | 03:01 |
| wenlock | mordred +1 thats a new blueprint for us, we'd host the blueprint source for that for sure if we wanted to start one | 03:02 |
| anteaya | about 80 accounts | 03:02 |
| anteaya | https://review.openstack.org/#/admin/groups/270,members | 03:02 |
| wenlock | mordred it would be clone redstone and give it a new name... call it some other cool name | 03:02 |
| anteaya | subgroup of above: https://review.openstack.org/#/admin/groups/91,members | 03:02 |
| wenlock | :P | 03:02 |
| clarkb | jhesketh: images are just about done. will let you know when they are all up to date. should be soon | 03:03 |
| jhesketh | awesome, thanks clarkb | 03:03 |
| anteaya | clarkb: hmmmm, I would really like an infra, travis, forj conf I think maybe | 03:03 |
| *** shashankhegde has joined #openstack-infra | 03:03 | |
| * StevenK tries to understand ZUUL_* | 03:04 | |
| *** MarkAtwood has quit IRC | 03:04 | |
| StevenK | Clearly, I can check ZUUL_PROJECT if it's set to something I care about, but then how do I get at the code, or is already on the machine then, and I just need to pip install --pre ? | 03:05 |
| anteaya | I'm wrong, 70 accounts: http://paste.openstack.org/show/85368/ | 03:05 |
| cody-somerville | anteaya: That's not a bad idea. Why don't you try and schedule something to happen during the OpenStack summit? | 03:05 |
| clarkb | StevenK: if you run gerrit git prep as the first part of your job all of that is taken care of | 03:05 |
| anteaya | oh god not during summit | 03:05 |
| clarkb | StevenK: if you aren't for whatever reason then the zuul url prefix + project tell you where to clone the zuul ref from | 03:06 |
| anteaya | krtaylor: has offered to do something in Texas next release | 03:06 |
| clarkb | s/clone/fetch/ | 03:06 |
| StevenK | clarkb: This is for tripleo jobs. One of things we run is pip install <every client> in pull-tools. If we're testing os-cloud-config or os-apply-config, I want to install that version so we actually use the code | 03:06 |
| anteaya | so perhaps late nov, early dec or maybe jan | 03:06 |
| clarkb | StevenK: that version being the one proposed? | 03:06 |
| anteaya | some sort of meetup targetted at the third party ci accounts, including stuffl like how to set up your system | 03:07 |
| StevenK | clarkb: Right | 03:07 |
| clarkb | anteaya: I really don't like the focus on "third party" fwiw | 03:07 |
| anteaya | clarkb: do expand | 03:07 |
| clarkb | anteaya: if you do something it really should be more general imo | 03:07 |
| anteaya | ah | 03:07 |
| clarkb | anteaya: third party openstack testing is so specific and imo not very interesting | 03:07 |
| anteaya | okay, what term do you suggest | 03:07 |
| cody-somerville | anteaya: What about a virtual meetup? | 03:07 |
| clarkb | its a thing we do to deal with a problem | 03:07 |
| clarkb | anteaya: I like the broader CI get together | 03:07 |
| anteaya | cody-somerville: oh no, these people need face to face | 03:07 |
| anteaya | clarkb: that works for me too, personally I would enjoy that more | 03:08 |
| anteaya | but we do need some form of physical meetup for folks running these systems | 03:08 |
| clarkb | StevenK: so ya gerrit git prep does that for you | 03:08 |
| clarkb | it will checkout the proposed version of the change in the repo for you | 03:08 |
| anteaya | cody-somerville: if virtual would work for them, they could get stuff done on irc | 03:08 |
| anteaya | cody-somerville: it is pulling teeth to get them on irc | 03:09 |
| anteaya | they want face to face | 03:09 |
| wenlock | anteaya what stops them? | 03:09 |
| anteaya | habits? | 03:09 |
| anteaya | proprietary fears? | 03:09 |
| cody-somerville | anteaya: What about video conferencing? | 03:09 |
| anteaya | they don't know who is listening | 03:09 |
| StevenK | clarkb: So I should just call pip install --pre for those projects and it will DTRT? | 03:09 |
| anteaya | cody-somerville: that sets the wrong message | 03:09 |
| clarkb | StevenK: I think so? what is the job running in response to? | 03:10 |
| anteaya | since video conferencing is proprieatry software | 03:10 |
| clarkb | StevenK: changes to $client? if yes then yes | 03:10 |
| anteaya | also I can't do that with this group | 03:10 |
| *** KarenZhu has quit IRC | 03:10 | |
| anteaya | many of them struggle with english | 03:10 |
| anteaya | many of them have no knowledge of opensource habits of communication | 03:10 |
| anteaya | they have no idea there is anyone there | 03:10 |
| anteaya | they don't trust irc | 03:10 |
| anteaya | they need to know someone is listenting | 03:11 |
| StevenK | clarkb: TripleO jobs in general -- I was going to check $ZUUL_PROJECT for os-{cloud,apply}-config first | 03:11 |
| wenlock | <<< never used irc till mordred told me it existed | 03:11 |
| anteaya | to start they wouldn't even talk to each other | 03:11 |
| *** bdpayne has joined #openstack-infra | 03:11 | |
| anteaya | they all had to talk to me | 03:11 |
| clarkb | StevenK: oh then no that won't do what you want | 03:11 |
| anteaya | it was exhausting | 03:11 |
| clarkb | StevenK: gerrit git prep will check out the os-foo stuff for you | 03:11 |
| anteaya | the meetings are helping, they are starting to talk to each other | 03:11 |
| clarkb | StevenK: not the clients | 03:11 |
| anteaya | and help each other, which is awesome | 03:11 |
| anteaya | but it is all work | 03:11 |
| anteaya | and foreign to this group | 03:12 |
| wenlock | anteaya that sounds like my first month in guadaljara trying to get my team to learn social (irc things).... i think we're still learning | 03:12 |
| StevenK | anteaya: It's pushing them uphill in the snow, and exhausting? | 03:12 |
| *** praneshp has quit IRC | 03:12 | |
| anteaya | wenlock: yup | 03:12 |
| anteaya | StevenK: yes | 03:12 |
| anteaya | now the ones that have stuck with it are learning | 03:12 |
| anteaya | and getting past their fears | 03:12 |
| wenlock | thats the win | 03:13 |
| anteaya | and talking to other people without checking with legal if they can type in irc | 03:13 |
| StevenK | clarkb: So the code should be checked out on disk, and I can pip install <path> | 03:13 |
| anteaya | wenlock: very much so | 03:13 |
| wenlock | and it's easy to regress to cave man habbits | 03:13 |
| anteaya | right | 03:13 |
| clarkb | StevenK: right but that codes isn't the client code. it is your os-* code | 03:13 |
| anteaya | so we have made some progress | 03:13 |
| clarkb | StevenK: maybe that is what you need? | 03:13 |
| anteaya | but I know the group wants a meetup | 03:13 |
| StevenK | clarkb: Right, which it is | 03:13 |
| clarkb | StevenK: maybe you need to use tox | 03:13 |
| anteaya | and it is time in K to give them one | 03:13 |
| clarkb | StevenK: rathe rthan reinventing the wheel we already have sorted out | 03:13 |
| StevenK | clarkb: Is it a well-known path? | 03:13 |
| clarkb | StevenK: if you add a tox.ini we do all this stuff for you | 03:14 |
| clarkb | and tox installs . | 03:14 |
| anteaya | and I'm open to what ever we call it | 03:14 |
| anteaya | and I agree with clark it should be broader | 03:14 |
| anteaya | but we have to include some tutorials like how to install jenkins | 03:14 |
| StevenK | clarkb: http://git.openstack.org/cgit/openstack/tripleo-incubator/tree/scripts/setup-clienttools is what I'm hcking | 03:14 |
| StevenK | *hacking | 03:14 |
| anteaya | since i know that is what they want | 03:14 |
| wenlock | ci for dummies? | 03:14 |
| anteaya | one track, sure | 03:15 |
| anteaya | but I can't use that term | 03:15 |
| wenlock | yeah TM and all | 03:15 |
| anteaya | since they feel vulnerable just being here | 03:15 |
| anteaya | more than that | 03:15 |
| clarkb | StevenK: yup looks like reinvented tox (there may be a reason for that but thats what it looks like) | 03:15 |
| wenlock | ooh, haha | 03:15 |
| anteaya | it is counter productive to the goal | 03:15 |
| anteaya | I *want* them to feel welcome | 03:15 |
| anteaya | I *want* them to contribute | 03:15 |
| anteaya | and I want to select language that conveys that | 03:16 |
| wenlock | our first contribution not from our team came from EU | 03:16 |
| anteaya | since it will take them a while to trust I really mean that anyway | 03:16 |
| wenlock | on forj | 03:16 |
| anteaya | nice | 03:16 |
| wenlock | second one came from China | 03:16 |
| anteaya | cool | 03:16 |
| StevenK | clarkb: We want those things installed every time, no matter what we're testing -- I want to extend it to install the proposed changes of os-apply-config or os-cloud-config if we're testing that to make sure it is used | 03:16 |
| clarkb | StevenK: right so we have a test framework that will do that for you | 03:17 |
| clarkb | it will install the change made and all of its dependencies then run stuff | 03:17 |
| StevenK | Hmm, not sure if we run pip freeze during toci builds | 03:17 |
| clarkb | oh this is for toci? | 03:18 |
| clarkb | if this is for toci then you are on your own | 03:18 |
| clarkb | because ya'll intentionally didn't use our stuff :) | 03:18 |
| nibalizer | toci? | 03:18 |
| nibalizer | oh tripple o ci | 03:18 |
| nibalizer | StevenK: do you work on toci? | 03:18 |
| StevenK | nibalizer: I do, yes. | 03:18 |
| nibalizer | I'm supposed to be learning the way of the toci | 03:18 |
| StevenK | clarkb: I said TripleO jobs early on, sorry I wasn't more clear | 03:19 |
| clarkb | StevenK: you should look at devstack-gate's functions.sh | 03:19 |
| cody-somerville | nibalizer: Why do you need to learn toci?? | 03:19 |
| * anteaya puts away her soapbox and prepares to go offline for the night | 03:20 | |
| clarkb | StevenK: there is a setup workspace function in there that should show you what you need | 03:20 |
| *** jyuso has joined #openstack-infra | 03:21 | |
| *** salv-orlando has quit IRC | 03:21 | |
| clarkb | jhesketh: waiting on two more builds to snapshot | 03:22 |
| anteaya | SergeyLukjanov: can you check and ensure you fulfilled David Pineau's gerrit ci account request? he sent an email to the list last wednesday but wasn't subscribed so it waited until Friday before it posted | 03:24 |
| clarkb | jhesketh: ok images built now we need them to cycle in | 03:25 |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for pypi mirror infra bug 1334550 https://review.openstack.org/104432 | 03:28 |
| uvirtbot | Launchpad bug 1334550 in openstack-ci "Could not find any downloads that satisfy the requirement X" [Low,Triaged] https://launchpad.net/bugs/1334550 | 03:28 |
| ianw | clarkb: please tell the hp horizon people that after uploading an image, a small red box with "Error: unable to create new image." isn't very helpful | 03:28 |
| *** annegent_ has joined #openstack-infra | 03:28 | |
| ianw | clarkb: and may lead to destruction of keyboards :) | 03:28 |
| clarkb | jhesketh: looks like there are two bare-precise ready images and they are 0.00 hours old | 03:28 |
| jhesketh | clarkb: cool... looking at the queue it might take a while | 03:28 |
| jhesketh | oh unless we play roulette | 03:29 |
| clarkb | jhesketh: so I think good to check experimental | 03:29 |
| jhesketh | how many nodes in waiting though? | 03:29 |
| clarkb | oh right how long have they been building /me looks | 03:29 |
| clarkb | there are a few buidling but they are mostly new | 03:29 |
| clarkb | I think if you play roulette your odds are good | 03:29 |
| clarkb | ianw: you can't upload images by default | 03:30 |
| clarkb | ianw: unless something changed | 03:30 |
| clarkb | ianw: and iirc that is a glance client bug | 03:30 |
| clarkb | because it did that to me on the CLI too | 03:30 |
| clarkb | \o/ | 03:30 |
| ianw | clarkb: hmm, then removing the UI to upload an image would be even more helpful | 03:31 |
| *** dims has quit IRC | 03:32 | |
| *** mriedem has quit IRC | 03:34 | |
| clarkb | jhesketh: oh hey it succeeded | 03:36 |
| clarkb | jhesketh: but 404s the file | 03:36 |
| jhesketh | yeah, I'm just trying to find the jenkins log | 03:36 |
| clarkb | jhesketh: so more debugging but progress | 03:36 |
| jhesketh | clarkb: https://jenkins05.openstack.org/job/experimental-swift-logs-config-pep8/8/console | 03:36 |
| jhesketh | this looks like it | 03:36 |
| clarkb | jhesketh: https://jenkins05.openstack.org/job/experimental-swift-logs-config-pep8/8/console ya that | 03:36 |
| jhesketh | now to figure out where in the world it went to | 03:36 |
| jhesketh | clarkb: actually, do you have access to the cloud files and could just see if it was uploaded? | 03:37 |
| clarkb | jhesketh: I suppose I should go look at that container directly? | 03:41 |
| jhesketh | clarkb: if you don't mind :-) | 03:41 |
| * clarkb logs into things | 03:41 | |
| jhesketh | clarkb: I'm also not sure if we configured os-loganlayze | 03:42 |
| jhesketh | so I'm checking that out | 03:42 |
| *** sweston has quit IRC | 03:42 | |
| jhesketh | it doesn't look like it, so that's probably why it's not being served u | 03:42 |
| jhesketh | *up | 03:42 |
| *** Ryan_Lane has joined #openstack-infra | 03:43 | |
| *** crc32 has joined #openstack-infra | 03:44 | |
| *** dimtruck is now known as zz_dimtruck | 03:44 | |
| clarkb | jhesketh: 16/104416/1/experimental/experimental-swift-logs-config-pep8/19640d0e1a0d4dbb99b568e6f0d6fa48consoleText is the path of the file in the container | 03:44 |
| clarkb | jhesketh: so I think we may not be getting that path 100% correct | 03:45 |
| jhesketh | hmm, missing a trailing slash | 03:45 |
| jhesketh | well after the folder | 03:45 |
| jhesketh | but successful push? that's something :-) | 03:45 |
| *** bdpayne has quit IRC | 03:45 | |
| openstackgerrit | Joshua Harlow proposed a change to openstack-infra/config: Add templated uploading jobs for doc8 project https://review.openstack.org/96631 | 03:45 |
| clarkb | jhesketh: yup :) also the we shroten the uuid | 03:46 |
| clarkb | jhesketh: on the logs.o.o side | 03:46 |
| clarkb | jhesketh: not sure that that matters | 03:46 |
| jhesketh | the return url needs fixing, yes | 03:47 |
| *** zhiyan_ is now known as zhiyan | 03:47 | |
| clarkb | but thats cool it worked | 03:51 |
| clarkb | I think I am going to call it an evening. Hoepfulyl that gives you enough to work with | 03:52 |
| *** nosnos has joined #openstack-infra | 03:52 | |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/config: Add in wsgi.conf for os-loganalyze https://review.openstack.org/104437 | 03:56 |
| jhesketh | clarkb: ^ | 03:56 |
| jhesketh | clarkb: oh, you're off.. no worries, I still have more to poke at and fix. Thanks heaps for your help! :-) | 03:57 |
| clarkb | jhesketh: no problem | 03:57 |
| clarkb | jhesketh: though quick question | 03:57 |
| *** dstanek is now known as dstanek_zzz | 03:57 | |
| clarkb | jhesketh: were we going to use a ro account for that? | 03:57 |
| jhesketh | clarkb: we probably should, but since I didn't have hiera keys etc I just reused for the moment | 03:58 |
| openstackgerrit | Ian Wienand proposed a change to openstack-infra/config: Revert "demote f20 to experimental" https://review.openstack.org/104438 | 03:58 |
| clarkb | jhesketh: ok | 03:58 |
| jhesketh | it seems reasonably harmless since we have a write account anyway | 03:58 |
| clarkb | ya I am not super worried about it especailly since this is still early stages | 03:58 |
| clarkb | jhesketh: I assume there will be a companion change to fix the path for the file in the container too? | 03:59 |
| *** annegent_ has quit IRC | 04:06 | |
| zaro | jhesketh, clarkb : help push a change along? https://review.openstack.org/#/c/94675/ | 04:06 |
| jhesketh | clarkb: yep, will be, but I've got to run for the moment so it'll be an hour or so | 04:09 |
| jhesketh | zaro: will do when I return | 04:09 |
| *** annegent_ has joined #openstack-infra | 04:09 | |
| *** zehicle_at_dell has quit IRC | 04:11 | |
| *** zehicle_at_dell has joined #openstack-infra | 04:11 | |
| *** jhesketh has quit IRC | 04:14 | |
| *** yongli has quit IRC | 04:21 | |
| *** jdahlin has quit IRC | 04:27 | |
| *** yfried has quit IRC | 04:27 | |
| *** blamar has quit IRC | 04:28 | |
| *** bdpayne has joined #openstack-infra | 04:28 | |
| notmyname | I'm trying to push patches to a backport, but jenkins is telling me it doesn't know the "proposed/2.0.0" branch name | 04:32 |
| notmyname | what do I do now? | 04:32 |
| openstackgerrit | Adrian Otto proposed a change to openstack-infra/config: Add solum-infra-guestagent repo to stackforge https://review.openstack.org/99705 | 04:32 |
| clarkb | notmyname: link? | 04:35 |
| notmyname | clarkb: https://jenkins07.openstack.org/job/check-tempest-dsvm-full/11846/ | 04:35 |
| notmyname | clarkb: https://review.openstack.org/#/c/104440/ | 04:36 |
| notmyname | clarkb: there are 3 total. that's the first one | 04:36 |
| clarkb | argle bargle | 04:36 |
| clarkb | its a test matrix thing again | 04:36 |
| notmyname | clarkb: how do you mean? | 04:37 |
| clarkb | notmyname: https://git.openstack.org/cgit/openstack-infra/devstack-gate/tree/test-matrix.py that script determines what services you are tested with during integration testing based on branch and other things | 04:38 |
| notmyname | clarkb: ah. so the same reason we had a feature/foo branch. but now the test matrix doesn't know about proposed/bar yet | 04:39 |
| notmyname | ? | 04:39 |
| *** talluri has joined #openstack-infra | 04:39 | |
| clarkb | ya | 04:39 |
| clarkb | and prposed is trickier because for most projects I think they would just do the stable thing maybe? | 04:40 |
| openstackgerrit | Adrian Otto proposed a change to openstack-infra/config: Add solum-infra-guestagent repo to stackforge https://review.openstack.org/99705 | 04:40 |
| clarkb | but swift doesn't do the stable/releasename thing | 04:40 |
| clarkb | or does it? | 04:40 |
| notmyname | I think this is new because we don't do milestone-proposed any more | 04:40 |
| clarkb | yup | 04:40 |
| clarkb | and we want to test propsed/foo against other proposed/foos. I think we want to basically treat these like stable/foo | 04:41 |
| notmyname | clarkb: and FWIW, we do have stable/* branches in swift | 04:41 |
| clarkb | and that way stable/foo is ready to go when we get ther | 04:41 |
| clarkb | notmyname: but you don't have them for every intermediate release iirc | 04:41 |
| *** pcrews has quit IRC | 04:42 | |
| notmyname | seeing as this is the first ever time it's been used, no | 04:42 |
| clarkb | well you didn't when you did them in the past too | 04:42 |
| notmyname | we did milestone-proposed then | 04:42 |
| clarkb | iirc you did two point releases per cycle, but only the second was branched off on stable/foo | 04:42 |
| notmyname | no, that's not quite right | 04:43 |
| *** annegent_ has quit IRC | 04:43 | |
| clarkb | this mostly worked before because milestone-proposed is/was a collapsed namespace | 04:43 |
| notmyname | regardless of what used to happen with m-p, now we've got an rc1 for this intermediate release, and we need to put some patches on it for rc2. | 04:44 |
| clarkb | sure so, I think the general case is easy. we remove the proposed/ prefix and treat things as eg juno | 04:45 |
| clarkb | that way we are testing the stable machinery before we are stable which is cool and one of the reasons I argued for an approach like this | 04:45 |
| clarkb | but swift isn't proposing stable today | 04:46 |
| clarkb | and that is where I am somewhat lost in how to deal with it | 04:46 |
| notmyname | correct. the stable releases are from the openstack integrated releases | 04:46 |
| clarkb | I think we want to treat this proposed release as master | 04:46 |
| clarkb | and other proposed releases as juno | 04:46 |
| notmyname | I don't follow what that means | 04:47 |
| clarkb | test-matrix is figuring out what setup to test a change against | 04:47 |
| clarkb | proposed/juno should be tested against a juno setup | 04:47 |
| clarkb | proposed/2.0.0 should probably be tested against a master setup beacuse it has no stable release equivalent | 04:48 |
| notmyname | ah. I see. also testing against icehouse (ie the most recent stable) makes sense to me too. either way is fine with me | 04:48 |
| *** devkulkarni has joined #openstack-infra | 04:48 | |
| clarkb | but I am not 100% on that and I am also not confident that a change I would write to test-matrix.py right now wouldn't come with all sorts of corner case breakage | 04:49 |
| *** blamar has joined #openstack-infra | 04:50 | |
| clarkb | I think the naive approach is to split on / and if string_list[1][0] is not in [a-z] then master else string_list[1] | 04:52 |
| clarkb | which seems like a huge hack | 04:52 |
| *** annegent_ has joined #openstack-infra | 04:52 | |
| *** jhesketh has joined #openstack-infra | 04:52 | |
| clarkb | but maybe it is good enough. I can throw that up and people can discuss it | 04:55 |
| *** Longgeek has joined #openstack-infra | 04:55 | |
| notmyname | clarkb: what I'm hearing is that these patches won't be backported until early next week then | 04:55 |
| *** amitpp has joined #openstack-infra | 04:55 | |
| clarkb | notmyname: I think we can get it through early tomorrow | 04:56 |
| clarkb | notmyname: mostly thinking of sdague and ttx's feedback here | 04:56 |
| * notmyname is frustrated | 04:58 | |
| notmyname | clarkb: to be clear, I'm not blaming you (or anyone) at all | 04:58 |
| notmyname | I've been trying to get these patches backported for 2 days, and jenkins + timezones have really slowed it down. and this is the end of the rc period and dates matter this time | 04:59 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 05:00 |
| clarkb | something like that | 05:00 |
| clarkb | notmyname: so we are always going to run into hiccups when trying something new | 05:01 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/config: Add in wsgi.conf for os-loganalyze https://review.openstack.org/104437 | 05:01 |
| *** nati_ueno has joined #openstack-infra | 05:02 | |
| notmyname | clarkb: ok, thanks. I should check back in the morning about status of it all? | 05:03 |
| *** bdpayne has quit IRC | 05:03 | |
| clarkb | notmyname: ya I think so | 05:03 |
| *** shashankhegde has quit IRC | 05:03 | |
| clarkb | maybe if ttx shows up early we can push 104448 earlyish | 05:03 |
| clarkb | jhesketh: thoughts on 104448? | 05:03 |
| *** blamar has quit IRC | 05:05 | |
| notmyname | FWIW, the end result is supposed to be https://github.com/notmyname/swift/commits/20bp | 05:05 |
| *** UtahDave has joined #openstack-infra | 05:05 | |
| jhesketh | clarkb: feedback inline | 05:07 |
| *** annegent_ has quit IRC | 05:07 | |
| clarkb | gah thanks clearly I am rushing | 05:07 |
| *** hogepodge has quit IRC | 05:07 | |
| clarkb | jhesketh: the isalpha thing is explained in the comment below and the commit message | 05:07 |
| *** hogepodge has joined #openstack-infra | 05:07 | |
| *** gyee has quit IRC | 05:09 | |
| jhesketh | oh right, got it... I was thinking it was related to an alpha-release not remembering that it's a builtin | 05:09 |
| jhesketh | clarkb: ^ | 05:09 |
| clarkb | ah | 05:10 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 05:10 |
| clarkb | that should do it | 05:10 |
| *** pcrews has joined #openstack-infra | 05:11 | |
| *** otherwiseguy has quit IRC | 05:13 | |
| jhesketh | clarkb: so how do you expect intermediate releases to be formatted... your example was 'swift' but I don't follow | 05:13 |
| *** maxbit has joined #openstack-infra | 05:14 | |
| clarkb | jhesketh: notmyname is working on proposed/2.0.0 | 05:14 |
| clarkb | jhesketh: because swift does mid cycle releases not tied to eg juno | 05:14 |
| clarkb | jhesketh: my biggest concern is I think swift may do one of those type of releases at juno release too | 05:14 |
| clarkb | and then my change will break | 05:14 |
| notmyname | yes, I think we've done as many as 4 in an openstack cycle before | 05:15 |
| clarkb | notmyname: so when you get to juno release will you have something like proposed/juno or proposed/2.x.x or both? | 05:16 |
| notmyname | and the last one is timed to match the integrated cycle. normally it's rare to have backports (well, less than half the time), but this is a big release | 05:16 |
| notmyname | clarkb: I don't know. we've never used anything but milestone-proposed before. I think this is new for everyone | 05:17 |
| *** wenlock has quit IRC | 05:17 | |
| clarkb | notmyname: in milestone proposed world you cut a stable/icehouse branch? | 05:17 |
| notmyname | clarkb: we should have a 2.x.y release for juno. eg icehouse had swift 1.13.1 | 05:17 |
| notmyname | after the release, yes | 05:17 |
| clarkb | notmyname: so old process was milestone-proposed until ready, tag 1.13.1 then create stable/icehouse branch? | 05:18 |
| notmyname | we did the milestone-proposed before the release (and it had a 1.13.1 version number in icehouse) | 05:18 |
| notmyname | clarkb: yes, that's it exactly as I understand it | 05:18 |
| jhesketh | clarkb: okay, makes sense to me | 05:19 |
| notmyname | clarkb: and the mid-cycle releases were the same except that there wasnt' a stable/* branch | 05:19 |
| jhesketh | only +1 though because I don't know much about how the test-matrix stuff works | 05:19 |
| clarkb | notmyname: ok in that case I think my change will mostly sort of work. there may be some interaction I am not considering though | 05:19 |
| *** basha has joined #openstack-infra | 05:19 | |
| clarkb | notmyname: the trickiest bit is going to be around juno release time | 05:21 |
| clarkb | notmyname: because we want to keep all the projects in sync and tested against each other with their proposed integrated release | 05:21 |
| notmyname | clarkb: right. seems that at that point everyone will have a proposed/* branch | 05:21 |
| notmyname | so you're concerned about what the "*" is, right? so we can test the right thing with the right thing? | 05:22 |
| clarkb | yup | 05:22 |
| clarkb | correct | 05:22 |
| notmyname | hmm...proposed/juno-2.0.0 seems reasonable to me for swift, at first glance. I don't really care too much about branch names. I care a _lot_ more about the version string in the code (eg 2.0.0) | 05:23 |
| notmyname | heh. "version string in the code". that's the other think I gotta get to at some point | 05:24 |
| notmyname | *sigh* so much to do | 05:24 |
| *** ihrachyshka has joined #openstack-infra | 05:26 | |
| *** yfried has joined #openstack-infra | 05:26 | |
| notmyname | clarkb: thanks. I'm turning in for the night. I'll check back in the morning | 05:26 |
| clarkb | ok | 05:27 |
| *** sabeen has joined #openstack-infra | 05:27 | |
| *** sabeen has quit IRC | 05:27 | |
| clarkb | ttx: sdague mordred jeblair and anyone else interested. test-matrix.py needs to be taught about proposed/* branches. https://review.openstack.org/104448 is a first stab at that. The current worry is whether or not we are treating branches like proposed/2.0.0 for swift correct as well as branches like proposed/juno for all the things | 05:27 |
| *** pcrews has quit IRC | 05:28 | |
| clarkb | we can probably limp along on my change to keep notmyname moving, but may need more thought before juno release time | 05:28 |
| *** marun_afk has quit IRC | 05:29 | |
| *** gokrokve has quit IRC | 05:30 | |
| *** bdpayne has joined #openstack-infra | 05:33 | |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/config: Ensure there is a slash in the object prefix https://review.openstack.org/104451 | 05:35 |
| clarkb | jhesketh: is that a job for os.path.join? | 05:39 |
| *** otter768 has quit IRC | 05:39 | |
| *** ihrachyshka has quit IRC | 05:39 | |
| jhesketh | clarkb: no because I'm not appending the tailing slash to the url | 05:39 |
| jhesketh | it's changing the filename to upload to be '/filename.txt' | 05:39 |
| clarkb | oh where does the prefix come in? is that set by the hmac and all that? | 05:40 |
| *** praneshp has joined #openstack-infra | 05:41 | |
| jhesketh | clarkb: the url is set by hmac | 05:41 |
| jhesketh | I can't set it there (I did look) because it's basically swift_destination + object_prefix | 05:41 |
| jhesketh | where object_prefix is LOG_PATH | 05:41 |
| clarkb | ya ok | 05:41 |
| clarkb | so you have to make sure that filename handles it | 05:41 |
| jhesketh | and LOG_PATH is consumed by Jenkins (and others?) expecting no trailing slash | 05:41 |
| jhesketh | clarkb: hmm, that's a question for the requests library and maybe swift | 05:43 |
| jhesketh | I don't have a test env available atm so I can't check :-( | 05:43 |
| jhesketh | do you think it's likely to work or shall I WIP it until I know | 05:44 |
| clarkb | maybe wip it? I do think requests/swift should handle it for us if possible | 05:44 |
| clarkb | because we are essential saying upload file to dir foo | 05:44 |
| clarkb | that should result in foo/file | 05:44 |
| clarkb | imo | 05:45 |
| clarkb | but swift is an object store not a file system | 05:45 |
| *** Longgeek has quit IRC | 05:45 | |
| *** ihrachyshka has joined #openstack-infra | 05:45 | |
| *** Longgeek has joined #openstack-infra | 05:45 | |
| clarkb | I am fine with what you currently have though | 05:46 |
| clarkb | +2 from me now and we can sort out how to handle it in the future | 05:46 |
| clarkb | because i need sleep soon | 05:46 |
| jhesketh | well it's not going to break anything that's working so maybe we just try it out and see if it fails | 05:46 |
| * jhesketh is being lazy though | 05:47 | |
| clarkb | that is fine with me | 05:47 |
| *** sweston has joined #openstack-infra | 05:47 | |
| clarkb | though I won't be around to rebuild images and check the results in the container until tomorrow | 05:47 |
| *** harlowja is now known as harlowja_away | 05:47 | |
| clarkb | so maybe we wait until then to merge anyways? | 05:47 |
| jhesketh | yep I'm good with that | 05:48 |
| jhesketh | clarkb: I'm about to push a change that needs a zuul restart anyway | 05:48 |
| jhesketh | although that won't stop us from checking it works, it'll just make the link left in gerrit correct | 05:48 |
| jhesketh | actually the only thing we need to check if it works is the credentials for os-loganalyze | 05:48 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/zuul: Fix LOG_PATH to use 7char uuid https://review.openstack.org/104453 | 05:49 |
| *** Longgeek has quit IRC | 05:50 | |
| *** bdpayne has quit IRC | 05:55 | |
| *** rcarrill` has joined #openstack-infra | 05:56 | |
| *** ihrachyshka has quit IRC | 05:56 | |
| *** otter768 has joined #openstack-infra | 05:57 | |
| *** rcarrillocruz has quit IRC | 05:57 | |
| *** doude has joined #openstack-infra | 05:58 | |
| *** basha has quit IRC | 05:59 | |
| *** shashankhegde has joined #openstack-infra | 06:00 | |
| *** gokrokve has joined #openstack-infra | 06:00 | |
| *** maxbit has quit IRC | 06:01 | |
| jhesketh | clarkb: so mattoliverau knows a bit about swift and the multipart-post middleware and I've done some brief reading on multipart data and it seems like we should be able to do that | 06:02 |
| jhesketh | ie my change | 06:03 |
| *** talluri has quit IRC | 06:03 | |
| jhesketh | perhaps I should emphasis the /should/ | 06:05 |
| *** ianw has quit IRC | 06:05 | |
| clarkb | cool | 06:05 |
| *** ianw has joined #openstack-infra | 06:05 | |
| *** basha has joined #openstack-infra | 06:08 | |
| *** mrmartin has joined #openstack-infra | 06:09 | |
| *** otter768 has quit IRC | 06:10 | |
| *** jhesketh has quit IRC | 06:11 | |
| *** baohua has joined #openstack-infra | 06:12 | |
| *** gokrokve has quit IRC | 06:16 | |
| *** jlibosva has joined #openstack-infra | 06:18 | |
| *** reed has quit IRC | 06:19 | |
| *** nati_ueno has quit IRC | 06:20 | |
| *** mrmartin has quit IRC | 06:24 | |
| *** chandan_kumar has joined #openstack-infra | 06:25 | |
| *** _nadya_ has joined #openstack-infra | 06:26 | |
| *** shashankhegde has quit IRC | 06:27 | |
| *** Longgeek has joined #openstack-infra | 06:27 | |
| *** _nadya_ has quit IRC | 06:27 | |
| *** camunoz has quit IRC | 06:29 | |
| *** jcoufal has joined #openstack-infra | 06:31 | |
| *** ildikov has joined #openstack-infra | 06:37 | |
| *** devkulkarni has quit IRC | 06:37 | |
| *** tkelsey has joined #openstack-infra | 06:38 | |
| *** penguinRaider has quit IRC | 06:41 | |
| *** rcarrillocruz has joined #openstack-infra | 06:42 | |
| *** rcarrill` has quit IRC | 06:44 | |
| *** unicell has quit IRC | 06:51 | |
| *** denis_makogon has joined #openstack-infra | 06:53 | |
| *** shashankhegde has joined #openstack-infra | 06:53 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Cleanup manual jobs https://review.openstack.org/103646 | 06:53 |
| *** e0ne has joined #openstack-infra | 06:57 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 06:58 | |
| *** e0ne has quit IRC | 07:01 | |
| *** jhesketh_ has joined #openstack-infra | 07:04 | |
| *** talluri has joined #openstack-infra | 07:04 | |
| *** e0ne has joined #openstack-infra | 07:05 | |
| *** e0ne has quit IRC | 07:07 | |
| *** _nadya_ has joined #openstack-infra | 07:08 | |
| *** talluri has quit IRC | 07:08 | |
| *** andreykurilin_ has joined #openstack-infra | 07:11 | |
| *** e0ne has joined #openstack-infra | 07:11 | |
| *** shashankhegde has quit IRC | 07:11 | |
| *** crc32 has quit IRC | 07:12 | |
| *** gokrokve has joined #openstack-infra | 07:12 | |
| *** HACKING-FACEBOOK has quit IRC | 07:14 | |
| *** e0ne has quit IRC | 07:15 | |
| *** mrmartin has joined #openstack-infra | 07:15 | |
| *** afazekas_ has joined #openstack-infra | 07:15 | |
| *** gokrokve has quit IRC | 07:17 | |
| *** basha has quit IRC | 07:18 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 07:18 | |
| *** e0ne has joined #openstack-infra | 07:19 | |
| *** ihrachyshka has joined #openstack-infra | 07:19 | |
| *** nati_ueno has joined #openstack-infra | 07:20 | |
| *** e0ne has quit IRC | 07:20 | |
| *** flaper87|afk is now known as flaper87 | 07:24 | |
| *** e0ne has joined #openstack-infra | 07:24 | |
| *** dizquierdo has joined #openstack-infra | 07:25 | |
| mattoliverau | clarkb: jhesketh is not here, but changing the filename to start with a '/' works.. as does adding a '/' to the end of the object_prefix. (works on my SAIO environment). Can't test it in curl (filenames with a '/', but via python and the requests package (like inzuul_swift_upload.py) worked successfully. | 07:27 |
| *** markmc has joined #openstack-infra | 07:27 | |
| *** e0ne has quit IRC | 07:29 | |
| *** basha has joined #openstack-infra | 07:31 | |
| *** e0ne has joined #openstack-infra | 07:31 | |
| *** HACKING-FACEBOOK has quit IRC | 07:31 | |
| *** nati_ueno has quit IRC | 07:33 | |
| *** jgallard has joined #openstack-infra | 07:33 | |
| *** e0ne has quit IRC | 07:35 | |
| *** mmaglana has quit IRC | 07:35 | |
| *** mmaglana has joined #openstack-infra | 07:35 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 07:38 | |
| *** e0ne has joined #openstack-infra | 07:39 | |
| openstackgerrit | afazekas proposed a change to openstack-infra/elastic-recheck: Add query for pypi index fetching issues https://review.openstack.org/104298 | 07:40 |
| *** mmaglana has quit IRC | 07:40 | |
| *** talluri has joined #openstack-infra | 07:41 | |
| *** Hal_ has joined #openstack-infra | 07:41 | |
| *** ildikov_ has joined #openstack-infra | 07:42 | |
| *** e0ne has quit IRC | 07:44 | |
| *** ildikov has quit IRC | 07:45 | |
| *** talluri has quit IRC | 07:45 | |
| *** HACKING-FACEBOOK has quit IRC | 07:48 | |
| *** e0ne has joined #openstack-infra | 07:48 | |
| SergeyLukjanov | anteaya, sure, I'll check requests today | 07:52 |
| *** andreykurilin_ has quit IRC | 07:53 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 07:53 | |
| *** roeyc has joined #openstack-infra | 07:55 | |
| *** e0ne has quit IRC | 07:55 | |
| *** yaguang has quit IRC | 07:55 | |
| *** basha has quit IRC | 07:57 | |
| *** hashar has joined #openstack-infra | 07:59 | |
| *** jpich has joined #openstack-infra | 07:59 | |
| *** e0ne has joined #openstack-infra | 07:59 | |
| *** asettle has quit IRC | 08:01 | |
| *** HACKING-FACEBOOK has quit IRC | 08:03 | |
| *** nosnos has quit IRC | 08:05 | |
| *** salv-orlando has joined #openstack-infra | 08:05 | |
| *** nosnos has joined #openstack-infra | 08:05 | |
| *** afazekas_ is now known as afzekas | 08:05 | |
| *** mrmartin has quit IRC | 08:05 | |
| *** derekh_ has joined #openstack-infra | 08:05 | |
| *** afzekas is now known as afazekas_ | 08:07 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 08:08 | |
| *** e0ne has quit IRC | 08:11 | |
| *** gokrokve has joined #openstack-infra | 08:13 | |
| *** guangjun has quit IRC | 08:14 | |
| *** e0ne has joined #openstack-infra | 08:15 | |
| *** HACKING-FACEBOOK has quit IRC | 08:18 | |
| *** gokrokve has quit IRC | 08:18 | |
| openstackgerrit | A change was merged to stackforge/python-jenkins: Add a gitreview file https://review.openstack.org/104261 | 08:19 |
| *** pblaho has joined #openstack-infra | 08:20 | |
| *** _nadya_ has quit IRC | 08:22 | |
| *** jp_at_hp has joined #openstack-infra | 08:22 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 08:25 | |
| pleia2 | woo translations meeting | 08:25 |
| * pleia2 sleep & | 08:26 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 08:27 | |
| *** jgallard has quit IRC | 08:28 | |
| *** gokrokve has joined #openstack-infra | 08:31 | |
| *** jgallard has joined #openstack-infra | 08:34 | |
| *** praneshp has quit IRC | 08:34 | |
| *** talluri has joined #openstack-infra | 08:34 | |
| *** HACKING-FACEBOOK has quit IRC | 08:35 | |
| *** gokrokve has quit IRC | 08:35 | |
| *** denis_makogon has quit IRC | 08:37 | |
| *** mrda is now known as mrda-away | 08:39 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 08:39 | |
| *** jcoufal has quit IRC | 08:41 | |
| *** jcoufal has joined #openstack-infra | 08:44 | |
| *** e0ne has quit IRC | 08:47 | |
| *** basha has joined #openstack-infra | 08:48 | |
| *** e0ne has joined #openstack-infra | 08:51 | |
| *** pelix has joined #openstack-infra | 08:55 | |
| *** Ryan_Lane has quit IRC | 09:04 | |
| *** Longgeek has quit IRC | 09:04 | |
| SergeyLukjanov | anteaya, I've created 3 accounts for CI today | 09:12 |
| *** yaguang has joined #openstack-infra | 09:16 | |
| *** rcarrill` has joined #openstack-infra | 09:17 | |
| *** e0ne has quit IRC | 09:19 | |
| *** rcarrillocruz has quit IRC | 09:19 | |
| *** e0ne has joined #openstack-infra | 09:23 | |
| *** e0ne has quit IRC | 09:27 | |
| s0nea | thank you for the openattic service account SergeyLukjanov :) | 09:28 |
| *** penguinRaider has joined #openstack-infra | 09:29 | |
| *** gokrokve has joined #openstack-infra | 09:31 | |
| *** rcarrillocruz has joined #openstack-infra | 09:33 | |
| *** yaguang has quit IRC | 09:33 | |
| *** rcarrill` has quit IRC | 09:35 | |
| *** mmaglana has joined #openstack-infra | 09:36 | |
| *** gokrokve has quit IRC | 09:36 | |
| *** e0ne has joined #openstack-infra | 09:37 | |
| *** hashar has quit IRC | 09:38 | |
| *** dmakogon_ is now known as denis_makogon | 09:38 | |
| *** mmaglana has quit IRC | 09:40 | |
| *** e0ne has quit IRC | 09:43 | |
| *** baohua has quit IRC | 09:46 | |
| *** yaguang has joined #openstack-infra | 09:47 | |
| *** yaguang_ has joined #openstack-infra | 09:54 | |
| *** yaguang has quit IRC | 09:55 | |
| *** talluri has quit IRC | 09:58 | |
| *** talluri has joined #openstack-infra | 09:58 | |
| *** jgallard has quit IRC | 10:05 | |
| *** dstanek_zzz is now known as dstanek | 10:14 | |
| *** pblaho has quit IRC | 10:21 | |
| *** hashar has joined #openstack-infra | 10:22 | |
| *** marun has joined #openstack-infra | 10:28 | |
| *** gokrokve has joined #openstack-infra | 10:32 | |
| *** cnesa2 has joined #openstack-infra | 10:37 | |
| *** gokrokve has quit IRC | 10:37 | |
| openstackgerrit | yolanda.robla proposed a change to openstack-infra/nodepool: Build images using diskimage-builder https://review.openstack.org/46482 | 10:39 |
| *** HACKING-FACEBOOK has quit IRC | 10:39 | |
| *** basha has quit IRC | 10:39 | |
| *** akuznetsova has left #openstack-infra | 10:42 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 10:43 | |
| *** e0ne has joined #openstack-infra | 10:45 | |
| *** cnesa2 has quit IRC | 10:46 | |
| *** roeyc has quit IRC | 10:50 | |
| *** e0ne has quit IRC | 10:50 | |
| *** Longgeek has joined #openstack-infra | 10:52 | |
| *** e0ne has joined #openstack-infra | 10:54 | |
| openstackgerrit | Radoslav Gerganov proposed a change to openstack/requirements: Update oslo.vmware to version 0.4 https://review.openstack.org/104514 | 10:55 |
| *** HACKING-FACEBOOK has quit IRC | 10:55 | |
| *** rcarrillocruz has quit IRC | 10:57 | |
| *** gnuoy has quit IRC | 10:57 | |
| *** jpeeler has quit IRC | 10:57 | |
| *** derekh_ has quit IRC | 10:59 | |
| *** rcarrillocruz has joined #openstack-infra | 10:59 | |
| *** gnuoy has joined #openstack-infra | 11:00 | |
| *** derekh_ has joined #openstack-infra | 11:00 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 11:00 | |
| *** HACKING-FACEBOOK has quit IRC | 11:00 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 11:02 | |
| *** e0ne has quit IRC | 11:02 | |
| *** HACKING-FACEBOOK has quit IRC | 11:02 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 11:04 | |
| *** e0ne has joined #openstack-infra | 11:06 | |
| *** ominakov has joined #openstack-infra | 11:07 | |
| *** rcarrill` has joined #openstack-infra | 11:12 | |
| *** rcarrillocruz has quit IRC | 11:14 | |
| *** ociuhandu has quit IRC | 11:15 | |
| openstackgerrit | Marton Kiss proposed a change to openstack-infra/config: Add groups.openstack.org instance https://review.openstack.org/104332 | 11:15 |
| *** HACKING-FACEBOOK has quit IRC | 11:15 | |
| *** e0ne has quit IRC | 11:19 | |
| *** yaguang_ has quit IRC | 11:19 | |
| *** mugsie has joined #openstack-infra | 11:20 | |
| *** _nadya_ has joined #openstack-infra | 11:22 | |
| *** Longgeek has quit IRC | 11:22 | |
| *** Longgeek has joined #openstack-infra | 11:23 | |
| *** e0ne has joined #openstack-infra | 11:23 | |
| *** e0ne has quit IRC | 11:25 | |
| *** tcammann_ has quit IRC | 11:26 | |
| *** _nadya_ has quit IRC | 11:26 | |
| *** e0ne has joined #openstack-infra | 11:29 | |
| *** rcarrillocruz has joined #openstack-infra | 11:29 | |
| *** jcoufal has quit IRC | 11:30 | |
| *** jcoufal has joined #openstack-infra | 11:30 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 11:31 | |
| *** HACKING-FACEBOOK has quit IRC | 11:31 | |
| *** _nadya_ has joined #openstack-infra | 11:31 | |
| *** rcarrill` has quit IRC | 11:32 | |
| *** gokrokve has joined #openstack-infra | 11:33 | |
| *** e0ne has quit IRC | 11:33 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 11:33 | |
| *** penguinRaider has quit IRC | 11:34 | |
| *** _nadya_ has quit IRC | 11:36 | |
| *** gokrokve has quit IRC | 11:37 | |
| *** e0ne has joined #openstack-infra | 11:37 | |
| *** skolekonov has joined #openstack-infra | 11:42 | |
| *** HACKING-FACEBOOK has quit IRC | 11:42 | |
| *** radez_g0n3 is now known as radez | 11:43 | |
| *** Longgeek_ has joined #openstack-infra | 11:43 | |
| *** mbacchi has joined #openstack-infra | 11:44 | |
| *** Longgeek has quit IRC | 11:47 | |
| *** penguinRaider has joined #openstack-infra | 11:47 | |
| *** liam_ has joined #openstack-infra | 11:49 | |
| *** liam_ is now known as Guest45753 | 11:49 | |
| *** Guest45753 has quit IRC | 11:50 | |
| *** mkerrin has quit IRC | 11:51 | |
| openstackgerrit | Julien Vey proposed a change to openstack-infra/config: Add solum-infra-guestagent repo to stackforge https://review.openstack.org/99705 | 11:52 |
| *** e0ne has quit IRC | 11:54 | |
| *** nosnos has quit IRC | 11:55 | |
| *** dprince has joined #openstack-infra | 11:55 | |
| *** e0ne has joined #openstack-infra | 11:56 | |
| *** nosnos_ has joined #openstack-infra | 11:58 | |
| *** yfried_ has joined #openstack-infra | 11:58 | |
| *** yfried has quit IRC | 11:59 | |
| *** mkerrin has joined #openstack-infra | 11:59 | |
| *** yfried__ has joined #openstack-infra | 12:00 | |
| *** ociuhandu has joined #openstack-infra | 12:00 | |
| *** yfried has joined #openstack-infra | 12:01 | |
| *** jcoufal has quit IRC | 12:01 | |
| *** ociuhandu_ has joined #openstack-infra | 12:02 | |
| *** yfried_ has quit IRC | 12:03 | |
| *** Longgeek_ has quit IRC | 12:04 | |
| *** jcoufal has joined #openstack-infra | 12:04 | |
| *** yfried__ has quit IRC | 12:04 | |
| *** dims has joined #openstack-infra | 12:04 | |
| *** ociuhandu has quit IRC | 12:04 | |
| *** ociuhandu_ is now known as ociuhandu | 12:04 | |
| *** Longgeek has joined #openstack-infra | 12:05 | |
| *** hashar has quit IRC | 12:05 | |
| *** Longgeek_ has joined #openstack-infra | 12:06 | |
| *** afazekas_ has quit IRC | 12:07 | |
| *** Longgeek has quit IRC | 12:09 | |
| *** jyuso has quit IRC | 12:11 | |
| *** johnthetubaguy is now known as zz_johnthetubagu | 12:12 | |
| *** pblaho has joined #openstack-infra | 12:13 | |
| *** e0ne has quit IRC | 12:13 | |
| *** e0ne has joined #openstack-infra | 12:15 | |
| *** dkranz has joined #openstack-infra | 12:16 | |
| *** yamahata has joined #openstack-infra | 12:16 | |
| *** e0ne has quit IRC | 12:18 | |
| *** amotoki has joined #openstack-infra | 12:19 | |
| *** e0ne has joined #openstack-infra | 12:22 | |
| openstackgerrit | Sean Dague proposed a change to openstack-infra/elastic-recheck: make ER only go back 10 days https://review.openstack.org/104533 | 12:23 |
| *** jcoufal has quit IRC | 12:24 | |
| *** jcoufal has joined #openstack-infra | 12:24 | |
| *** afazekas_ has joined #openstack-infra | 12:24 | |
| *** rfolco has joined #openstack-infra | 12:24 | |
| *** dtantsur has quit IRC | 12:24 | |
| *** dizquierdo has quit IRC | 12:24 | |
| *** dtantsur has joined #openstack-infra | 12:24 | |
| *** weshay has joined #openstack-infra | 12:26 | |
| anteaya | SergeyLukjanov: thank you for creating the accounts, and for updating me | 12:28 |
| *** e0ne has quit IRC | 12:31 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 12:32 | |
| *** e0ne has joined #openstack-infra | 12:33 | |
| *** gokrokve has joined #openstack-infra | 12:33 | |
| *** afazekas_ has quit IRC | 12:34 | |
| *** yfried_ has joined #openstack-infra | 12:35 | |
| *** fifieldt has quit IRC | 12:35 | |
| *** yfried__ has joined #openstack-infra | 12:35 | |
| *** yfried__ has quit IRC | 12:35 | |
| *** zigo has quit IRC | 12:37 | |
| *** zigo has joined #openstack-infra | 12:37 | |
| *** talluri has quit IRC | 12:38 | |
| *** gokrokve has quit IRC | 12:38 | |
| *** talluri has joined #openstack-infra | 12:39 | |
| *** yfried has quit IRC | 12:39 | |
| *** ArxCruz has joined #openstack-infra | 12:41 | |
| *** hashar has joined #openstack-infra | 12:41 | |
| *** hashar has quit IRC | 12:41 | |
| *** zigo has quit IRC | 12:41 | |
| *** zigo has joined #openstack-infra | 12:42 | |
| sdague | SergeyLukjanov / fungi - any idea what's up with - https://review.openstack.org/#/c/104010/ | 12:42 |
| *** hashar has joined #openstack-infra | 12:42 | |
| *** adalbas has joined #openstack-infra | 12:43 | |
| *** talluri has quit IRC | 12:43 | |
| SergeyLukjanov | anteaya, np | 12:44 |
| SergeyLukjanov | sdague, looking | 12:44 |
| sdague | it looks like it was completely lost by jenkins | 12:45 |
| *** zigo has quit IRC | 12:45 | |
| *** mtreinish has quit IRC | 12:45 | |
| SergeyLukjanov | sdague, yeah... looks like zuul forget about this changes, no ideas why | 12:46 |
| sdague | ok | 12:46 |
| *** afazekas_ has joined #openstack-infra | 12:47 | |
| *** rlandy has joined #openstack-infra | 12:48 | |
| *** jamielennox is now known as jamielennox|away | 12:50 | |
| *** Longgeek_ has quit IRC | 12:51 | |
| *** Longgeek has joined #openstack-infra | 12:52 | |
| *** yaguang has joined #openstack-infra | 12:56 | |
| *** miqui has joined #openstack-infra | 12:56 | |
| *** e0ne has quit IRC | 12:57 | |
| *** dims_ has joined #openstack-infra | 12:58 | |
| fungi | yeah, some time between 07-02 01:18 and now, it fell out of the gate queue. checking to see when we restarted zuul... | 12:59 |
| *** pbelanyi_ is now known as pbelanyi | 12:59 | |
| anteaya | sdague: at one point I was given a gerrit url for a dashboard you had created (I believe) used by a third party ci to assess their results in gerrit (the link broke in firefox and worked in chrome and was really long) I thought I had a copy of it somewhere and can't seem to find it, do you know what I'm talking about? | 13:00 |
| *** yfried__ has joined #openstack-infra | 13:00 | |
| *** rcarrillocruz has quit IRC | 13:01 | |
| *** e0ne has joined #openstack-infra | 13:02 | |
| *** dims has quit IRC | 13:02 | |
| *** rcarrillocruz has joined #openstack-infra | 13:02 | |
| *** yfried has joined #openstack-infra | 13:02 | |
| *** yfried_ has quit IRC | 13:02 | |
| *** e0ne has quit IRC | 13:04 | |
| *** yfried__ has quit IRC | 13:05 | |
| sdague | anteaya: yes, we've fixed the dashboard creator to work in firefox now | 13:06 |
| sdague | anteaya: https://github.com/stackforge/gerrit-dash-creator | 13:06 |
| sdague | I don't know the link in particular that you are discussing, however I do know that we've fixed the firefox url bug | 13:07 |
| *** e0ne has joined #openstack-infra | 13:08 | |
| fungi | last zuul restart was on july 1st, so prior to that... | 13:10 |
| fungi | 2014-07-02 02:29:03,457 INFO zuul.DependentPipelineManager: Dequeuing change <Change 0x7fdb594fab90 104010,1> because it can no longer merge | 13:11 |
| *** Longgeek_ has joined #openstack-infra | 13:12 | |
| fungi | sdague: that's the point at which its parent change, 104009, fell out of the gate ahead of it | 13:13 |
| anteaya | sdague: thanks I'll dig around and see if I can find the url | 13:13 |
| anteaya | and yay for fixing the firefox url bug! | 13:13 |
| *** yaguang has quit IRC | 13:13 | |
| sdague | fungi: ok, but the parent change is now merged | 13:13 |
| sdague | which I expected would retrigger | 13:14 |
| fungi | sdague: so the question is why when the parent change re-entered the gate at 07-02 13:42 did that one not get dragged back in with it? | 13:14 |
| sdague | fungi: yeh | 13:14 |
| fungi | which i'm hunting for the answer to now | 13:14 |
| *** oomichi has quit IRC | 13:14 | |
| *** Longgeek has quit IRC | 13:15 | |
| fungi | ohhhhhhhhhh | 13:17 |
| *** amitpp has quit IRC | 13:17 | |
| *** pcrews has joined #openstack-infra | 13:18 | |
| anteaya | sounds like you found something | 13:18 |
| *** chandan_kumar is now known as chandankumar | 13:18 | |
| fungi | sdague: http://paste.openstack.org/show/85413/ | 13:18 |
| sdague | fungi: yeh, that looks like an actual issue :) | 13:19 |
| fungi | it failed to report a -1 on the change when that happened, and as a result had no prior check result so the clean check requirement was never met to reenqueue it later | 13:19 |
| *** alexpilotti has joined #openstack-infra | 13:20 | |
| *** mriedem has joined #openstack-infra | 13:20 | |
| fungi | so i wonder what's up with the LOG_PATH key not existing | 13:20 |
| fungi | we hit that exception 5 times that day and 13 times the day before | 13:21 |
| *** julim has joined #openstack-infra | 13:21 | |
| *** Longgeek has joined #openstack-infra | 13:21 | |
| fungi | and quite a bit for the week prior to that as well | 13:21 |
| fungi | but not in the couple days since | 13:22 |
| *** dizquierdo has joined #openstack-infra | 13:23 | |
| fungi | and whatever error that is aside, we probably want to treat an parent enqueuing when there's no verify vote from zuul on the child change the same as a recheck on the child change | 13:23 |
| fungi | otherwise these are left in limbo until someone notices and comments on them | 13:24 |
| fungi | jeblair: once you're up and about, thoughts appreciated ^ | 13:24 |
| *** Longgeek_ has quit IRC | 13:24 | |
| fungi | i'll get a couple zuul bugs opened and dig some more | 13:29 |
| *** e0ne has quit IRC | 13:29 | |
| anteaya | sdague: on the readme it never says your irc nick, it says find me on irc but no nick | 13:29 |
| fungi | clearly intentional ;) | 13:30 |
| anteaya | it appears so | 13:30 |
| anteaya | wander around aimlessly calling into the woods, I'll find you | 13:30 |
| *** dangers_away is now known as dangers | 13:31 | |
| fungi | YorikSar: hashar: i'm tagging git-review 1.24 here in just a bit... any last-minute comments/additions for the release notes? https://etherpad.openstack.org/p/git-review-1.24 | 13:31 |
| *** zz_gondoi is now known as gondoi | 13:31 | |
| hashar | fungi: I haven't followed git-review development lately :/ | 13:32 |
| fungi | hashar: fair enough! | 13:32 |
| *** basha has joined #openstack-infra | 13:32 | |
| hashar | only got involved again recently because of the hash mismatch that was hitting wikimedia gerrit as well :D | 13:32 |
| fungi | hashar: yeah, i suspect we'll fix that in our tests once they release 2.8.7 | 13:33 |
| *** e0ne has joined #openstack-infra | 13:33 | |
| anteaya | sdague: this was the one I was talking about: http://paste.openstack.org/show/85416/ | 13:34 |
| *** gokrokve has joined #openstack-infra | 13:34 | |
| anteaya | sdague: did you create that? | 13:34 |
| *** jaypipes has joined #openstack-infra | 13:35 | |
| *** nosnos_ has quit IRC | 13:37 | |
| *** homeless has joined #openstack-infra | 13:39 | |
| *** gokrokve has quit IRC | 13:39 | |
| mordred | morning all | 13:41 |
| *** e0ne has quit IRC | 13:41 | |
| *** wenlock has joined #openstack-infra | 13:42 | |
| *** basha_ has joined #openstack-infra | 13:44 | |
| *** e0ne has joined #openstack-infra | 13:46 | |
| *** basha has quit IRC | 13:47 | |
| *** basha_ is now known as basha | 13:47 | |
| hashar | I had a Nodepool question for you guys. I notice there is a prepare_node.sh script, are the instances being booted stock images of distro that then need a lot of preparation work before being pooled in? | 13:47 |
| hashar | I was expecting you to have fully prepared images on the cloud provider to speed up the pooling of a new instance | 13:48 |
| fungi | hashar: sort of. those prepare scripts are actually run on a template instance during image update | 13:48 |
| mordred | hashar: yah. so, it's both things | 13:48 |
| fungi | hashar: after they complete, the instance is stopped and its filesystem snapshotted, then new nodes are booted from the snapshot and added to the pool | 13:49 |
| *** nosnos has joined #openstack-infra | 13:49 | |
| hashar | ahhh | 13:49 |
| fungi | hashar: so we usually only run those prepare scripts once a day per label per provider | 13:49 |
| fungi | since that's the frequency of our scheduled image updates | 13:49 |
| hashar | http://ci.openstack.org/nodepool/ might need a 30'000 feet overview :-D | 13:50 |
| hashar | so node pool boot a stock image, run the prep_node.sh script, snapshot that image and new instances are booted from that snapshot isn't it? | 13:51 |
| fungi | hashar: basically anything we can accomplish in advance within reason for the capabilities of an instance and without compromising the integrity of the jobs which will be run on the resulting nodes is done in advance and cached on the image | 13:51 |
| hashar | (rephrasing to make sure I understood) | 13:51 |
| fungi | yep | 13:51 |
| fungi | so that's why we have scripts in there updating distro packages, pre-caching git repositories, et cetera | 13:52 |
| hashar | and node pool as a ssh private key to ssh to the insntance, it then upload the scripts and execute them there right? | 13:52 |
| hashar | snapshoting being done over OpenStack API I guess | 13:52 |
| fungi | correct. those scripts run in the context of the template instance, and are put there and started by the nodepool server | 13:52 |
| *** talluri has joined #openstack-infra | 13:52 | |
| *** mriedem has quit IRC | 13:54 | |
| fungi | hashar: keep in mind that some of the complexity evolved to deal with the fact that our providers lacked any means for us to upload images of our own. there is an effort underway currently to improve this to build-once-upload-everywhere now that we have access to glance services | 13:54 |
| hashar | I am probably going to use nodepool/ openstack myself | 13:54 |
| hashar | though that will be using our own cloud, so it is quite easy to provide custom images that already have everything | 13:54 |
| *** jamespage_ has joined #openstack-infra | 13:55 | |
| hashar | I am going to have a bit of fun installing/setting up nodepool | 13:55 |
| fungi | hashar: right, you could probably have a prepare_node.sh which is effectively a no-op | 13:55 |
| *** mriedem has joined #openstack-infra | 13:56 | |
| *** jamespage__ has joined #openstack-infra | 13:56 | |
| hashar | fungi: thanks much helpful. | 13:56 |
| *** arnaud__ has joined #openstack-infra | 13:57 | |
| *** jamespage_ has quit IRC | 13:57 | |
| *** zigo has joined #openstack-infra | 13:58 | |
| fungi | looks like we don't (yet) have a slide presentation which covers much of nodepool? that would be a good one to add | 13:59 |
| *** nosnos has quit IRC | 13:59 | |
| *** dkranz has quit IRC | 14:00 | |
| *** xchu has joined #openstack-infra | 14:01 | |
| *** markwash has joined #openstack-infra | 14:02 | |
| *** dkehnx has quit IRC | 14:02 | |
| *** ominakov has quit IRC | 14:04 | |
| *** ominakov has joined #openstack-infra | 14:04 | |
| adalbas | hi! i'm having issues when running stack.sh, where the upgrade of setuptools seems to be not working properly. It remains in version 0.6c11 instead of the latest (5.3). | 14:04 |
| adalbas | this causes failure when installing prettytable. | 14:05 |
| sandywalsh | hey guys ... how do I remove a branch dependency in gerrit? | 14:05 |
| adalbas | Has anyone seen this? | 14:05 |
| adalbas | (i'm running that on a CI environment) | 14:05 |
| *** wenlock has quit IRC | 14:05 | |
| YorikSar | fungi: Yay! :) | 14:08 |
| *** luqas has joined #openstack-infra | 14:08 | |
| *** jamespage__ has quit IRC | 14:08 | |
| *** alexpilotti_ has joined #openstack-infra | 14:08 | |
| YorikSar | fungi: btw, we should verify README if it's still right... | 14:08 |
| anteaya | sandywalsh: I'm foggy, you want a branch removed from a repo? or your repo is dependant on the branch of another repo? | 14:09 |
| *** alexpilotti has quit IRC | 14:09 | |
| *** alexpilotti_ is now known as alexpilotti | 14:09 | |
| *** yfried_ has joined #openstack-infra | 14:10 | |
| fungi | YorikSar: good point. i'll give it a re-read since it's been nearly a year from the last release | 14:12 |
| *** BobBall has quit IRC | 14:12 | |
| *** eharney has joined #openstack-infra | 14:13 | |
| *** yfried has quit IRC | 14:13 | |
| *** dizquierdo has quit IRC | 14:13 | |
| sandywalsh | anteaya: there was a review branch that was dependent on another review branch ... but we don't need the parent branch yet and wanted to remove the dependency. | 14:16 |
| sandywalsh | anteaya: but I think we have it via a rebase off master and git review -X | 14:16 |
| fungi | sandywalsh: just rebase the change on master | 14:16 |
| fungi | or cherry-pick onto it, yeah | 14:16 |
| *** mtreinish has joined #openstack-infra | 14:16 | |
| sandywalsh | fungi: cool, thanks | 14:16 |
| *** dims_ has quit IRC | 14:16 | |
| anteaya | adalbas: how old is your copy of devstack? | 14:17 |
| adalbas | anteaya, the latest. I get it from upstream. | 14:18 |
| anteaya | have you successfully run stack.sh previously? | 14:18 |
| adalbas | anteaya, and i am running it on a fedora image. | 14:18 |
| anteaya | what version of fedora? | 14:18 |
| adalbas | anteaya, yes, this was running on the PowerKVM CI, and suddenly it started those errors. F19 | 14:19 |
| *** gokrokve has joined #openstack-infra | 14:19 | |
| anteaya | is this the powerkvm issue that started on Tuesday? | 14:19 |
| adalbas | yes | 14:19 |
| anteaya | okay so start looking at new releases of packages that affect f19 that were released just prior to your errors | 14:20 |
| anteaya | ianw: any input on difficulty with upgrading setuptools on f19? | 14:20 |
| *** jgallard has joined #openstack-infra | 14:21 | |
| *** xchu has quit IRC | 14:21 | |
| *** msabramo has quit IRC | 14:21 | |
| *** prad has joined #openstack-infra | 14:21 | |
| adalbas | anteaya, sure. on another hand, ArxCruz is trying to get this work on F20 as well | 14:22 |
| *** andreaf has joined #openstack-infra | 14:22 | |
| anteaya | great | 14:22 |
| anteaya | that was going to be my next suggestion, I'm glad you are way ahead of me on this | 14:22 |
| *** dkehn_ has joined #openstack-infra | 14:22 | |
| anteaya | adalbas: if ianw shows up, he is going to be your best fedora resource | 14:23 |
| anteaya | that I know of | 14:23 |
| adalbas | anteaya, tks! | 14:23 |
| *** dkehn_ is now known as dkehnx | 14:23 | |
| anteaya | adalbas: thanks for asking, I hope you find the issue, do share once you figure it out | 14:23 |
| *** alexpilotti has quit IRC | 14:23 | |
| adalbas | anteaya, yes, as soon as I have more info, I ll share it on this channel | 14:24 |
| *** prad has quit IRC | 14:24 | |
| *** yfried_ is now known as yfried | 14:24 | |
| krtaylor | anteaya, adalbas I remember just recent prettytable discussions, don't remember specifics though | 14:24 |
| krtaylor | just a day ot two ago? | 14:24 |
| adalbas | krtaylor, yes | 14:24 |
| mordred | ok. I fixed and resubmitted my PR for adding logging config to bandersnatch | 14:26 |
| anteaya | I think that was canada day for me | 14:26 |
| anteaya | hey mordred | 14:26 |
| mordred | hey anteaya | 14:26 |
| anteaya | so I'm going to try to get a walk in before more rain shows up | 14:27 |
| *** jgrimm has joined #openstack-infra | 14:27 | |
| anteaya | back in a while | 14:27 |
| *** dkranz has joined #openstack-infra | 14:27 | |
| *** wenlock has joined #openstack-infra | 14:28 | |
| *** pcm_ has joined #openstack-infra | 14:28 | |
| openstackgerrit | Lucas Eznarriaga proposed a change to openstack-infra/config: Added advise to use dns over hardcoded ip in third_party https://review.openstack.org/104565 | 14:29 |
| *** mfer has joined #openstack-infra | 14:29 | |
| fungi | mordred: you said pbr should be unversioned as a setup_requires yes? if so, i'll fix http://git.openstack.org/cgit/openstack-infra/git-review/tree/setup.py | 14:29 |
| *** mtreinish has quit IRC | 14:30 | |
| *** mtreinish has joined #openstack-infra | 14:30 | |
| *** nelsnelson has joined #openstack-infra | 14:30 | |
| mordred | fungi: yes | 14:31 |
| mordred | fungi: setup_requires has no way to recover from a version mismatch | 14:31 |
| *** mtreinish has quit IRC | 14:31 | |
| openstackgerrit | yolanda.robla proposed a change to openstack-infra/nodepool: Build images using diskimage-builder https://review.openstack.org/46482 | 14:31 |
| yolanda | jeblair, i pushed some changes according to your review. Will be good if you take a look, i'm still testing it | 14:32 |
| *** gokrokve has quit IRC | 14:32 | |
| fungi | mordred: thanks | 14:32 |
| openstackgerrit | Chuck Short proposed a change to openstack-infra/config: Add python34 to hacking gate. https://review.openstack.org/104568 | 14:32 |
| *** otter768 has joined #openstack-infra | 14:32 | |
| *** prad has joined #openstack-infra | 14:32 | |
| *** mtreinish has joined #openstack-infra | 14:32 | |
| *** msabramo has joined #openstack-infra | 14:33 | |
| *** skolekonov has quit IRC | 14:33 | |
| *** devkulkarni has joined #openstack-infra | 14:33 | |
| mordred | dstufft: requests.request("PURGE", "http://pypi.python.org/pypi/skil/json") | 14:34 |
| *** msabramo has quit IRC | 14:34 | |
| *** msabramo has joined #openstack-infra | 14:34 | |
| mordred | dstufft: is that the right thing I want to purge if I get a stale? or do I want to purge without the json? | 14:34 |
| *** hashar has quit IRC | 14:36 | |
| Alex_Gaynor | mordred: you send a PUREGE to whichever URL you want to purge | 14:36 |
| mordred | Alex_Gaynor: yah | 14:36 |
| Alex_Gaynor | if you want the JSON afresh, that one, if you want the simple inex, hit that | 14:36 |
| mordred | Alex_Gaynor: so if I get this: | 14:37 |
| mordred | 2014-07-01 18:10:01,692 ERROR: Stale serial for package skll | 14:37 |
| Alex_Gaynor | Probably easiest to just hit the trio of /pypi/skil, /pypi/skil/json, /simple/skil | 14:37 |
| mordred | that's not enough to make the call... I need to look at 2014-07-01 18:10:01,692 DEBUG: Expected PyPI serial 1143447 for request https://pypi.python.org/pypi/skll/json but got 1143446 | 14:37 |
| *** _nadya_ has joined #openstack-infra | 14:37 | |
| mordred | Alex_Gaynor: I'll try that | 14:38 |
| *** mostly_d34dh0r53 is now known as d34dh0r53 | 14:38 | |
| *** chandankumar has quit IRC | 14:38 | |
| *** dkehn_ has joined #openstack-infra | 14:39 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Update the README to mention dependencies https://review.openstack.org/104570 | 14:39 |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Update requirements to OpenStack's recommendations https://review.openstack.org/104571 | 14:39 |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Update homepage on PyPI https://review.openstack.org/104572 | 14:39 |
| fungi | some minor polish to squeeze in prior to release ^ | 14:39 |
| *** gokrokve has joined #openstack-infra | 14:40 | |
| *** skolekonov has joined #openstack-infra | 14:41 | |
| *** dkehn_ has quit IRC | 14:42 | |
| *** dkehn has quit IRC | 14:42 | |
| *** otter768 has quit IRC | 14:42 | |
| *** alexpilotti has joined #openstack-infra | 14:42 | |
| yolanda | jeblair, mordred, i don't like the way i use for detecting if there is an image already built or a image building in process, i'd prefer to add a table into database | 14:42 |
| *** _nadya_ has quit IRC | 14:42 | |
| *** dkehn_ has joined #openstack-infra | 14:42 | |
| mordred | yolanda: that seems like it's in-line with how other things keep track of state | 14:43 |
| *** dkehnx has quit IRC | 14:43 | |
| *** ihrachyshka has quit IRC | 14:43 | |
| *** gargola has quit IRC | 14:43 | |
| *** rcarrill` has joined #openstack-infra | 14:43 | |
| *** dkehn_ is now known as dkehnx | 14:43 | |
| yolanda | mordred, what do you mean? | 14:43 |
| *** dkehn has joined #openstack-infra | 14:43 | |
| mordred | yolanda: I mean, other things keep track of things in the database, so that seems reasonable | 14:43 |
| yolanda | yes, at the moment just looking at images dir to see if there is a file matching with the image name pattern there to detect if is built | 14:44 |
| yolanda | or using a set of items, to check if a image is in building process | 14:44 |
| *** gargola has joined #openstack-infra | 14:44 | |
| yolanda | doesn't look very robust | 14:44 |
| *** rcarrillocruz has quit IRC | 14:45 | |
| *** talluri has quit IRC | 14:47 | |
| *** dkranz has quit IRC | 14:49 | |
| *** otherwiseguy has joined #openstack-infra | 14:49 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Update homepage on PyPI https://review.openstack.org/104572 | 14:55 |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Update requirements to OpenStack's recommendations https://review.openstack.org/104571 | 14:55 |
| Kiall | Question - olsosphinx in the docs jobs, do they use a pypi or git version of oslosphix? | 14:55 |
| fungi | Kiall: you should be able to tell from the log | 14:56 |
| *** zz_dimtruck is now known as dimtruck | 14:56 | |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Trap for stale packages and purge them https://review.openstack.org/104574 | 14:56 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Trap for stale packages and purge them https://review.openstack.org/104574 | 14:57 |
| Kiall | fungi: annoyingly, the docs jobs dont upload the log to l.o.o ;) | 14:57 |
| mordred | Alex_Gaynor, dstufft: ^^ santiy check me? | 14:57 |
| *** gokrokve has quit IRC | 14:57 | |
| fungi | Kiall: sure they do | 14:57 |
| Kiall | http://docs-draft.openstack.org/37/104337/1/check/gate-designate-docs/88189da/ - ? | 14:57 |
| Kiall | Oh - docs-draft | 14:58 |
| Kiall | Found it | 14:58 |
| * Kiall hangs his head.. | 14:58 | |
| fungi | Kiall: they're uploaded by the git commit sha since they're run after merge (either in post or (pre)release pipelines generally) | 14:58 |
| *** pcm_ has quit IRC | 14:58 | |
| *** pcm_ has joined #openstack-infra | 14:59 | |
| fungi | Kiall: the url looks like http://logs.openstack.org/xx/yyyy where yyyy is the full git sha (of the merge commit, not usually the original commit unless it was a fast-forward), and xx is the first two digits of the same sha | 15:00 |
| *** james_li has joined #openstack-infra | 15:00 | |
| *** thedodd has joined #openstack-infra | 15:00 | |
| fungi | Kiall: http://logs.openstack.org/eb/eba6479f3fdc68b3c8bed51544b3274aad6b8010/post/nova-docs/344171f/console.html#_2014-07-03_06_28_21_897 | 15:01 |
| fungi | Kiall: so that particular nova post-merge docs build used oslosphinx==2.1.0 from pypi according to the pip freeze at the end of the job | 15:01 |
| Kiall | fungi: Ah, nice. Okay.. so oslosphinx comes from pypi | 15:02 |
| *** rcarrillocruz has joined #openstack-infra | 15:02 | |
| Kiall | dhellmann: about? I've love if you could cut a olsosphix release that includes https://github.com/openstack/oslosphinx/commit/c144be899391901c34d8ad5a87e457325d8774fb :) | 15:02 |
| Kiall | That would mostly unblock Designate from getting the docs jobs merged | 15:02 |
| dhellmann | Kiall: I'll be doing that Monday (no releases before a long weekend :-) | 15:03 |
| *** rcarrill` has quit IRC | 15:04 | |
| Kiall | dhellmann: totally and utterly unacceptable. Always deploy/release immediately prior to long weekends! ;) | 15:04 |
| *** dims_ has joined #openstack-infra | 15:04 | |
| Kiall | Thanks - I'll keep an eye out for it | 15:04 |
| *** UtahDave has left #openstack-infra | 15:05 | |
| dhellmann | Kiall: So I can put your contact info in the announcement email, and let everyone talk to you if the release causes issues? :-) | 15:05 |
| Kiall | Sure - It's kiall@dev.null :D | 15:05 |
| dhellmann | I'll send email to the -dev list like with the other libs | 15:05 |
| dhellmann | heh | 15:05 |
| *** mmitchell_ has quit IRC | 15:06 | |
| Kiall | I'd have no clue about release issues with oslosphinx to be fair ;) real email or kiall@dev.null - end result would basically be the same :D | 15:06 |
| *** doddstack has joined #openstack-infra | 15:06 | |
| dhellmann | lol | 15:07 |
| fungi | Kiall: it's not as if you benefit from dhellmann's long weekend anyway | 15:07 |
| Kiall | Day off on Friday for me too ;) | 15:08 |
| *** thedodd has quit IRC | 15:08 | |
| fungi | oh, awesome! | 15:08 |
| *** dims_ has quit IRC | 15:08 | |
| fungi | my july 4th preparations include bringing everything from my deck into the living room, taping my trash can lid tightly shut, making sure all my flashlights have fresh batteries, and filling lots of containers with potable water | 15:10 |
| Kiall | Sounds like you're going to be shouting "Get off my lawn!" all weekend too? ;) | 15:11 |
| fungi | bwahahaha | 15:11 |
| *** mmitchell has joined #openstack-infra | 15:11 | |
| dhellmann | fungi: are they expecting the main storm to hit land, or mostly peripheral effects? | 15:11 |
| Kiall | Oh.. Storm? I don't keep up with US weather forecasts ;) | 15:12 |
| Kiall | Sounded like protection from local kids on a 4tj of July rampage! | 15:12 |
| Kiall | 4th* | 15:12 |
| Shrews | fungi: that's what you get for moving! | 15:12 |
| *** markwash has quit IRC | 15:12 | |
| fungi | dhellmann: latest track has it coming mostly over top of us, but they haven't called for evacuation here on bodie island yet. hatteras island has been under mandatory evacuation for all residents since yesterday morning | 15:13 |
| *** pbelanyi has quit IRC | 15:13 | |
| fungi | Shrews: yes, nature is taking its earliest opportunity to test my resolve as a beach bum | 15:14 |
| *** basha has quit IRC | 15:14 | |
| *** pblaho has quit IRC | 15:17 | |
| *** sabeen has joined #openstack-infra | 15:17 | |
| dhellmann | fungi: I'm reading up a little now, and I love the Gov's statement to tourists about going to the beach this weekend: "Don't put your stupid hat on." | 15:17 |
| *** moted has joined #openstack-infra | 15:17 | |
| fungi | yeah, first 'cane of the season and it comes right up the coast at us | 15:18 |
| fungi | barely even summer yet. doesn't bode well for the next several months | 15:18 |
| dhellmann | if I believed in portents... | 15:18 |
| *** radez is now known as radez_g0n3 | 15:20 | |
| *** dkranz has joined #openstack-infra | 15:21 | |
| *** dimtruck is now known as zz_dimtruck | 15:22 | |
| *** matrohon has quit IRC | 15:22 | |
| *** jpeeler has joined #openstack-infra | 15:23 | |
| openstackgerrit | A change was merged to openstack-infra/git-review: Update the README to mention dependencies https://review.openstack.org/104570 | 15:24 |
| *** asselin has joined #openstack-infra | 15:24 | |
| *** gokrokve has joined #openstack-infra | 15:24 | |
| *** doude has quit IRC | 15:25 | |
| *** mtreinish has quit IRC | 15:25 | |
| *** david-ly_ has joined #openstack-infra | 15:27 | |
| anteaya | oh yeah, you are all off tomorrow - stay safe fungi | 15:28 |
| anteaya | I should make sure I run my errands today then | 15:28 |
| *** signed8bit has joined #openstack-infra | 15:29 | |
| fungi | anteaya: i expect to be around, though that in part depends on availability of electricity and other necessities | 15:29 |
| *** devkulkarni has quit IRC | 15:29 | |
| *** skolekonov has quit IRC | 15:29 | |
| anteaya | right | 15:29 |
| anteaya | so happy to have your help (as always) but I will expect some interruptions in service | 15:30 |
| *** reed has joined #openstack-infra | 15:30 | |
| anteaya | I hope I don't get hit with another storm so I have to go offline | 15:30 |
| *** jcoufal has quit IRC | 15:31 | |
| *** yaguang has joined #openstack-infra | 15:31 | |
| *** yfried has quit IRC | 15:32 | |
| *** praneshp has joined #openstack-infra | 15:33 | |
| clarkb | sdague mordred ttx fungi did you see the test matrix thing from last night? | 15:35 |
| *** e0ne has quit IRC | 15:35 | |
| ttx | clarkb: yes, commented on the review | 15:35 |
| mordred | clarkb: no | 15:35 |
| ttx | clarkb: not sure why this is working worse than milestone-proposed though | 15:35 |
| *** mfer has quit IRC | 15:36 | |
| jeblair | clarkb: link? | 15:36 |
| clarkb | https://review.openstack.org/#/c/104448/ | 15:36 |
| ttx | jeblair: looks like you were wrong about that "should just work" thing :) | 15:36 |
| clarkb | ttx because its new | 15:36 |
| clarkb | from after last milestone proposed iirc | 15:37 |
| ttx | clarkb: ah. So proposed/foo is not exactly to blame :) | 15:37 |
| ttx | we'd have had same issue with milestone-proposed | 15:37 |
| *** e0ne has joined #openstack-infra | 15:37 | |
| ttx | jeblair: I retract what i just said | 15:37 |
| fungi | clarkb: hrm... proposed/$1 and stable/$1 should be tested with proposed/$1 and stable/$1 of other projects when they exist, or with master if not | 15:38 |
| *** MarkAtwood has joined #openstack-infra | 15:38 | |
| clarkb | fungi: no | 15:38 |
| fungi | clarkb: what corner case am i missing there? | 15:38 |
| clarkb | brcause projects will go from proposed to stable at different times | 15:38 |
| *** msabramo has quit IRC | 15:39 | |
| fungi | clarkb: right, which is why i said "and" | 15:39 |
| fungi | at any point in time there will be a mix of proposed and stable branches, or proposed and master branches, or possibly proposed, stable and master | 15:39 |
| fungi | (because some projects don't use branches other than master ever) | 15:40 |
| clarkb | right so my change allows for this | 15:40 |
| clarkb | its about picking test setups | 15:40 |
| clarkb | not about which branches are tested together | 15:40 |
| fungi | clarkb: right, so my point was i don't understand the need for an isalpha distinction to single-out numeric release branches | 15:40 |
| *** habib has joined #openstack-infra | 15:41 | |
| clarkb | numeric release branches are essentially meaningless in our existing matrix | 15:41 |
| fungi | since they would fall back to being tested with master of other projects anyway (as there would be no matching stable or proposed branch names for other projects) | 15:41 |
| clarkb | swift is a snowflake here and was not considered | 15:41 |
| clarkb | they would not fall back on master | 15:41 |
| luqas | anteaya: topic third-party means creating a branch third-party, right? | 15:42 |
| clarkb | fungi: for branch selection it would. but test matrix is about test env | 15:42 |
| notmyname | would a proposed/<prefix>-* help? | 15:42 |
| notmyname | eg proposed/juno-2.0.0 | 15:42 |
| *** matrohon has joined #openstack-infra | 15:42 | |
| sdague | clarkb: I proposed a different approach on it | 15:43 |
| clarkb | fungi: test matrix has no fallback. it is explicit and error otherwise | 15:43 |
| clarkb | sdague ya I think yours makes sense except it would mask the lack of eg juno in our matrix | 15:43 |
| clarkb | I still want to fail if we dont know how to juno | 15:44 |
| fungi | clarkb: oh, okay. i was focusing entirely on our branch selection algorithm, yes. i suppose there are $reasons we don't explicitly map environment selection to the branch selection result | 15:44 |
| zaro | morning | 15:45 |
| *** arnaud__ has quit IRC | 15:45 | |
| sdague | clarkb: ok, well I'm intermitent today as there is a multi step transit to vermont for the weekend, will look at it in more detail later | 15:45 |
| clarkb | notmyname: possibly and we could drop the numbers | 15:46 |
| clarkb | and treat it as juno | 15:47 |
| notmyname | clarkb: hmm..but next month when we have 2.0.1 or 2.1.0 then what? also, does this impact client libraries? | 15:47 |
| clarkb | no I mean drop numbers in test matrix | 15:47 |
| notmyname | clarkb: ok, so just the pattern propose/<release>-*? | 15:48 |
| *** jaypipes_ has joined #openstack-infra | 15:48 | |
| clarkb | ya | 15:49 |
| *** jaypipes_ has quit IRC | 15:49 | |
| *** msabramo has joined #openstack-infra | 15:50 | |
| *** jpich has quit IRC | 15:50 | |
| notmyname | seems fine to me. just that I need to repropose the patches when the new branch is set up if that's what's happening | 15:50 |
| clarkb | I am not sure it is necessary yet. | 15:51 |
| clarkb | curious to see what mordred and jeblair say | 15:51 |
| *** pballand has joined #openstack-infra | 15:52 | |
| *** gyee has joined #openstack-infra | 15:52 | |
| fungi | clarkb: a couple of quick and easy +2's for last-minute polish on git-review before i tag 1.24... https://review.openstack.org/104571 and https://review.openstack.org/104572 | 15:53 |
| *** rcarrill` has joined #openstack-infra | 15:53 | |
| *** david-ly_ is now known as david-lyle | 15:54 | |
| anteaya | luqas: it can yes, you can also click the square icon in the gui next to topic and edit it there | 15:54 |
| anteaya | and git review -t <topic> I believe changes the topic in the patch to be what you like | 15:54 |
| anteaya | but let me confirm that | 15:55 |
| *** rcarrillocruz has quit IRC | 15:55 | |
| *** dizquierdo has joined #openstack-infra | 15:55 | |
| *** ihrachyshka has joined #openstack-infra | 15:56 | |
| clarkb | fungi done | 15:56 |
| *** nati_ueno has joined #openstack-infra | 15:56 | |
| anteaya | luqas: yes git review -t <topic> works as well | 15:56 |
| luqas | anteaya: done with the square button of the gui :) thanks | 15:56 |
| anteaya | np | 15:56 |
| fungi | clarkb: thanks | 15:57 |
| anteaya | that method requires to you edit the topic each new patchset | 15:57 |
| anteaya | and thanks for offering that patch, luqas | 15:57 |
| luqas | anteaya: yw | 15:57 |
| *** ildikov has joined #openstack-infra | 15:59 | |
| clarkb | I think elasticsearch hates me | 16:00 |
| *** mtreinish has joined #openstack-infra | 16:00 | |
| clarkb | everything great for a day andbam no more | 16:00 |
| anteaya | I think elasticsearch hates everybody | 16:00 |
| anteaya | I haven't seen it play favourites | 16:00 |
| fungi | jeblair: ran into a fun new (i think) bug in zuul, which i'm digging into now... https://storyboard.openstack.org/#!/story/147 | 16:00 |
| jeblair | fungi: ok, lemme know if you want help | 16:01 |
| *** ildikov_ has quit IRC | 16:02 | |
| *** mmaglana has joined #openstack-infra | 16:02 | |
| clarkb | jeblair: were you going to weigh in on 104448? I will probably go with sdagues suggestion if no one else has something better | 16:03 |
| *** ildikov has quit IRC | 16:03 | |
| *** mtreinish has quit IRC | 16:04 | |
| *** dhellmann is now known as dhellmann_ | 16:04 | |
| openstackgerrit | A change was merged to openstack-infra/git-review: Update requirements to OpenStack's recommendations https://review.openstack.org/104571 | 16:04 |
| *** mtreinish has joined #openstack-infra | 16:04 | |
| *** atiwari has joined #openstack-infra | 16:04 | |
| *** Hal_ has quit IRC | 16:05 | |
| *** dims_ has joined #openstack-infra | 16:05 | |
| fungi | jeblair: separate from that, it exposed a workflow hole with the new pipeline requirements. specifically if zuul fails to vote on a change, then it will fail to reconsider that change later if a parent change is reenqueued into a dependent pipeline (basically same as if it had voted -1). perhaps we should be treating that condition as a recheck? | 16:05 |
| *** rcarrill` has quit IRC | 16:06 | |
| *** anode has joined #openstack-infra | 16:08 | |
| *** dims_ has quit IRC | 16:09 | |
| *** _nadya_ has joined #openstack-infra | 16:10 | |
| *** rcarrillocruz has joined #openstack-infra | 16:11 | |
| jeblair | clarkb: i'm okay with either way; they both solve the immediate problem, the only question is whether we want hard-fail on new proposed branches before we've added them to the config file. | 16:11 |
| *** praneshp has quit IRC | 16:11 | |
| jeblair | fungi: hrm. that would be really hard to implement. | 16:12 |
| fungi | jeblair: yeah, perhaps we just try not to let that happen in the first place ;) | 16:12 |
| *** flaper87 is now known as flaper87|afk | 16:12 | |
| jeblair | fungi: let's do that :) | 16:12 |
| clarkb | jeblair: I am going to go ahead and add juno in that change and do what sdague suggests | 16:12 |
| jeblair | clarkb: ok | 16:13 |
| fungi | did we have a standard/convention yet for referring to storyboard tasks in commit message headers? | 16:14 |
| jeblair | fungi: i've been using "Story: 123" | 16:15 |
| fungi | jeblair: wfm, thanks | 16:16 |
| jeblair | fungi: it's likely going to have to change at some point when we figure out whether and how to address tasks directly | 16:16 |
| fungi | i guess ultimately tasks in a story will... | 16:16 |
| fungi | yeah | 16:16 |
| fungi | i couldn't find a usable task-specific identifier and was just going to us ethe story id number | 16:16 |
| *** nati_ueno has quit IRC | 16:17 | |
| *** Hal_ has joined #openstack-infra | 16:17 | |
| *** nati_uen_ has joined #openstack-infra | 16:17 | |
| *** tkelsey has quit IRC | 16:18 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:19 |
| clarkb | sdague: jeblair notmyname ttx ^ how is that? | 16:19 |
| jeblair | i like the reverse alphabetical order there :) | 16:20 |
| notmyname | clarkb: seems that it would mean that there would be many proposed/juno branches during the juno cycle | 16:20 |
| notmyname | clarkb: ie exact match of "juno",not .startswith(juno) | 16:21 |
| clarkb | notmyname: ya I don't think your change is necessary to do juno-2.0.0 | 16:21 |
| zaro | msabramo: ping | 16:21 |
| clarkb | unless wait | 16:21 |
| clarkb | are we doing proposed/juno* for the milestones? | 16:21 |
| *** dims_ has joined #openstack-infra | 16:21 | |
| clarkb | I thought we were just tagging those? | 16:21 |
| notmyname | clarkb: point being, isn't this the same situation as milestone-proposed where it would need to be deleted and recreated? | 16:21 |
| clarkb | notmyname: no because I thought we agreed we would only tag in these situations | 16:22 |
| notmyname | clarkb: right, tagging, but (eg with swift now) what if there is a backport to that marker | 16:22 |
| clarkb | notmyname: you wouldn't | 16:22 |
| clarkb | notmyname: you would tag a new thing in the future | 16:22 |
| clarkb | but swift si different because swift treats it as an actual release | 16:22 |
| fungi | clarkb: what ttx was saying is that he thought better of that, and that there might after all need to be some backported releases to inter-cycle milestones for other projects as well, in which case they and swift inter-cycle releases could use a similar branch pattern | 16:23 |
| *** devkulkarni has joined #openstack-infra | 16:23 | |
| jeblair | fungi: nova might do a backport to realase an icehouse-milestone2.1? | 16:24 |
| fungi | er, backported patches | 16:24 |
| *** yaguang has quit IRC | 16:24 | |
| msabramo | hey zaro | 16:24 |
| zaro | msabramo: i forgot to check doc generation on the change to introduce pbr to python-jenkins. anyways it's failing for me. getting the following sphinx error api.rst:4: SEVERE: Duplicate ID: "module-jenkins" | 16:24 |
| jeblair | to be set expectations, that's insane and there's only one right answer to my question. :) | 16:24 |
| fungi | jeblair: it sounded like ttx was saying nova might backport a change or two to proposed/juno-2 before tagging juno-2 | 16:25 |
| fungi | ir icehouse-2 or whatever | 16:25 |
| msabramo | zaro: oh hmmm let me see | 16:25 |
| zaro | msabramo: i'm not sure how to fix that. full log is here: http://paste.openstack.org/show/85432/ | 16:25 |
| jeblair | fungi: hrm, that's in contradiction to what he said at the summit | 16:25 |
| jeblair | i wonder what changed his mind? | 16:25 |
| clarkb | fungi: " We said we would just re-tag on master, but alternatively we could do the swift model" is what ttx said. I am interpretting that as a listing of options but I don't really see him changing his mind | 16:25 |
| fungi | jeblair: yes, he acknowledged that it was a contradition to what he said at the summit | 16:25 |
| jeblair | fungi: at the summit, he was pretty sure that never happened in practice | 16:25 |
| * fungi finds scrollback from an hour ago | 16:26 | |
| *** pcm_ has quit IRC | 16:26 | |
| clarkb | oh I am reading from change | 16:26 |
| jeblair | okay i think i'm caught up now | 16:26 |
| fungi | ahh, it was indeed in the change comment, not in irc. my bad | 16:26 |
| zaro | msabramo: google indicates that there's a sphinx bug. | 16:26 |
| *** pcm_ has joined #openstack-infra | 16:26 | |
| jeblair | so, yeah, if we wanted to support that case, the notmyname's suggestion of 'startswith' would be necessary | 16:27 |
| jeblair | heh "the notmyname". "..then notmyname's..." | 16:27 |
| fungi | i do agree that reusing proposed/juno several times in a cycle more or less takes us back to the arguments for getting rid of the reused milestone-proposed branch | 16:28 |
| clarkb | fungi: correct we do not want to do that | 16:28 |
| jeblair | fungi: yeah, i don't think we would do that regardless | 16:28 |
| msabramo | zaro: ok well I reproduced it at least | 16:28 |
| clarkb | fungi: but what we agreed at the summit to do was only have one porposed/foo branch | 16:28 |
| clarkb | at release time | 16:28 |
| *** _nadya_ has quit IRC | 16:28 | |
| clarkb | and use tags otehrwise | 16:28 |
| notmyname | seems that startswith(proposed/juno) handles both cases | 16:28 |
| fungi | but proposed/juno-1 or proposed/juno-2 seems sane for inter-cycle milestones and inter-cycle releases | 16:28 |
| jeblair | fungi: the current patch is written for the summit plan which is "tag milestones, proposed/foo branch is created only for final release" | 16:28 |
| clarkb | notmyname: except the mapping goes the other way but sure | 16:29 |
| fungi | jeblair: yep, agreed | 16:29 |
| notmyname | clarkb: pseudocode ;-) | 16:29 |
| *** zhiyan is now known as zhiyan_ | 16:29 | |
| *** bdpayne has joined #openstack-infra | 16:29 | |
| fungi | jeblair: which brings us back to devstack-gate's test matrix didn't consider swift inter-cycle releases | 16:29 |
| *** rcarrill` has joined #openstack-infra | 16:30 | |
| fungi | unless swift goes back to the plan of only tagging those on master and not backporting anything to them | 16:30 |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Formatting the json for readability https://review.openstack.org/104612 | 16:30 |
| jeblair | fungi: the current patch fixes that without "fixing" it for the rest of openstack | 16:30 |
| *** zz_dimtruck is now known as dimtruck | 16:31 | |
| clarkb | I am writing the other patch and we can decide which we prefer | 16:31 |
| *** markmc has quit IRC | 16:31 | |
| fungi | jeblair: the current patch only fixes it if swift reuses proposed/juno for releases between now and the integrated juno release, unless i'm misreading | 16:31 |
| jeblair | clarkb: i'm 40% of a mind to say leave it as-is and if ttx wants to change the plan, then make a patch to change the plan... but i'm 60% of a mind to say, oh well, we may as well just allow for it now and it's ready if he wants to use it. | 16:32 |
| jeblair | fungi: the current patch is designed to support "proposed/2.0.0" for swift | 16:32 |
| clarkb | I am going to use for else just for mordred | 16:32 |
| notmyname | jeblair: no, it's proposed/juno for swift, right? | 16:32 |
| fungi | jeblair: oh, true | 16:32 |
| *** rcarrillocruz has quit IRC | 16:32 | |
| *** luqas has quit IRC | 16:32 | |
| fungi | jeblair: but only if 2.0.0 gets added to the allowed branches list though, right? | 16:33 |
| *** mfer has joined #openstack-infra | 16:33 | |
| jeblair | fungi: no, it's because 2.0.0 isn't in allowed branches that this works | 16:33 |
| fungi | gah, right, the branch = GRID['branches']['default'] fallthrough there | 16:33 |
| jeblair | yep | 16:33 |
| fungi | okay, i'm back to being okay with this plan | 16:33 |
| openstackgerrit | Ian Cordasco proposed a change to openstack/requirements: Add rfc3986 for Nova URI validations https://review.openstack.org/103892 | 16:34 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:34 |
| clarkb | I think that will do it | 16:34 |
| jeblair | notmyname: i'm pretty sure that in no case we want to reuse a branch in the future, so proposed/juno would only happen for swift for the actual juno release (if at all) | 16:34 |
| *** rickerc has joined #openstack-infra | 16:35 | |
| notmyname | jeblair: ack. and for participating in in the integrated release, i'd expect that we'd (swift) use proposed/juno if that's what everyone else is using. the fall-though of default or event the startswith (current patch?) seems to work overall | 16:36 |
| fungi | clarkb: yeah, and that supports both the namespaced proposed/juno-2 case of matching with juno branches (if they exist) and the proposed/11.0 matching master | 16:36 |
| *** matrohon has quit IRC | 16:36 | |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Formatting the json for readability https://review.openstack.org/104612 | 16:36 |
| jeblair | clarkb: is it worth a comment explaining the intended behavior for the startswith match? | 16:37 |
| zaro | jeblair: tried pushing tag to gerrit (for python-jenkins project) but got a "prohibited by Gerrit" message. Would you be able to fix that? | 16:37 |
| clarkb | jeblair: oh probably. I will add that | 16:37 |
| msabramo | zaro: :noindex: got rid of the warning | 16:37 |
| msabramo | let me check that it rendered OK | 16:37 |
| jeblair | zaro: done; add others to https://review.openstack.org/#/admin/groups/323,members as you wish | 16:37 |
| notmyname | clarkb: jeblair: ok, so looks like we need a new branch in swift? the proposed/juno-2.0.0 branch? or do we stick with proposed/2.0.0 and it gets treated as the default compare against master? | 16:38 |
| zaro | jeblair: thanks! | 16:38 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 16:38 |
| msabramo | zaro: looks ok to me; I will send a review request | 16:38 |
| clarkb | notmyname: I think you can stick with what you have | 16:38 |
| jeblair | notmyname: proposed/2.0.0 -- i don't think we want to use something like 'juno-2.0.0' -- it doesn't really semantically make sense (at least, not to anyone not _extremely_ versed in our version mechanics) | 16:39 |
| openstackgerrit | Marc Abramowitz proposed a change to stackforge/python-jenkins: Fix `tox -e doc` duplicate id error https://review.openstack.org/104615 | 16:39 |
| notmyname | clarkb: ok. also seems that in the future we'd use proposed/release-ver | 16:39 |
| *** amitpp has joined #openstack-infra | 16:39 | |
| jeblair | notmyname: (and possibly not even us ;) | 16:39 |
| notmyname | jeblair: oh. s not that ^^ | 16:39 |
| msabramo | zaro: https://review.openstack.org/104615 | 16:39 |
| * anteaya can't focus with all the construction noise - moving to the library | 16:40 | |
| notmyname | jeblair: well, it makes sense to me I think. it's the swift 2.0.0 release area for backports, if needed. and it's developed during the openstack juno integrated cycle | 16:40 |
| *** penguinRaider has quit IRC | 16:40 | |
| notmyname | jeblair: it's not that juno and 2.0.0 are semantically linked. it's a temporal thing | 16:40 |
| *** anode has quit IRC | 16:40 | |
| jeblair | notmyname: it's not necessary. do you _want_ to call it juno-2.0.0? | 16:40 |
| notmyname | jeblair: ie emphasizing that 2.0.0 happend during the juno integrated cycle | 16:41 |
| jeblair | notmyname: i don't know what the 2.0.0 version of juno is, so i'm thoroughly confused by it. :) | 16:41 |
| notmyname | jeblair: for today, not too important I think. it should "just work". but for the future, I think I do want to have the cycle name in there. just so that we can show and test against the other integrated projects at the right time | 16:41 |
| *** Longgeek has quit IRC | 16:41 | |
| jeblair | notmyname: no other projects will have juno branches at that point | 16:42 |
| notmyname | jeblair: ya, I think the "-" is the separator, not a qualifier. | 16:42 |
| notmyname | jeblair: ah, good point. so master makes sense then | 16:42 |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Formatting the json for readability https://review.openstack.org/104612 | 16:43 |
| *** katyafervent__ is now known as katyafervent_awa | 16:43 | |
| *** markwash has joined #openstack-infra | 16:44 | |
| *** dhellmann_ is now known as dhellmann | 16:44 | |
| *** Hal_ has quit IRC | 16:44 | |
| zaro | msabramo: thanks. | 16:44 |
| msabramo | zaro: once we fix this, do you want to cut some releases? | 16:45 |
| zaro | msabramo: i just created a 0.3.0 release on commit before pbr into. | 16:45 |
| *** anode has joined #openstack-infra | 16:45 | |
| zaro | *into/intro | 16:45 |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Formatting the json for readability https://review.openstack.org/104353 | 16:46 |
| *** rakhmerov has quit IRC | 16:46 | |
| notmyname | clarkb: jeblair: so it seems the current patch set is good? once it lands I can recheck the proposed swift patches? | 16:46 |
| *** afazekas_ has quit IRC | 16:47 | |
| clarkb | notmyname: yup | 16:48 |
| jeblair | notmyname: i think so, unless clarkb wants to add that comment. can always add it in a followup though. :) | 16:48 |
| clarkb | jeblair: I added the comment | 16:48 |
| notmyname | jeblair: he already did | 16:48 |
| jeblair | oh neato | 16:48 |
| msabramo | zaro: oh cool | 16:48 |
| jeblair | need to make gerritbot use blinking text | 16:48 |
| pleia2 | good morning | 16:49 |
| clarkb | jeblair: rainbow colors | 16:49 |
| notmyname | jeblair: clarkb: after it's approved, what's the eta to being usable? minutes? hours? (I gotta reboot to take care of some updates, so if it's soon, I'll wait) | 16:49 |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Updating Horizon core with recent changes https://review.openstack.org/104353 | 16:50 |
| fungi | immediate i think, since it will get updated from git (as part of devstack-gate) | 16:50 |
| jeblair | aprvd. i'm assuming sdague will be okay with it based on his earlier comment, but since he's in and out i don't want to block on his review. | 16:50 |
| clarkb | jeblair: ++ we can always adjust later | 16:50 |
| jeblair | clean check is the blocker here... | 16:51 |
| fungi | well, not immediately after it's approved, but immediately after it merges | 16:51 |
| notmyname | ok | 16:51 |
| jeblair | if you stacked a reverify behind it, it would use it in the gate, but you can't do that in the check queue and you can't get into the gate queue without a clean check :( | 16:51 |
| jeblair | so yeah, it has to land and then recheck and/or reverify | 16:52 |
| jeblair | or reapprove. or whatever :) | 16:52 |
| *** arnaud has joined #openstack-infra | 16:52 | |
| notmyname | re-something | 16:52 |
| jeblair | exactly | 16:52 |
| *** devkulkarni has quit IRC | 16:52 | |
| fungi | so assuming no random job failures and no bumrush on the gate between now and then, ~2.5 hours | 16:52 |
| notmyname | fungi: before I can recheck the patches? | 16:53 |
| *** rakhmerov has joined #openstack-infra | 16:53 | |
| notmyname | or before my patches land? | 16:53 |
| fungi | notmyname: before you can recheck | 16:53 |
| notmyname | fungi: ah. so 2.5 hours before clarkb's -infra patch takes effect | 16:53 |
| notmyname | right? | 16:54 |
| clarkb | ya, devstack-gate is gated by tempest tests so has to go through that | 16:54 |
| *** rcarrill` has quit IRC | 16:54 | |
| notmyname | got it. | 16:55 |
| notmyname | ok, getting offline to get ready/commute/etc | 16:55 |
| notmyname | clarkb: fungi: jeblair: thanks! | 16:55 |
| jeblair | cross-project depends would probably help in this case | 16:55 |
| jeblair | notmyname: you're welcome | 16:55 |
| *** jgallard has quit IRC | 16:56 | |
| *** adalbas has quit IRC | 16:56 | |
| *** d34dh0r53 is now known as mostly_d34dh0r53 | 16:56 | |
| *** mugsie has quit IRC | 16:57 | |
| *** ominakov has quit IRC | 16:57 | |
| *** penguinRaider has joined #openstack-infra | 16:57 | |
| *** derekh_ has quit IRC | 16:58 | |
| fungi | okay, so it looks like the zuul bug i'm squashing occurs when the scheduler reports on a change which never had any jobs launched (which happens for some changes dequeued from the gate). since Gearman.launch is never called, none of the builds have a LOG_PATH parameter set yet | 17:00 |
| *** rcarrillocruz has joined #openstack-infra | 17:00 | |
| fungi | or i guess it would happen if any one of the builds for the change had not yet been launched | 17:00 |
| *** mostly_d34dh0r53 is now known as d34dh0r53 | 17:01 | |
| fungi | i guess i should start by writing the test case and seeing if i can reproduce the same exception using zuul's unit testing framework | 17:02 |
| jeblair | fungi: that's what i usually do -- much easier than trying to repro for real | 17:02 |
| jeblair | fungi: there are probably tests that already excercise the "remove from queue and report before any jobs are launched" path... | 17:03 |
| jeblair | fungi: the novelty here may be the succes/failure url depending on a parameter set in a custom function (which hasn't run) | 17:04 |
| fungi | jeblair: oh... | 17:04 |
| jeblair | fungi: i don't think the standard configuration for most of the tests does that | 17:04 |
| clarkb | I found my german power socket adaptor last night. that means I am ready to go right? | 17:04 |
| fungi | good point | 17:04 |
| jeblair | clarkb: let me know if you find the gearman power socket adapter | 17:05 |
| clarkb | jeblair: spelling is hard | 17:05 |
| fungi | gearman should have been one of the megaman villains | 17:05 |
| clarkb | jeblair: like columbia and colombia. Living in the northwest it is almost impossible to type colombia | 17:06 |
| clarkb | fungi: this! | 17:06 |
| jeblair | fungi: nice | 17:06 |
| fungi | i imagine he'd look like metalman but with gears instead of circular saw blades | 17:06 |
| clarkb | fungi: and have a bunch of little worker minions | 17:07 |
| *** dimtruck is now known as zz_dimtruck | 17:07 | |
| *** cdub has joined #openstack-infra | 17:08 | |
| *** shashankhegde has joined #openstack-infra | 17:08 | |
| *** jlibosva has quit IRC | 17:08 | |
| clarkb | I am having a hard time not starting MURICA weekend early | 17:10 |
| *** adalbas has joined #openstack-infra | 17:10 | |
| jeblair | i'm planning on starting it a little early this afternoon :) | 17:10 |
| nibalizer | jeblair: if you want to work on the p3 master again today let me know | 17:10 |
| jeblair | nibalizer: you have time now? | 17:11 |
| nibalizer | jeblair: sure | 17:11 |
| clarkb | jeblair: ya I will probably end up doing that too. I will stick around to make sure notmyname's thing is handled though | 17:11 |
| mgagne | guys, I'm a bit discouraged here... I've been trying to merge this patch for weeks and I always hit transient bugs. I probably reverified against them all https://review.openstack.org/#/c/91727/ | 17:11 |
| nibalizer | i only suggested the hand wavy future because I didn't want to cramp your style | 17:11 |
| *** ihrachyshka has quit IRC | 17:12 | |
| jeblair | nibalizer: are there outstanding patches? (that second one?) | 17:12 |
| SpamapS | I'm having a weird problem with git review | 17:12 |
| SpamapS | Rebasing a change against master shows a conflict.. | 17:12 |
| SpamapS | I resolve.. | 17:12 |
| SpamapS | then review.. | 17:13 |
| SpamapS | and it is rejected | 17:13 |
| SpamapS | remote: (W) No changes between prior commit 328e1c6 and new commit cfa14c5 | 17:13 |
| clarkb | mgagne: ok? sorry its a nova change so not sure where to start | 17:13 |
| morganfainberg | SpamapS, i had that issue recently, where the parent was seeing a change due to an odd rebase | 17:13 |
| openstackgerrit | A change was merged to openstack-infra/zuul: Fix up fetching jquery.visibility https://review.openstack.org/91316 | 17:13 |
| SpamapS | this is after having git review -d'd the change | 17:13 |
| jeblair | mgagne: when there are enough transient bugs like that, sometimes people start rallying folks on the mailing list to try to fix the top ones in http://status.openstack.org/elastic-recheck/ | 17:13 |
| mgagne | clarkb: I'm not sure I'm on the right channel too | 17:13 |
| *** praneshp has joined #openstack-infra | 17:13 | |
| clarkb | mgagne: ok :) well do let us know if there is more to that that you think infra can help out whith | 17:13 |
| SpamapS | perhaps I need to rebuild the patch stream from cherry picks | 17:13 |
| clarkb | (eg maybe some of the flakyness was our fault) | 17:14 |
| morganfainberg | SpamapS, my solution, checkout the parent, git rebase origin/master then cherry-pick your change you're updating on the tip | 17:14 |
| nibalizer | jeblair: this should be merged eventually https://review.openstack.org/#/c/104352/ | 17:14 |
| morganfainberg | SpamapS, yeah. | 17:14 |
| SpamapS | morganfainberg: ok that was my next step, ty | 17:14 |
| morganfainberg | SpamapS, sure thing | 17:14 |
| jeblair | mgagne: you may want to talk to jogo_away, or mtreinish in #openstack-qa | 17:14 |
| clarkb | SpamapS: that means the diff between the two commits is nil | 17:14 |
| clarkb | SpamapS: and gerrit says no | 17:14 |
| jeblair | clarkb, fungi: can you review https://review.openstack.org/#/c/104352/ ? | 17:15 |
| morganfainberg | clarkb, i've seen that happen with a wonky rebase | 17:15 |
| *** bnemec is now known as bnemec_appt | 17:15 | |
| SpamapS | clarkb: but why does _git_ think there's a change necessary? | 17:15 |
| openstackgerrit | A change was merged to openstack-infra/git-review: Update homepage on PyPI https://review.openstack.org/104572 | 17:15 |
| morganfainberg | clarkb, git rebases, gerrit sees no change, but the hash updated because of the rebase. | 17:15 |
| fungi | jeblair: reviewing | 17:15 |
| SpamapS | the conflict is legit | 17:15 |
| clarkb | SpamapS: there may be a conflict there, you just happen to essentailly revert everything in your rebase | 17:15 |
| clarkb | jeblair: yup | 17:15 |
| *** devkulkarni has joined #openstack-infra | 17:16 | |
| morganfainberg | SpamapS, i only have had this happen once or twice in 4+ years of heavily using git | 17:16 |
| clarkb | SpamapS: it is possible this may be a subtle gerrit bug. They have good intentions but don't handle this one corner case | 17:16 |
| *** denis_makogon has quit IRC | 17:16 | |
| morganfainberg | SpamapS, it's an edge case where the commit and conflict etc net odd behavior like this. | 17:17 |
| SpamapS | clarkb: ah | 17:17 |
| SpamapS | https://review.openstack.org/#/c/94527/ is the review anyway | 17:17 |
| SpamapS | (with the first unmerged patch in the stream being the one that conflicts) | 17:17 |
| SpamapS | I'm rebuilding the stream as cherry picks | 17:17 |
| clarkb | jeblair: nibalizer +2. have we considered what the eventual CA server move will look like? | 17:17 |
| fungi | nibalizer: does undef evaluate false? i take it this is a preferable construct to '' or false? | 17:18 |
| clarkb | jeblair: nibalizer: just a wholesale rsync and config update at some point? | 17:18 |
| clarkb | fungi: I think puppetism is to use undef | 17:18 |
| fungi | k | 17:18 |
| SpamapS | ah the middle one says outdated | 17:18 |
| SpamapS | I think thats the issue | 17:18 |
| clarkb | fungi: I don't like undef because iirc it does prettyprint to something like 'undef' | 17:18 |
| *** MarkAtwood has quit IRC | 17:18 | |
| fungi | clarkb: and yeah, sounds like copy all the server certs over as well as the hiera file | 17:18 |
| *** otherwiseguy has quit IRC | 17:19 | |
| *** blamar has joined #openstack-infra | 17:20 | |
| SpamapS | gah | 17:20 |
| fungi | new errors i don't recognize in http://logs.openstack.org/52/104352/3/check/gate-config-puppet-apply-precise/419298d/console.html | 17:20 |
| *** jp_at_hp has quit IRC | 17:20 | |
| SpamapS | cherry picks are not working either | 17:20 |
| clarkb | SpamapS: I didn't expect them to | 17:20 |
| nibalizer | fungi: yea i tested that templating pretty extensively | 17:21 |
| clarkb | SpamapS: if the problem is you are doing a revert in your conflict resolution then it may need a revert commit | 17:21 |
| nibalizer | fungi: but maybe it should be further testsed | 17:22 |
| SpamapS | got it | 17:22 |
| SpamapS | clarkb: no I just had to start with a pull, then cherry pick the next 2 | 17:22 |
| SpamapS | clarkb: I think the problem was the middle patch was based on a previous revision | 17:23 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Trap for stale packages and purge them https://review.openstack.org/104574 | 17:23 |
| SpamapS | so needed to keep the latest hash for patch 1, then update the parent for 2 | 17:23 |
| jeblair | clarkb: yeah, i think the idea is we flag-day move the ca to the new server when we're ready | 17:23 |
| SpamapS | clarkb: thanks for the sounding board though. :) | 17:24 |
| mordred | jeblair: the patch above should address the stale pypi stuff | 17:24 |
| *** whoops has joined #openstack-infra | 17:24 | |
| openstackgerrit | David Lyle proposed a change to openstack-infra/reviewstats: Updating Horizon core with recent changes https://review.openstack.org/104353 | 17:25 |
| fungi | nibalizer: none of the new errors in the puppet apply job seemed to be related to manifests your patch altered, so i'm fine with it i suppose | 17:26 |
| nibalizer | what does the puppet-apply test do? | 17:27 |
| *** Ajaeger has joined #openstack-infra | 17:27 | |
| fungi | nibalizer: a --noop apply of all the modules on a precise slave | 17:28 |
| clarkb | nibalizer: for node in site.pp do puppet apply --noop | 17:28 |
| Ajaeger | good evening ;) | 17:28 |
| clarkb | and it does it in parallel is is baller | 17:28 |
| nibalizer | fungi: you can play with this if you want to convince yourself that it will work http://paste.ubuntu.com/7742943/ | 17:28 |
| Ajaeger | This morning the translation jobs were still using old slave scripts - and not the current ones. | 17:28 |
| clarkb | https://jenkins05.openstack.org/job/check-grenade-dsvm-partial-ncpu/3570/console arg that hit the d-g change | 17:28 |
| clarkb | Ajaeger: we only update the slave images once per day at 1400 UTC | 17:29 |
| fungi | nibalizer: http://git.openstack.org/cgit/openstack-infra/config/tree/test.sh | 17:29 |
| *** jdahlin has joined #openstack-infra | 17:29 | |
| Ajaeger | clarkb: The patch for it merged three days ago | 17:29 |
| clarkb | oh wait those jobs run on the priveleged slaves | 17:30 |
| clarkb | I bet puppet is stuck on them | 17:30 |
| *** habib has quit IRC | 17:30 | |
| nibalizer | fungi: ah okay | 17:30 |
| jeblair | clarkb: puppet seems dead on the priv slaves | 17:30 |
| mordred | jeblair: also, my patch to bandersnatch to add logging configurability was merged | 17:30 |
| fungi | Ajaeger: i think puppet wasn't properly updating on the nodepool server until after the image updates ran yesterday? did it correct itself a few hours ago? | 17:30 |
| clarkb | http://puppetboard.openstack.org/node/proposal.slave.openstack.org | 17:30 |
| clarkb | jeblair: ya | 17:30 |
| clarkb | unreported | 17:30 |
| *** pcm_ has quit IRC | 17:30 | |
| openstackgerrit | A change was merged to openstack-infra/config: Allow site.pp to manage ca and ca_sever in puppet.conf https://review.openstack.org/104352 | 17:30 |
| mordred | jeblair: so we can officially view that upstream as "Responsive" | 17:31 |
| mordred | clarkb: why would it be stuck? | 17:31 |
| *** pcm_ has joined #openstack-infra | 17:31 | |
| clarkb | mordred: I don't know | 17:31 |
| *** nati_uen_ has quit IRC | 17:31 | |
| mordred | clarkb: I mean, by that - there is nothing that should stick | 17:31 |
| *** nati_ueno has joined #openstack-infra | 17:31 | |
| clarkb | http://puppetboard.openstack.org/fact/puppetversion that is why | 17:31 |
| jeblair | it ran at 17:22 | 17:31 |
| clarkb | it has puppet 3.6 | 17:32 |
| clarkb | something is I don't even why | 17:32 |
| mordred | oh. well. that's clearly not right | 17:32 |
| clarkb | and 3.6 won't talk to our master | 17:32 |
| *** maxbit has joined #openstack-infra | 17:32 | |
| mordred | nope | 17:32 |
| *** nelsnelson has quit IRC | 17:32 | |
| Ajaeger | It was still wrong at 2014-07-03 15:04:36.857 but if the image was updated at 17:22, then I'll check again later... | 17:33 |
| jeblair | ii puppet 2.7.25-1puppetlabs1 Centralized configuration management - agent startup and compatibility scripts | 17:33 |
| mordred | clarkb: proposal.slave.openstack.org has the proper pin in installed | 17:33 |
| *** Ryan_Lane has joined #openstack-infra | 17:33 | |
| mordred | mordred@proposal:~$ which puppet | 17:33 |
| clarkb | what do people think about directly enqueing my d-g change into the gate? | 17:33 |
| mordred | /usr/local/bin/puppet | 17:33 |
| clarkb | it has only failed partial-ncpu and all of the other things seems happy | 17:33 |
| mordred | did something GEM INSTALL puppet? | 17:33 |
| *** radez_g0n3 is now known as radez | 17:33 | |
| jeblair | Ajaeger: 17:22 wasn't for your conversation | 17:33 |
| Ajaeger | jeblair: ah... | 17:33 |
| clarkb | mordred: probably | 17:33 |
| clarkb | mordred: I think gozer land ran into this too | 17:33 |
| clarkb | mordred: and on our single use slaves that run the puppet apply test they use puppet 3 too | 17:33 |
| clarkb | something is weird | 17:33 |
| mordred | install_puppet.sh: gem install hiera hiera-puppet | 17:34 |
| mordred | this is what we gem install - I wonder if that's triggering puppet install now? | 17:34 |
| fungi | gem dependencies? | 17:34 |
| * mordred testing | 17:34 | |
| mordred | ok. gem install heira-puppet did not result in puppet being installed on my laptop | 17:35 |
| clarkb | mordred: you are brave doing that on your laptop :) | 17:36 |
| *** signed8bit is now known as signed8bit_ZZZzz | 17:36 | |
| *** rcarrill` has joined #openstack-infra | 17:36 | |
| *** rcarrillocruz has quit IRC | 17:36 | |
| clarkb | mordred: we can probably consult our image build logs to see if we log when it gets pulled in | 17:36 |
| *** yfried has joined #openstack-infra | 17:36 | |
| mordred | clarkb: (I'm already managing my laptop with puppet ...) | 17:37 |
| mordred | clarkb: nod | 17:37 |
| jesusaurus | you can always gem remove... oh wait | 17:37 |
| mordred | anybody have a problem with me running gem uninstall puppet ? | 17:37 |
| clarkb | mordred: no, but we should figure out why it happens first maybe? | 17:37 |
| *** maxbit_ has joined #openstack-infra | 17:37 | |
| *** maxbit_ has quit IRC | 17:37 | |
| clarkb | a quick grep in the image log for puppet3 doesn't show anything | 17:38 |
| *** maxbit has quit IRC | 17:38 | |
| mordred | clarkb: grep for gem | 17:38 |
| *** praneshp_ has joined #openstack-infra | 17:38 | |
| *** blamar has quit IRC | 17:39 | |
| *** maxbit has joined #openstack-infra | 17:39 | |
| jeblair | clarkb: why di you think the apply test uses puppet3? | 17:39 |
| clarkb | jeblair: at one point it gave off a bunch of errors from a puppet3 path | 17:40 |
| clarkb | jesusaurus: ^ do you happen to have links to that? | 17:40 |
| jesusaurus | clarkb: im looking | 17:40 |
| jeblair | this recent run doesn't seem to: http://logs.openstack.org/52/104352/3/check/gate-config-puppet-apply-precise/419298d/console.html | 17:40 |
| clarkb | the only puppet-3 stuff I see if for trusty and f20 image builds | 17:40 |
| jeblair | maybe it was transient? | 17:40 |
| clarkb | jeblair: could be | 17:40 |
| jesusaurus | jeblair: there were deprecation warnings being thrown with file paths indicating puppet3 | 17:40 |
| clarkb | jeblair: and since we rebuild images they are fine | 17:40 |
| mordred | yah - maybe we landed a change at some point which borked this fora second | 17:40 |
| clarkb | but the long lived slave would get stuck | 17:40 |
| jeblair | ya | 17:40 |
| mordred | when we were adding support for trusty | 17:41 |
| mordred | I'm going to run this unless someone says no: ansible *slave.openstack.org -a 'gem uninstall puppet' | 17:41 |
| *** praneshp has quit IRC | 17:42 | |
| *** praneshp_ is now known as praneshp | 17:42 | |
| nibalizer | mordred: can you grab 'gem list' before? | 17:42 |
| nibalizer | maybe we can try to figure out which gem pulled in puppet | 17:42 |
| mordred | nibalizer: sure | 17:42 |
| mordred | nibalizer: http://paste.openstack.org/show/85440/ | 17:43 |
| nibalizer | mordred: hrm i suspect puppet_spec_helped | 17:44 |
| nibalizer | er helper | 17:44 |
| nibalizer | er puppetlabs_spec_helper | 17:44 |
| jeblair | image builds failed universally between jun5-10 inclusive; the file date on the puppet binary on mirror27 is june 6 | 17:44 |
| *** amitpp has quit IRC | 17:45 | |
| jeblair | (i have no idea why they failed) | 17:45 |
| nibalizer | hrm puppetlabs_spec_helper seems like it is innocent | 17:46 |
| mordred | nibalizer: yes. that is correct | 17:46 |
| mordred | mordred@proposal:~$ sudo gem uninstall puppet | 17:46 |
| mordred | You have requested to uninstall the gem: | 17:46 |
| mordred | puppet-3.6.1 | 17:46 |
| mordred | puppetlabs_spec_helper-0.5.1 depends on [puppet (>= 0)] | 17:46 |
| *** yfried has quit IRC | 17:47 | |
| nibalizer | oh | 17:47 |
| nibalizer | i mannegd to install ppuppetlabs_spec_helper version 0.6 and it dosen't deppend on puppent | 17:47 |
| mordred | ah. so we should change that depend to puppetlabs_spec_helper >= 0.6 ? | 17:48 |
| *** dizquierdo has quit IRC | 17:48 | |
| nibalizer | confirming that hypothesis now | 17:48 |
| nibalizer | yuuup | 17:48 |
| jesusaurus | ah so we were probably seeing the strange behaviour before 0.6 was released | 17:48 |
| mordred | gotcha | 17:49 |
| nibalizer | so puppetlabs_spec_helper 0.5 pulls in puppet, but 0.6 doesn't | 17:49 |
| nibalizer | so can we safely upgrade that? | 17:49 |
| *** signed8bit_ZZZzz is now known as signed8bit | 17:49 | |
| mordred | don't see any reason why not | 17:49 |
| *** doude has joined #openstack-infra | 17:49 | |
| jeblair | 2014-07-03 14:25:32,529 INFO nodepool.image.build.rax-ord.bare-precise: notice: /Stage[main]/Openstack_project::Thick_slave/Package[puppetlabs_spec_helper]/ensure: created | 17:49 |
| jeblair | 2014-07-03 14:24:34,951 INFO nodepool.image.build.rax-ord.bare-trusty: Notice: /Stage[main]/Openstack_project::Thick_slave/Package[puppetlabs_spec_helper]: Dependency Package[rubygems] has failures: true | 17:50 |
| jeblair | 2014-07-03 14:24:34,952 INFO nodepool.image.build.rax-ord.bare-trusty: Warning: /Stage[main]/Openstack_project::Thick_slave/Package[puppetlabs_spec_helper]: Skipping because of failed dependencies | 17:50 |
| jeblair | seems to be unhappy on trusty | 17:50 |
| *** rkukura has joined #openstack-infra | 17:51 | |
| *** arnaud has quit IRC | 17:52 | |
| sdague | jeblair: I don't believe rubygems is a package on trusty | 17:52 |
| sdague | because gem is bundled in ruby 1.9 | 17:52 |
| nibalizer | ya | 17:52 |
| jeblair | sdague: indeed: http://paste.openstack.org/show/85441/ | 17:52 |
| mordred | yah. I I thought we had a switch for that | 17:52 |
| nibalizer | sdague: also ruby-bundler became bundler | 17:52 |
| *** dkehn_ has joined #openstack-infra | 17:53 | |
| Ajaeger | clarkb, fungi: You were so kind to review and approve again https://review.openstack.org/#/c/103646/ . If you could put the request again on your queue, it would be really appreciated. | 17:53 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New StoryBoard Logo https://review.openstack.org/99217 | 17:53 |
| jeblair | nibalizer: okay let's back-burner that for after we fix the spec_helper, but before we try to launch the new master | 17:53 |
| mordred | puppet should be fixed on *slave.openstack.org now | 17:54 |
| nibalizer | jeblair: *nod* | 17:54 |
| mordred | and I've done a run of puppet agent --test on each to confirm | 17:54 |
| *** maxbit_ has joined #openstack-infra | 17:54 | |
| *** maxbit has quit IRC | 17:54 | |
| *** dkehn has quit IRC | 17:55 | |
| *** mburned_out has joined #openstack-infra | 17:55 | |
| jeblair | okay, i think the rest of the unreporting there make sense | 17:55 |
| nibalizer | is anyone working on a commit to fix spec_helper? I'm able to do it | 17:55 |
| jeblair | nibalizer: nope, go for it | 17:55 |
| mordred | nibalizer: it's fine - upstream has made new release | 17:55 |
| mordred | wait - unless there is something else? | 17:55 |
| *** dkehnx has quit IRC | 17:56 | |
| nibalizer | maybe im confused on the issue | 17:56 |
| nibalizer | i thought our plan was to pin it to 0.6 | 17:56 |
| mordred | I may also be confused on the issue | 17:56 |
| mordred | well, 0.6 is what will naturally get installed now | 17:56 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New StoryBoard Logo https://review.openstack.org/99217 | 17:56 |
| nibalizer | thats true | 17:56 |
| jeblair | so we're done with thing 1 | 17:56 |
| mordred | we can add the pin - but it will be a no op | 17:56 |
| mordred | yah | 17:56 |
| jeblair | nibalizer: now let's work on trusty | 17:56 |
| nibalizer | jeblair: okay | 17:56 |
| clarkb | jeblair: fungi mordred what do you think about enqueuing 104448 in the gate queue? | 17:56 |
| clarkb | or would you prefer it gets rechecked? | 17:56 |
| mordred | jeblair: rubygems has the proper logic in instal_puppet | 17:57 |
| *** dkehn has joined #openstack-infra | 17:57 | |
| *** Ryan_Lane has quit IRC | 17:57 | |
| jeblair | clarkb: you physically can't enqueue it in the gate queue | 17:57 |
| *** marun is now known as marun_afk | 17:57 | |
| clarkb | oh right we broke that | 17:57 |
| jeblair | clarkb: and also, no way. | 17:57 |
| mordred | jeblair: I'm confused why the above error is happening | 17:57 |
| clarkb | jeblair: you could in the past | 17:57 |
| mordred | jeblair: beacuse we also have a switch on the rubygems name in modules/openstack_project/manifests/jenkins_params.pp | 17:57 |
| jeblair | clarkb: yeah, i understand that. i'm saying "no way to we break policy for that change" | 17:58 |
| clarkb | jeblair: sure | 17:58 |
| clarkb | sorry I interpretted no way as there was no physical way | 17:58 |
| jeblair | no, it's me being very clear on the fact that we don't break our policy because we're impatient | 17:58 |
| jeblair | our job is not to enqueue things into the gate queue | 17:59 |
| jeblair | we have computers for that | 17:59 |
| afazekas | clarkb: Which repos needs to be modified in order to add a multi-node job ? | 18:00 |
| nibalizer | jeblair: so i agree that the only place rubygems/ruby is a problem any more looks like install_jenkins_slave.sh | 18:00 |
| nibalizer | pretty easy fix | 18:00 |
| openstackgerrit | A change was merged to openstack-infra/config: Cleanup manual jobs https://review.openstack.org/103646 | 18:00 |
| *** matrohon has joined #openstack-infra | 18:00 | |
| *** devkulkarni has quit IRC | 18:01 | |
| Ajaeger | fungi: Thanks! | 18:01 |
| jeblair | nibalizer: i don't think that's the problem here | 18:01 |
| clarkb | afazekas: openstack-infra/config. You need to add a multinode label to our cloud providers and you need add a job that runs on multinode and you need to tell zuul to run it | 18:01 |
| mordred | jeblair: BWAHAHAHAAHAHAHAHAH | 18:01 |
| mordred | jeblair: you're going to be amused by the patch I'm about to send | 18:01 |
| clarkb | afazekas: so basically the same as all job additions with the additional step of ensuring nodepool gives you the nodes | 18:02 |
| *** d34dh0r53 is now known as mostly_d34dh0r53 | 18:02 | |
| jeblair | Ajaeger: the proposal slave should have updated scripts now, so the next time jobs run they should be what you expect | 18:02 |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/config: Rubygems/Ruby package rename handled in jenkins_slave.sh https://review.openstack.org/104631 | 18:02 |
| jeblair | nibalizer: i don't think we run that script anywhere | 18:02 |
| afazekas | clarkb: will start something on both node ? | 18:02 |
| nibalizer | jeblair: oh, and there i was trying to fix it, hah | 18:02 |
| clarkb | afazekas: no. There is one primary node that is connected to jenkins. All other nodes are subnodes | 18:02 |
| afazekas | clarkb: how can I discover the other nodes on one node ? | 18:02 |
| *** arnaud has joined #openstack-infra | 18:02 | |
| *** arnaud has quit IRC | 18:02 | |
| clarkb | afazekas: so you have to use the primary node for control. | 18:03 |
| openstackgerrit | Monty Taylor proposed a change to openstack-infra/config: Put the trusty test in the Ubuntu section https://review.openstack.org/104632 | 18:03 |
| clarkb | afazekas: every node has files in /etc/nodepool telling it what its role is and where the otehr nodes are located | 18:03 |
| mordred | jeblair, clarkb, nibalizer ^^ enjoy that one | 18:03 |
| clarkb | afazekas: http://ci.openstack.org/nodepool/scripts.html#scripts | 18:03 |
| jesusaurus | mordred: wow | 18:03 |
| mordred | jesusaurus: right? :) | 18:03 |
| clarkb | I am amused by the error in http://logs.openstack.org/48/104448/5/check/check-grenade-dsvm-partial-ncpu/6b93538/console.html we use a fake virt driver | 18:03 |
| clarkb | how does that error? | 18:03 |
| afazekas | clarkb: So at the beginig I will be the jnkins user, how can I ssh the other nodes ? | 18:04 |
| jesusaurus | mordred: im guessing thats my bad | 18:04 |
| * anteaya is on lagging wifi at the library | 18:04 | |
| mordred | jesusaurus: I may have done it too | 18:04 |
| jesusaurus | from the slave refactor stuff | 18:04 |
| mordred | oh - neat. I'll blame you | 18:04 |
| * afazekas read the fine manual | 18:04 | |
| clarkb | afazekas: that page I linked you lists the files in /etc/nodepool. | 18:04 |
| afazekas | :) | 18:04 |
| clarkb | afazekas: the keys are in there | 18:04 |
| afazekas | MTU size ? | 18:04 |
| afazekas | Nic number ? | 18:04 |
| Ajaeger | jeblair: Ok, will check later. Thanks! | 18:05 |
| clarkb | afazekas: why do those matter? | 18:05 |
| *** zz_dimtruck is now known as dimtruck | 18:05 | |
| afazekas | for GRE or VXLAN if the external net has lower than 1500 byte mtu, I need to hack it lower in the jobs | 18:05 |
| clarkb | afazekas: you may have to discover that as it will probably vary between rax and hpcloud | 18:05 |
| *** amcrn has joined #openstack-infra | 18:05 | |
| clarkb | I don't know why nic number matters | 18:05 |
| clarkb | rax nodes get two nics | 18:06 |
| clarkb | hpcloud nodes get one nic | 18:06 |
| fungi | clarkb: well, we affinity subnodes into the same nodepool provider as the main node | 18:06 |
| clarkb | fungi: right but I don't want to ahrdcode a value that is in theory discoverable | 18:06 |
| fungi | clarkb: so they should be local-ish depending on the topology within the provider | 18:06 |
| *** pcm_ has quit IRC | 18:06 | |
| afazekas | clarkb: I assume first it should be created as an experimental job | 18:06 |
| fungi | clarkb: agreed. the job should figure out what it needs | 18:07 |
| clarkb | afazekas: yes that is probably the best place to start | 18:07 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 18:07 | |
| jeblair | yeah, i think we can assume that for a given run, the values will be the same -- but we should not assume a given run will happen on one or the other | 18:07 |
| clarkb | jeblair: yup | 18:07 |
| *** pcm_ has joined #openstack-infra | 18:07 | |
| fungi | ahh, you said "vary between providers" not "communicate between providers" (i skimmed poorly, sorry!) | 18:07 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New StoryBoard Logo https://review.openstack.org/99217 | 18:08 |
| afazekas | First may be we should start with simple topology, when the main node has all services (may be excluding compute) and all extra nodes (based on the configured number) has the same 2th set of services (compute with neutron and ceilometer agents) | 18:08 |
| mordred | the info should all be in teh node metadata | 18:09 |
| *** MarkAtwood has joined #openstack-infra | 18:09 | |
| clarkb | afazekas: I think sdague wants all services on one node including compute and compute on the other node | 18:09 |
| *** whoops has quit IRC | 18:10 | |
| *** liusheng has quit IRC | 18:10 | |
| jeblair | mordred: should nodepool write out ansible compatible node lists? | 18:10 |
| jeblair | (was thinking that "ansible subnode* do-something" would be a nice interface) | 18:11 |
| *** liusheng has joined #openstack-infra | 18:11 | |
| jeblair | (for multinode testing) | 18:11 |
| afazekas | clarkb: Can we use f20 nodes ? | 18:12 |
| clarkb | afazekas: I think trusty would be preferred | 18:12 |
| sdague | afazekas: yeh, trusty definitely prefered | 18:12 |
| clarkb | as we have to throw away f20 in the not so distant future | 18:12 |
| clarkb | whereas work done on trusty is good for years | 18:12 |
| mordred | jeblair: ooh. that would be neat | 18:12 |
| sdague | I don't want to have to spend time debugging distro differences between multinode and tempest full single node | 18:12 |
| clarkb | sdague: is http://logs.openstack.org/48/104448/5/check/check-grenade-dsvm-partial-ncpu/6b93538/logs/old/screen-n-cpu.txt.gz#_2014-07-03_17_15_26_819 a known thing? | 18:13 |
| clarkb | sdague: ++ | 18:13 |
| *** devkulkarni has joined #openstack-infra | 18:13 | |
| sdague | clarkb: looking | 18:13 |
| mordred | jeblair: turns out, the list can just be a plain list in a file - but also additionally can have groupings | 18:13 |
| sdague | clarkb: no, that's new to me | 18:13 |
| clarkb | sdague: cool I will file a nova bug | 18:13 |
| clarkb | or search over there then do that | 18:13 |
| jeblair | mordred: what's the path for the default node list? | 18:14 |
| mordred | jeblair: http://paste.openstack.org/show/85442/ | 18:14 |
| mordred | jeblair: /etc/ansible/hosts | 18:14 |
| jeblair | no ansible on precise, right? | 18:15 |
| fungi | jeblair: what's the reason for having some zuul test layouts in tests/fixtures and some in tests/fixtures/layouts? is there a semantic distinction, or just historical and never reorganized? | 18:16 |
| mordred | jeblair: pip install ansible works fine | 18:16 |
| jeblair | fungi: tests/fixtures/layouts is for the layout validation tests; tests/fixtures is for the rest of the tests | 18:18 |
| openstackgerrit | A change was merged to openstack-infra/config: Put the trusty test in the Ubuntu section https://review.openstack.org/104632 | 18:18 |
| mordred | jeblair: actually, ansible 1.4.4 is available through apt on precise | 18:18 |
| clarkb | sdague: https://bugs.launchpad.net/nova/+bug/1337472 | 18:18 |
| uvirtbot | Launchpad bug 1337472 in nova "UnboundLocalError: local variable 'domain' referenced before assignment" [Undecided,New] | 18:18 |
| afazekas | clarkb: log collection: node1 node2 noden sub-folder for node specific logs sounds ok ? | 18:19 |
| fungi | jeblair: makes sense, thanks | 18:19 |
| clarkb | afazekas: ya that should work | 18:19 |
| *** doddstack has quit IRC | 18:19 | |
| mordred | jeblair: it doesn't look like the 1.5 features would matter to us for nodepool actions | 18:19 |
| Ajaeger | jeblair: yes, proposal slaves are updated with the changes I've been looking for! | 18:20 |
| afazekas | clarkb: first we should start with serial install , than we can try to run devstack in parallel on each node | 18:20 |
| mordred | jeblair: main changes that matter in 1.5 are ssh pipelining and improved speed in openssh transport, as well as not needing to add localhost to hosts to get local_action commands to work properly | 18:20 |
| mordred | jeblair: I don't think those would be a problem for us | 18:20 |
| *** MarkAtwood has quit IRC | 18:21 | |
| clarkb | afazekas: ya, I think that is fine. We probably won't make this voting until it is parallel though simply due to time involved | 18:21 |
| clarkb | afazekas: but it will be much easier to boostrap the serial case | 18:21 |
| clarkb | and I think to debug any issues the come up too | 18:21 |
| greghaynes | /window splith | 18:22 |
| greghaynes | wooops | 18:22 |
| clarkb | greghaynes: there lies the path to ww | 18:22 |
| greghaynes | fact | 18:22 |
| *** Ryan_Lane has joined #openstack-infra | 18:22 | |
| *** Ryan_Lane has joined #openstack-infra | 18:22 | |
| clarkb | dansmith: https://bugs.launchpad.net/nova/+bug/1337472 maybe of interest to you as it involves that upgrade path you and jogo sorted out and are testing with taht test | 18:23 |
| uvirtbot | Launchpad bug 1337472 in nova "UnboundLocalError: local variable 'domain' referenced before assignment" [Undecided,New] | 18:23 |
| afazekas | clarkb: So the devstack-gate can be used for adding the multi-node scripts, right ? | 18:23 |
| clarkb | afazekas: yup | 18:23 |
| *** maxbit_ has quit IRC | 18:23 | |
| clarkb | afazekas: initially we can probably get away with the pre_test hook and post_test hook | 18:23 |
| dansmith | clarkb: the day before a holiday, very few things interest me | 18:24 |
| afazekas | clarkb: thx | 18:24 |
| clarkb | afazekas: but at some point we will probably want to merge stuff back into d-g | 18:24 |
| clarkb | dansmith: well ok | 18:24 |
| *** MarkAtwood has joined #openstack-infra | 18:24 | |
| dansmith | clarkb: that further doesn't interest me because it looks like a bug | 18:24 |
| clarkb | dansmith: are stocked up and ready to blow things up? | 18:24 |
| clarkb | dansmith: you live on the westside which basically turns into mortars mortars everywhere while TVFR runs around fining people | 18:25 |
| clarkb | dansmith: yes I would say it is a bug :) | 18:25 |
| dansmith | clarkb: heh, I don | 18:25 |
| dansmith | clarkb: I don't generally participate in the blowing up of things | 18:26 |
| dansmith | clarkb: I'm too busy scolding my neighbors for being idiots | 18:26 |
| *** otherwiseguy has joined #openstack-infra | 18:26 | |
| *** mostly_d34dh0r53 is now known as d34dh0r53 | 18:26 | |
| jeblair | afazekas, clarkb: the more you can do in devstack-gate itself, the less you need to run through a config review/merge/update/test cycle | 18:27 |
| *** arnaud__ has joined #openstack-infra | 18:27 | |
| jeblair | afazekas, clarkb: since devstack-gate changes are self testing | 18:27 |
| clarkb | jeblair: good point | 18:28 |
| *** mburned_out has quit IRC | 18:28 | |
| clarkb | afazekas: so the experimental job should probably be added to devstack-gate, devstack, and tempest to start? | 18:28 |
| jeblair | so probably an experimental job that sets up the basics, but you'll probably want to keep most of the actual work inside devstack-gate for ease of development | 18:28 |
| *** mfer has quit IRC | 18:28 | |
| *** doude has quit IRC | 18:28 | |
| clarkb | dansmith: when I lived there I would just sit out in the driveway and watch them try to burn down the neighborhood | 18:30 |
| clarkb | dansmith: much fun was had | 18:30 |
| dansmith | clarkb: heh | 18:30 |
| *** MarkAtwood has quit IRC | 18:31 | |
| *** gondoi is now known as zz_gondoi | 18:31 | |
| *** nati_ueno has quit IRC | 18:32 | |
| ttx | clarkb: o/ | 18:32 |
| ttx | need any clarification from me ? | 18:32 |
| clarkb | ttx: maybe you can rereview the change, but it should handle all the cases now | 18:33 |
| clarkb | ttx: it is currently working through testing and will hopefully merge at some point | 18:33 |
| ttx | link? | 18:33 |
| clarkb | ttx: https://review.openstack.org/#/c/104448/ | 18:33 |
| *** MarkAtwood has joined #openstack-infra | 18:33 | |
| ttx | I thought about it again, and actually my use case (proposed/juno-2 is not really valid | 18:34 |
| *** zz_gondoi is now known as gondoi | 18:34 | |
| *** mburned_out has joined #openstack-infra | 18:35 | |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/config: Removing unused jenkins_slave script https://review.openstack.org/104642 | 18:35 |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/config: Removing extra parens https://review.openstack.org/104643 | 18:36 |
| ttx | clarkb: I think that should work | 18:36 |
| wenlock | question: jenkins::plugin doesn't seem to support updating a plugin by removing the version that was previously there when it doesn't match the version being installed. is there something i'm missing, or is that just not something that is desirable? | 18:38 |
| clarkb | wenlock: somewhat not desireable. upgrading plugins requires a ton of orchestration | 18:38 |
| wenlock | clarkb : remove the folder and file, is there more to it? | 18:39 |
| clarkb | wenlock: restart jenkins | 18:39 |
| wenlock | clarkb right subscribe service jenkins | 18:39 |
| *** arnaud__ has quit IRC | 18:40 | |
| clarkb | wenlock: that won't work. you will kill all your jobs everytime you make a change | 18:40 |
| clarkb | wenlock: instead you want to put the node in shutdown mode, wait for all jobs to finish, remove any slaves that are derpy because jenkins will just readd them after a restart, then stop jenkins, check that it actually stopped, then start jenkins | 18:40 |
| wenlock | clarkb gotcha and you want to schedule that | 18:41 |
| clarkb | yup | 18:41 |
| clarkb | wenlock: and when you have more than one master you want to do it in a rolling fashion | 18:42 |
| wenlock | clarkb so when you update a plugin, you also manually remove it | 18:42 |
| * ttx disappears | 18:42 | |
| wenlock | on all masters | 18:42 |
| clarkb | wenlock: yes | 18:42 |
| wenlock | clarkb ok thanks for the explain on that, makes sense for a very large environment. I think smaller environments it's probably non issue sense builds arn't non-stop, and rechecks are cheap | 18:43 |
| *** Ajaeger has quit IRC | 18:44 | |
| *** maxbit has joined #openstack-infra | 18:44 | |
| *** MarkAtwood has quit IRC | 18:44 | |
| wenlock | clarkb plus cant we just update the plugin and schedule the restart for loading the new plugin later | 18:44 |
| clarkb | wenlock: sure. but that doesn't change the need for orchestration | 18:45 |
| clarkb | wenlock: and may make it hard to actually know when things will converge on a version | 18:45 |
| clarkb | wenlock: but jenkins will handle that properly | 18:45 |
| zaro | anybody know what's this error is about, invalid command 'bdist_wheel'? https://jenkins05.openstack.org/job/python-jenkins-tarball/1/console | 18:46 |
| *** e0ne has quit IRC | 18:46 | |
| wenlock | zaro missing that command from the os? | 18:46 |
| fungi | zaro: might be that you can't actually build and upload wheels that way until after you switch to pbr | 18:47 |
| zaro | ohh, ok. | 18:48 |
| *** e0ne has joined #openstack-infra | 18:48 | |
| anteaya | my headache is gone!! | 18:48 |
| anteaya | woooooo | 18:48 |
| anteaya | yay | 18:48 |
| *** mfer has joined #openstack-infra | 18:48 | |
| clarkb | anteaya: yay | 18:48 |
| anteaya | thanks | 18:48 |
| clarkb | anteaya: is it still noah like outside? | 18:48 |
| wenlock | anteaya happy times woot | 18:48 |
| fungi | zaro: so we might need a tarball-only job instead of the one which tries to build tarballs and wheels | 18:48 |
| anteaya | clarkb: no, just mist | 18:48 |
| anteaya | clarkb: but I had 3 bobcats, 1 mini-x, a tree chipper and 3 chainsaws at home | 18:49 |
| anteaya | so I had to run away | 18:49 |
| anteaya | folks are fixing driveways and removing fallen trees (in addition to building houses) | 18:49 |
| zaro | fungi: so this is a build of python-jenkins, which will be the last one before giving it pbr. it's already been tagged. | 18:49 |
| anteaya | library is better | 18:49 |
| zaro | fungi: what do you suggest? | 18:49 |
| anteaya | wenlock: thanks | 18:49 |
| clarkb | zaro: fungi: may be simplest to just manually upload an sdist? or am I giving up too quickly? | 18:50 |
| fungi | zaro: i can manually trigger the upload job which will upload the tarball and then fail to find and upload the wheel, but should sort itself out | 18:50 |
| fungi | dhellmann: mordred: ^ unforseen consequence of unconditionally adding wheel building to the tarball job | 18:51 |
| zaro | fungi: can't upload without having the tarball first right? | 18:51 |
| fungi | zaro: yeah, but i think the tarball already got put on the tarballs site didn't it? | 18:51 |
| * fungi checks | 18:51 | |
| fungi | no, i guess not | 18:51 |
| fungi | hrm... we might want to fix that problem and then retrigger the tarball job, followed by the upload job | 18:52 |
| dhellmann | fungi: why don't we want a wheel of that? | 18:52 |
| * dhellmann doesn't know what python-jenkins is | 18:52 | |
| zaro | fungi: i think what happened is that it built (not correct) version of tarball. in the logs | 18:52 |
| fungi | dhellmann: well, the packaging on it doesn't know how to build a wheel (yet) | 18:52 |
| dhellmann | ah | 18:53 |
| *** otter768 has joined #openstack-infra | 18:53 | |
| fungi | this is a release immediately prior to pbr-ization | 18:53 |
| zaro | fungi: the job built version 2.1 | 18:53 |
| dhellmann | fungi: when we did the work to add the wheel thing I think you and I talked about splitting those steps into different jobs, or at least making wheels optional in some way | 18:53 |
| zaro | fungi: I tagged it as 0.3.0 so i thought it woudl build that version. | 18:53 |
| fungi | zaro: ahh, so it did. i bet because it doesn't use pbr and so didn't look at the tag to determine its version | 18:54 |
| zaro | ahh | 18:54 |
| zaro | not sure what to do now. | 18:54 |
| fungi | zaro: so in actuality, what you'll need to do is edit whatever in python-jenkins identifies the version (perhaps in setup.cfg or setup.py) and commit that, then tag it as 3.0.1 | 18:55 |
| *** MarkAtwood has joined #openstack-infra | 18:55 | |
| fungi | oh, except new complication there | 18:56 |
| fungi | you tagged a historical commit, so you'd need a separate branch or several reverts first | 18:56 |
| fungi | since master is now several commits ahead of what you're going to want to modify and release | 18:57 |
| fungi | s/3.0.1/0.3.1/ or whatever | 18:57 |
| *** sweston has quit IRC | 18:57 | |
| zaro | fungi: i think there is already a job to build from a branch so i'll go the branch route. | 18:58 |
| fungi | zaro: in that case we'll need to create a branch named something like 0.3-fixes from the same commit as the 0.3.0 tag and then update it with the commit which adjusts the version in the source to 0.3.1 | 18:59 |
| fungi | and we probably also need to add some tarball-only jobs | 19:00 |
| zaro | fungi: just to confirm the steps. 1. create branch, 2. patch branch with version info, 3. create a tag from the branch, 4. build will fail on wheels, 5. you manually upload tarball to pypi | 19:00 |
| fungi | zaro: well, it actually looks like the tarball job failed before it ever even tried to upload anything to pypi anyway | 19:01 |
| *** james_li_ has joined #openstack-infra | 19:01 | |
| fungi | because the wheel couldn't be built, the job aborted before even sticking the tarball on the tarballs site | 19:01 |
| *** pelix has quit IRC | 19:02 | |
| zaro | fungi: did you mean pypi or tarballs.o.o? | 19:02 |
| fungi | so i think we need to make some adjustments to the jobs, or make some new jobs | 19:02 |
| fungi | zaro: the pypi upload job is dependent on the tarball job, and the tarball job aborted before uploading anything at all to tarballs.openstack.org | 19:03 |
| *** MarkAtwood has quit IRC | 19:03 | |
| *** james_li has quit IRC | 19:04 | |
| *** james_li_ is now known as james_li | 19:04 | |
| *** mburned_out is now known as mburned | 19:04 | |
| zaro | fungi: ohh i see, upload to tarballs doesn't happen until wheels thing completes. | 19:05 |
| *** otter768 has quit IRC | 19:06 | |
| *** ociuhandu has quit IRC | 19:06 | |
| fungi | right, we could switch the ordering in the tarball job so that it uploads the tarball after it builds it, then tries to build the wheel | 19:06 |
| fungi | but i think that's just papering over the real issue | 19:06 |
| *** arnaud__ has joined #openstack-infra | 19:07 | |
| zaro | fungi: ok. will do the following: 1. create branch, 2. patch branch with version info, 3. create new jobs to build tarball only, 4. create a tag from the branch, 5. job runs and tarball gets uploaded to pypi | 19:07 |
| *** zzelle has joined #openstack-infra | 19:07 | |
| fungi | which is that the tarball job either needs to know that it shouldn't build a wheel in some cases, or we need a variant of it which only does tarballs | 19:07 |
| *** markwash has quit IRC | 19:08 | |
| *** mburned has left #openstack-infra | 19:08 | |
| fungi | zaro: yep, and then 6. get one of the gerrit admins to delete the old 0.3-fixes branch | 19:08 |
| *** doddstack has joined #openstack-infra | 19:08 | |
| *** dims_ has quit IRC | 19:09 | |
| clarkb | woo the normal grenade test failed this time | 19:09 |
| fungi | speaking of releases, git-review 1.24, complete with wheel! https://pypi.python.org/pypi/git-review (time to send the release announcement now) | 19:09 |
| anteaya | woooo | 19:09 |
| clarkb | this time ssh timed out so at least it wasn't the same fail | 19:10 |
| *** anode has quit IRC | 19:10 | |
| clarkb | fungi: \o/ | 19:10 |
| fungi | example install from wheel on pypi... http://paste.openstack.org/show/85447/ | 19:11 |
| notmyname | clarkb: https://review.openstack.org/#/c/104448/ failed grenade (other tests still runnin) | 19:11 |
| clarkb | notmyname: ya | 19:11 |
| clarkb | notmyname: because ssh failed | 19:11 |
| notmyname | clarkb: ah. you're on top of it :-) | 19:11 |
| clarkb | recheck bug 1298472 is what we will want | 19:12 |
| uvirtbot | Launchpad bug 1298472 in nova "SSHTimeout in tempest trying to verify that computes are actually functioning" [Critical,Confirmed] https://launchpad.net/bugs/1298472 | 19:12 |
| openstackgerrit | Edward Raigosa proposed a change to openstack-infra/config: provide an option in plugin to pre-update a plugin with puppet https://review.openstack.org/104652 | 19:12 |
| zaro | fungi: i don't have permission to create a branch for python-jenkins. would you be able to help? | 19:15 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New StoryBoard Logo https://review.openstack.org/99217 | 19:15 |
| fungi | zaro: sure, though if you expect to do this in the future it could also have its acl extended to allow branch creation by the -ptl group for it | 19:15 |
| anteaya | fungi: git review -u is just a git review command, yeah? | 19:16 |
| fungi | zaro: is 3.0-fixes an acceptable name, or do you need something else there? | 19:16 |
| zaro | fungi: wfm | 19:17 |
| *** shashankhegde has quit IRC | 19:17 | |
| *** arnaud__ has quit IRC | 19:18 | |
| *** mfer has quit IRC | 19:18 | |
| fungi | zaro: done | 19:20 |
| fungi | anteaya: yeah, the manpage for git-review looks like it could use some love though (maybe we should be updating it with the output of git-review --help since git's behavior of displaying the manpage on git review --help masks the usage details) | 19:21 |
| clarkb | I like that | 19:21 |
| fungi | also the troff formatting in the current manpage is atrocious | 19:21 |
| clarkb | might also reduce confusion | 19:21 |
| anteaya | fungi: yes I used git review --help to find -u | 19:22 |
| anteaya | where do I find man for getting the latest git review? | 19:22 |
| fungi | please rephrase... parsing error | 19:22 |
| anteaya | sorry | 19:23 |
| anteaya | lagging wifi here | 19:23 |
| anteaya | I want the new git review | 19:23 |
| anteaya | what commands do I run to get it | 19:23 |
| *** mfer has joined #openstack-infra | 19:23 | |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Renamed module folders for consistency https://review.openstack.org/103182 | 19:24 |
| wenlock | pip install -U git-review | 19:24 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Tag Complete directive has proper focus behavior https://review.openstack.org/103183 | 19:24 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Moved template files into related modules https://review.openstack.org/103181 | 19:24 |
| fungi | anteaya: it depends on how you currently run git-review. i personally avoid pip installing anything into the global system context of my machines, so i make a dedicate virtualenv for git-review and pip install it in there, then i make sure my environment is set up with ~/bin in my execution path and i add a symlink from ~/bin/git-review to the git-review executable in the virtualenv | 19:25 |
| anteaya | ah | 19:25 |
| anteaya | I don't do that | 19:25 |
| fungi | anteaya: though to link the manpage, you have to do a similar dance with the manpath in your environment | 19:25 |
| anteaya | awesome | 19:26 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Added focus directive https://review.openstack.org/103345 | 19:26 |
| anteaya | I think I use pip without the venv | 19:26 |
| fungi | anteaya: though in theory if you're not running git-review from distro packages you can just 'sudo pip install git-review' and it should simply work | 19:26 |
| wenlock | anteaya yep, cloud account with ubuntu does the trick for me | 19:27 |
| fungi | or more likely sudo pip install -U git-review since you may already have an older one pip-installed | 19:27 |
| anteaya | okay thanks | 19:27 |
| clarkb | ok I rechecked 104448 going to afk now for lunch | 19:27 |
| clarkb | will check in on it later | 19:27 |
| clarkb | also elasticsearch :( why | 19:27 |
| anteaya | thanks, suod with -U worked | 19:28 |
| clarkb | just looking at graphs it seems like it should be healthy but something is off | 19:28 |
| anteaya | wenlock: I'm just on my work laptop | 19:28 |
| *** e0ne has quit IRC | 19:28 | |
| *** _nadya_ has joined #openstack-infra | 19:29 | |
| anteaya | sounds like it is only a matter of time before it collapes due to conflicts | 19:29 |
| * pleia2 updates too | 19:29 | |
| *** maxbit has quit IRC | 19:30 | |
| fungi | anteaya: one up-side to using a virtualenv for something like git-review is you can install many versions of it side-by-side and test/compare them fairly painlessly | 19:30 |
| anteaya | ah | 19:30 |
| fungi | harder to do when you install it system-wide | 19:30 |
| anteaya | sounds like a to do for my next hard drive | 19:30 |
| anteaya | yes | 19:31 |
| wenlock | just updated to, went fast (system wide) :D | 19:31 |
| *** mfer has quit IRC | 19:31 | |
| anteaya | so far for me, it outputs the version number as expected | 19:31 |
| fungi | this way i can just 'pip install .' into a virtualenv from my current git checkout and then move my symlink to that one (or keep the venv activated) to test proposed changes | 19:31 |
| anteaya | right | 19:31 |
| anteaya | makes sense | 19:31 |
| anteaya | it is 3:30pm here, I'm trying to figure out what time the noisemakers will decide they have had enough and go home | 19:32 |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New Story Autofocus https://review.openstack.org/104342 | 19:33 |
| *** _nadya_ has quit IRC | 19:33 | |
| openstackgerrit | Dan Prince proposed a change to openstack-infra/nodepool: Add support for network labels https://review.openstack.org/98287 | 19:33 |
| pleia2 | anteaya: yeesh, isn't canada day over? :) | 19:34 |
| pleia2 | we're just starting the noise around here, big fireworks show at the ballpark last night | 19:35 |
| anteaya | pleia2: long ago | 19:37 |
| anteaya | see that i have no issue with | 19:37 |
| anteaya | announced with times and date | 19:37 |
| anteaya | you don't want to listen, you have lots of warneing to do something elese | 19:38 |
| * Mithrandir waves at pleia2 | 19:38 | |
| *** deny__ has joined #openstack-infra | 19:38 | |
| pleia2 | Mithrandir: hey, home safe I assume? :) | 19:39 |
| pleia2 | Mithrandir: chatted with paulproteus the other day, he's super excited about CI stuff so I don't think we'll be wanting for informal chats/topics at debconf | 19:39 |
| Mithrandir | pleia2: coolie. Got home safe and sound without any problems. | 19:40 |
| Mithrandir | trying to figure out flights to debconf now. | 19:40 |
| * pleia2 nods | 19:40 | |
| Mithrandir | I can choose between tedious, tedious and otherwise tedious. | 19:40 |
| Mithrandir | this planet's pretty big | 19:40 |
| pleia2 | yeah | 19:40 |
| pleia2 | I haven't decided what days to attend yet, but I'll probably book in a couple weeks (short flight for me though) | 19:40 |
| anteaya | I think I am going to try going hom e now | 19:41 |
| pleia2 | anteaya: good luck! | 19:41 |
| anteaya | thanks | 19:42 |
| anteaya | I'm so tired of noise, I am shaking and nautious | 19:42 |
| *** deny_ has quit IRC | 19:42 | |
| pleia2 | yeah, hope it gets better :( | 19:42 |
| *** enikanorov_ has joined #openstack-infra | 19:43 | |
| adalbas | ArxCruz, anteaya : the version of setuptools installed by pip is causing it to break when installing prettytable. I see there is an option PIP_NO_SETUPTOOLS to use for devstack on localrc | 19:44 |
| *** enikanorov__ has quit IRC | 19:47 | |
| *** pblaho has joined #openstack-infra | 19:49 | |
| *** blamar has joined #openstack-infra | 19:51 | |
| *** dims_ has joined #openstack-infra | 19:53 | |
| openstackgerrit | Khai Do proposed a change to openstack-infra/config: Give PTLs permission to create branches https://review.openstack.org/104663 | 19:54 |
| openstackgerrit | Dolph Mathews proposed a change to openstack-infra/elastic-recheck: add query for dox publish failure https://review.openstack.org/104664 | 19:54 |
| *** jdahlin has quit IRC | 19:57 | |
| *** cnesa2 has joined #openstack-infra | 20:01 | |
| *** gondoi is now known as zz_gondoi | 20:03 | |
| *** arnaud has joined #openstack-infra | 20:03 | |
| *** blamar has quit IRC | 20:03 | |
| *** dkehn__ has joined #openstack-infra | 20:04 | |
| *** otherwiseguy has quit IRC | 20:06 | |
| *** dprince has quit IRC | 20:07 | |
| *** dkehn_ has quit IRC | 20:07 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/git-review: Build universal wheels https://review.openstack.org/104667 | 20:08 |
| *** pblaho has quit IRC | 20:08 | |
| fungi | mordred: ^ sane? | 20:08 |
| *** pblaho has joined #openstack-infra | 20:08 | |
| *** zz_gondoi is now known as gondoi | 20:08 | |
| fungi | i tested the hypothesis with 2.7 and 3.3 venvs | 20:09 |
| *** mmaglana has quit IRC | 20:11 | |
| *** devkulkarni has quit IRC | 20:11 | |
| *** rcarrillocruz has joined #openstack-infra | 20:12 | |
| *** rcarrill` has quit IRC | 20:14 | |
| *** dimtruck is now known as zz_dimtruck | 20:15 | |
| *** zz_dimtruck is now known as dimtruck | 20:16 | |
| *** sweston has joined #openstack-infra | 20:18 | |
| *** Sukhdev has joined #openstack-infra | 20:18 | |
| *** dkehn__ is now known as dkehnx | 20:19 | |
| *** zz_johnthetubagu is now known as johnthetubaguy | 20:20 | |
| *** pblaho has quit IRC | 20:21 | |
| *** pcm_ has quit IRC | 20:21 | |
| *** james_li_ has joined #openstack-infra | 20:23 | |
| *** james_li has quit IRC | 20:26 | |
| *** james_li_ is now known as james_li | 20:26 | |
| *** harlowja_at_home has joined #openstack-infra | 20:26 | |
| openstackgerrit | Sergey Skripnick proposed a change to openstack-infra/config: Publish generated html tests results https://review.openstack.org/104677 | 20:26 |
| *** bnemec_appt is now known as bnemec | 20:27 | |
| mordred | fungi: yup | 20:28 |
| fungi | mordred: thanks! | 20:28 |
| *** jgrimm has quit IRC | 20:28 | |
| fungi | clarkb: notmyname: 104448,5 seems to have found its way into the gate finally... current eta is 48 minutes (hopefully the nova change ahead of it doesn't cause delays) | 20:29 |
| *** sandywalsh has quit IRC | 20:29 | |
| *** jaypipes has quit IRC | 20:29 | |
| *** sandywalsh has joined #openstack-infra | 20:30 | |
| clarkb | woot | 20:31 |
| JayF | So I'm going to write a patch, but it looks like our IPA builder put the image in a strange place (http://tarballs.openstack.org/ironic-python-agent/coreos/ipa-coreos.tar.gz/ipa-coreos.tar.gz) -- I want to clean up the infra config so it doesn't put ipa-coreos.tar.gz inside a dir named ipa-coreos.tar.gz, but I think I'll need that dir removed first as dirs and files can't share a name | 20:31 |
| *** marun_afk is now known as marun | 20:32 | |
| clarkb | JayF: I can do that | 20:33 |
| clarkb | JayF: but maybe you should update the job first otherwise it will just redo the same thing | 20:33 |
| JayF | clarkb: wait until my patch is ready to merge | 20:33 |
| JayF | yeah, exactly :) | 20:33 |
| clarkb | ya | 20:33 |
| *** denis_makogon has joined #openstack-infra | 20:33 | |
| openstackgerrit | Jay Faulkner proposed a change to openstack-infra/config: Fix upload to not have a redundant directory https://review.openstack.org/104681 | 20:35 |
| JayF | clarkb: ^ I believe that should do the trick | 20:35 |
| JayF | clarkb: otherwise it built and worked great, thanks for the suggestion to do it first-party and providing the bare-trusty image | 20:35 |
| clarkb | cool | 20:35 |
| clarkb | fungi: mordred jeblair ^ want to give that a relatively quick review so that I can do the corresponding update on static.o.o? | 20:36 |
| jeblair | ack | 20:36 |
| *** anode has joined #openstack-infra | 20:36 | |
| * JayF is very happy to have ipa images building | 20:37 | |
| jeblair | aprvd | 20:37 |
| fungi | lgtm but not approving since it sounds like jeblair's looking too | 20:37 |
| fungi | and there it goes ;) | 20:37 |
| clarkb | thanks! | 20:37 |
| JayF | TYVM guys :) | 20:38 |
| clarkb | JayF: how does that look now? | 20:39 |
| clarkb | I moved the existing tarball to where it should have gone | 20:40 |
| JayF | oh hey, that's great | 20:40 |
| JayF | that's even better than I expected! I thought I'd have to kick off a fake build :) | 20:40 |
| *** mbacchi has quit IRC | 20:41 | |
| openstackgerrit | A change was merged to openstack-infra/config: Fix upload to not have a redundant directory https://review.openstack.org/104681 | 20:43 |
| *** ihrachyshka has joined #openstack-infra | 20:44 | |
| *** fabo has quit IRC | 20:47 | |
| clarkb | jeblair: before you long weekend, did yo usee jhesketh's changes for swift uploads? | 20:47 |
| *** fabo has joined #openstack-infra | 20:48 | |
| jeblair | clarkb: no | 20:48 |
| clarkb | jeblair: https://review.openstack.org/#/c/104437/ and https://review.openstack.org/#/c/104453/ and https://review.openstack.org/#/c/104451/ | 20:48 |
| clarkb | the last one is probably the most pressing one as it will help us continue to test. But the other two will need to go in at some point | 20:48 |
| clarkb | for the one that adds a wsgi.conf I didn't know if we wanted to use a ro account for that access | 20:49 |
| clarkb | fungi: ^ | 20:49 |
| clarkb | I thought we did | 20:49 |
| fungi | clarkb: yeah, i thought i created two... one for the log display server (which was ro) and one for the artifact uploads (which was rw) | 20:50 |
| * clarkb looks at hiera | 20:50 | |
| bodepd | I missed the notification that you guys were asking me about beaker | 20:50 |
| bodepd | happy to say what I know although others know way more | 20:51 |
| clarkb | fungi: yup its in there I will go ahead and amend that change for jhesketh_ | 20:51 |
| bodepd | it ties node definitions into provisioning systems and has rspec syntax to refer to actions to run on those nodes | 20:51 |
| fungi | clarkb: coolies | 20:51 |
| bodepd | but it uses serverspec to do the *stuff* | 20:51 |
| clarkb | fungi: jeblair: the really cool thing is the upload worked | 20:51 |
| clarkb | so we are very very close to having this working | 20:51 |
| *** devkulkarni has joined #openstack-infra | 20:52 | |
| jeblair | bodepd: thanks; i think with all the machinery we have around provisioning, i'm mostly instered in just doing "the stuff" | 20:53 |
| jeblair | interested, even. | 20:53 |
| *** signed8bit is now known as signed8bit_ZZZzz | 20:53 | |
| *** ildikov has joined #openstack-infra | 20:53 | |
| *** otherwiseguy has joined #openstack-infra | 20:54 | |
| fungi | i find it amusing that now when lp times out on bug status changes through the webui, half the time it sends you to a "private" page for which you lack viewing permission | 20:54 |
| fungi | "Sorry, you don't have permission to access this page or the information in this page is not shared with you." | 20:54 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Ensure there is a slash in the object prefix https://review.openstack.org/104451 | 20:55 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Add in wsgi.conf for os-loganalyze https://review.openstack.org/104437 | 20:55 |
| clarkb | fungi: jeblair ^ that should do it | 20:55 |
| fungi | lookin | 20:55 |
| bodepd | jeblair: yep, that is called serverspec :) | 20:55 |
| mordred | jeblair: I'm more instered myself | 20:55 |
| bodepd | jeblair: if I have time I want to wrap my head and heat based provisioning and serverspec | 20:56 |
| *** mestery_ has joined #openstack-infra | 20:56 | |
| jeblair | bodepd: cool thanks; mordred: then you should probably inster serverspec | 20:56 |
| bodepd | jeblair: and decide whether or not beaker has a place in a heat based rspec framework for serverspec | 20:56 |
| jeblair | bodepd: sounds reasonable | 20:56 |
| fungi | in good news, i found where the zuul keyerror was originating. docs-drafts failure message patterns dig deep into build.parameters and look for keys which may not exist at all. probably need to have some overloaded get for those which return a default value | 20:57 |
| bodepd | jeblair: I've got reusalble heat templates for Puppet. | 20:57 |
| jeblair | fungi: you are the keymaster | 20:57 |
| fungi | nah, that's vinz! | 20:57 |
| bodepd | but I have some concerns about heat as a provisioning solution in general... | 20:57 |
| jeblair | i like that lsb_release behaves differently on different platforms | 20:58 |
| fungi | (except on non-lsb platforms, where it behaves very consistently!) | 20:59 |
| zaro | fungi: i'm finding it to be a lot of work to setup jobs for this one time python-jenkins fix. do you think we need to setup? | 21:01 |
| clarkb | a lot of work? wouldn;t the jobs be the same/ | 21:02 |
| bodepd | actually. having thought about it. beaker does allow for ssh into provisioned machines, and you still need something to orchestrate serverspec (at least in as much as I understand it) | 21:02 |
| zaro | clarkb: python-jenkins uses the jobs in python-jobs.yaml | 21:02 |
| clarkb | right those should just work? | 21:03 |
| zaro | clarkb: those jobs are common across most of the projects. | 21:03 |
| fungi | zaro: well, the question we need to ask is whether we want to still be able to use the python tarball jobs to build sdists for projects which don't have the mechanisms needed to build wheels | 21:03 |
| jeblair | zaro: this is just for the pre-pbr release of python-jenkins? | 21:03 |
| zaro | jeblair: yes | 21:03 |
| fungi | we can work around it by manually uploading an sdist for python-jenkins, but it's probably not the only project in that boat | 21:03 |
| jeblair | fungi: yeah, that's what i'd do | 21:04 |
| zaro | clarkb: right now 'tarball' jobs are doing tarball and wheels. so what would you call a job that's exclusive to just tarball? | 21:04 |
| jeblair | fungi: no openstack projects are in that boat; we don't need to solve it generally | 21:04 |
| fungi | fair enough! | 21:05 |
| bodepd | I've got a generic jenkins question. I've been struggling to get adding slaves automated | 21:05 |
| bodepd | basically, this line of code: https://github.com/openstack-infra/nodepool/blob/master/nodepool/jenkins_manager.py#L29 | 21:05 |
| bodepd | are you guys passing credential ids or private keys? | 21:05 |
| bodepd | (where is the calling code for that? | 21:05 |
| zaro | jeblair, fungi : cool. i'll stop work on creating the jobs. | 21:05 |
| jeblair | bodepd: credential ids; nodepool.py | 21:05 |
| fungi | zaro: so you'll want to manually build an sdist from your version update commit on the 0.3-fixes branch, and then manually upload that to pypi | 21:06 |
| zaro | fungi: ok. let me try that. | 21:06 |
| jeblair | zaro: what's your pypi username? | 21:06 |
| zaro | jeblair: zaro0508 | 21:07 |
| bodepd | does it even work without credential ids? | 21:07 |
| *** otherwiseguy has quit IRC | 21:07 | |
| bodepd | and how are those credential ids create? I couldn't find an API for it | 21:07 |
| jeblair | bodepd: used to; dunno anymore... | 21:07 |
| jeblair | bodepd: webui | 21:07 |
| bodepd | I don't mean to swear, but F*CK jenkins | 21:07 |
| bodepd | oh wait, I did mean to swear | 21:08 |
| *** shashankhegde has joined #openstack-infra | 21:08 | |
| jeblair | bodepd: i think they are in their own file, so if there's no api, you may still be able to write that file and have jenkins reload the config | 21:08 |
| jeblair | bodepd: but we use the same key on all our hosts | 21:08 |
| bodepd | yeah, config.xml of whatever | 21:08 |
| jeblair | bodepd: no, not that file | 21:08 |
| *** johnthetubaguy is now known as zz_johnthetubagu | 21:09 | |
| bodepd | I can fidn the file. I was overwritting xml to create simple login users | 21:09 |
| jeblair | bodepd: that file has slaves defined in it, which means editing it directly while having something add/remove slaves is racy | 21:09 |
| jeblair | bodepd: there's a credentials.xml or something i think | 21:09 |
| bodepd | oh, maybe it's sane then | 21:09 |
| openstackgerrit | Spencer Krum proposed a change to openstack-infra/config: First draft of enabling beaker tests to run https://review.openstack.org/102071 | 21:09 |
| bodepd | I just want a heat template that can create a single master with a connected slave | 21:09 |
| jeblair | bodepd: anyway, we defined the one credential with an ssh key and then just reference that everywhere | 21:09 |
| jeblair | bodepd: ah, gotcha | 21:10 |
| *** melwitt has joined #openstack-infra | 21:10 | |
| bodepd | yeah. it's hard for bootstrapping. b/c you need to build the master, then manuall edit stuff before adding hte salve | 21:10 |
| bodepd | I'll look into that file. | 21:10 |
| jeblair | bodepd: have you found that the alternate form in nodepool does not work? | 21:10 |
| *** otherwiseguy has joined #openstack-infra | 21:10 | |
| bodepd | it looks like the non-credential way doesn't work any more | 21:10 |
| nibalizer | so i think i've got a rube goldberg machine that lets beaker tests run on the nodepool machines | 21:10 |
| bodepd | I wasn't using nodepool | 21:10 |
| jeblair | bodepd: last time i tried it, i think it created credential entries automatically by doing that | 21:10 |
| *** s0nea has quit IRC | 21:10 | |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: Handle proposed/ branches in test-matrix.py https://review.openstack.org/104448 | 21:10 |
| bodepd | jeblair: oh. sorry I misunderstood | 21:10 |
| nibalizer | jeblair: hogepodge https://review.openstack.org/#/c/102020/ and https://review.openstack.org/102071 are the relevant reviews | 21:11 |
| bodepd | I have found it to no longer work | 21:11 |
| jeblair | bodepd: right, the 'provide an ssh key' method | 21:11 |
| jeblair | hrm | 21:11 |
| bodepd | stack traces, something about null pointers | 21:11 |
| * nibalizer reads scrollback | 21:11 | |
| bodepd | I had some puppet resources that could add slaves, they didn't work | 21:11 |
| bodepd | the I found the jenkins_api_client gem and wrote types/providers for it | 21:11 |
| bodepd | that didn't work for private keys | 21:11 |
| bodepd | then I read the code you guys used to make sure it was the same | 21:12 |
| bodepd | so according to that round-about logic, I *believe* that method does not work | 21:12 |
| *** rfolco has quit IRC | 21:12 | |
| clarkb | notmyname: https://review.openstack.org/#/c/104448 merged | 21:12 |
| clarkb | nibalizer: I am looking for your change now to recheck | 21:12 |
| jeblair | bodepd: bummer | 21:12 |
| bodepd | and the errors are pretty consistent between the 2 failed methods. talked about certain versions of things not working between jenkins version, cred version, ssh slave ver etc | 21:13 |
| bodepd | yeah, it would have beeen easy without external creds | 21:13 |
| jeblair | zaro: i can't add users to python-jenkins on pypi | 21:13 |
| *** s0nea has joined #openstack-infra | 21:13 | |
| bodepd | but you have re-enforced that is the way to do it | 21:13 |
| jeblair | zaro: but i can build and upload manually using openstackci's creds | 21:13 |
| *** yfried has joined #openstack-infra | 21:13 | |
| zaro | jeblair: ok. i guess this just needs to get approved first: https://review.openstack.org/#/c/104675 | 21:14 |
| clarkb | notmyname: and rechecked let us see how that does | 21:14 |
| jeblair | zaro: is there a 0.3.0? | 21:14 |
| jeblair | zaro: ok, let me know when you're ready | 21:15 |
| hogepodge | nibalizer looking | 21:15 |
| fungi | jeblair: there was a 0.3.0 tag pushed for the parent of that | 21:15 |
| zaro | jeblair: i already tagged 0.3.0 but that failed to release. i will tag 0.3.1 after above merges | 21:15 |
| nibalizer | clarkb: i can't parse what you said to me | 21:16 |
| zaro | jeblair: doc for python-jenkins is here: http://pythonhosted.org//python-jenkins | 21:16 |
| *** radez is now known as radez_g0n3 | 21:16 | |
| clarkb | nibalizer: bad tab complete. sorry | 21:16 |
| zaro | jeblair: I have no idea how to update that on a new release | 21:16 |
| clarkb | nibalizer: i and o are right next to each other on this qwerty keyboard | 21:16 |
| nibalizer | ah gotcha | 21:17 |
| fungi | clarkb: commented on 104437. any reason not to do that? | 21:20 |
| *** otherwiseguy has quit IRC | 21:20 | |
| clarkb | fungi: good catch | 21:21 |
| clarkb | fungi: though its read only... meh | 21:21 |
| clarkb | fungi: I am hcecking the group now | 21:22 |
| fungi | my brain is a little twisty on the code in 104451 too | 21:22 |
| clarkb | www-data looks good | 21:22 |
| fungi | oh, nevermind i get 104451 now | 21:22 |
| *** otherwiseguy has joined #openstack-infra | 21:23 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Ensure there is a slash in the object prefix https://review.openstack.org/104451 | 21:24 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Add in wsgi.conf for os-loganalyze https://review.openstack.org/104437 | 21:24 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 21:26 |
| jeblair | clarkb: why are you uplading new versions of those? | 21:28 |
| jeblair | oh | 21:28 |
| *** rcarrill` has joined #openstack-infra | 21:28 | |
| jeblair | i see you changed something in the second one | 21:28 |
| clarkb | ya, I don;t know why jhesketh_ stacked them like that | 21:28 |
| clarkb | I can unstack them if that helps | 21:28 |
| jeblair | nah, i just aprvd the 1st | 21:28 |
| clarkb | cool | 21:28 |
| jeblair | i was just going through lots of empty inter-patchset diffs :) | 21:29 |
| openstackgerrit | Boris Pavlovic proposed a change to openstack/requirements: Bump osprofiler version to >=0.2.0 https://review.openstack.org/104691 | 21:29 |
| *** andreykurilin_ has joined #openstack-infra | 21:29 | |
| *** rcarrillocruz has quit IRC | 21:30 | |
| jeblair | clarkb: second lgtm; note that it's a production configuration change so i don't personally intend to approve it now :) | 21:31 |
| clarkb | ok | 21:31 |
| clarkb | I will probably approve when jhesketh shows up | 21:31 |
| clarkb | it is exciting stuff | 21:32 |
| fungi | in that case i unapproved 104437 just now ;) | 21:32 |
| clarkb | fungi: oh well if you want ot babysit go for it :) | 21:32 |
| jeblair | fungi: well, i mean if you want to make sure it doesn't break logs.o.o, go for it :) | 21:32 |
| clarkb | notmyname: the jobs for your swift change look good so far | 21:33 |
| fungi | nah, it makes sense to wait until jhesketh_ is up and about anyway | 21:33 |
| clarkb | we will need to rebuild images for the script change | 21:33 |
| clarkb | just bare-precise for our experimental job | 21:33 |
| clarkb | jeblair: did you get a chance to see the zuul change too? | 21:34 |
| jeblair | clarkb: do you grok the errors in http://logs.openstack.org/90/104690/1/check/gate-config-puppet-apply-precise/78755d5/console.html ? | 21:34 |
| clarkb | I haven't reviewed that one yet | 21:34 |
| *** prad has quit IRC | 21:35 | |
| clarkb | jeblair: the database stuff is the mysql module not working in noop mode I think | 21:35 |
| *** SumitNaiksatam has quit IRC | 21:35 | |
| clarkb | I do not know why the fqdn fact isn't working | 21:35 |
| *** james_li has quit IRC | 21:35 | |
| jeblair | so there's more than just operating system mismatches | 21:37 |
| *** jhesketh_ is now known as jhesketh | 21:37 | |
| clarkb | ya I think the db thing might be difficult to resolve | 21:37 |
| clarkb | there are bits in there that expect the user to exist | 21:37 |
| clarkb | but since we are in noop mode the db user isn't created | 21:38 |
| clarkb | and the subsequent things don't grok that in a noop context | 21:38 |
| clarkb | we might be able to fix that though | 21:38 |
| clarkb | bodepd: ^ | 21:38 |
| *** james_li has joined #openstack-infra | 21:38 | |
| *** devkulkarni has quit IRC | 21:39 | |
| clarkb | by the way I restared the logstash-indexer process on the indexers again | 21:39 |
| clarkb | and that seems to have helped | 21:39 |
| clarkb | so I think they are still getting themselves wedged over time :/ | 21:39 |
| *** devkulkarni has joined #openstack-infra | 21:39 | |
| clarkb | I may need to break out the strace and try not to cry when that runs on the jvm | 21:39 |
| openstackgerrit | A change was merged to openstack-infra/zuul: Fix LOG_PATH to use 7char uuid https://review.openstack.org/104453 | 21:40 |
| bodepd | clarkb: yeah, I hate that about puppet | 21:40 |
| bodepd | if fqdn fails, then it doesn't try to use hostname | 21:41 |
| bodepd | it returns empty string | 21:41 |
| clarkb | bodepd: I was hoping you would be all like "oh ya you just do this and that and bam it works" | 21:41 |
| clarkb | :) | 21:41 |
| bodepd | fqdn fact needs to resolve | 21:41 |
| bodepd | it's a pain | 21:41 |
| bodepd | I think the easiest fix it to set 127.0.1.1 | 21:41 |
| clarkb | oh maybe if we fix fqdn fact it fixes the db? | 21:41 |
| bodepd | in etc hosts | 21:41 |
| clarkb | ok I grok better now | 21:41 |
| bodepd | yep | 21:41 |
| bodepd | hold on. lemme find the facter code | 21:41 |
| bodepd | I think it runs dnsdomainname | 21:42 |
| clarkb | jeblair: so the facter fact for fqdn is silly. it won't treat hostname. as an fqdn | 21:42 |
| *** dkranz has quit IRC | 21:42 | |
| bodepd | and then checks resolve.conf | 21:42 |
| clarkb | you must have an actual domain in front of your hostname | 21:42 |
| *** mestery__ has joined #openstack-infra | 21:42 | |
| bodepd | for search paths | 21:42 |
| clarkb | because no one ever sets up home dns with a root of . | 21:42 |
| bodepd | or something | 21:42 |
| bodepd | yeah, I'm pretty sure setting 127.0.1.1 will make it all work. | 21:42 |
| *** mestery_ has quit IRC | 21:42 | |
| bodepd | lemme double check that | 21:42 |
| clarkb | jeblair: so first step in that test should probably be ^ | 21:42 |
| openstackgerrit | Boris Pavlovic proposed a change to openstack/requirements: Bump osprofiler version to >=0.2.1 https://review.openstack.org/104691 | 21:43 |
| bodepd | yeah, we had to do it for our openstack tests b/c it was hard to make guarentees about fqdn being set in test envs | 21:43 |
| bodepd | echo "127.0.1.1 \$(hostname).\$(get_meta domain) \$(hostname)" >> /etc/host | 21:44 |
| bodepd | :) | 21:44 |
| bodepd | well, not quite tat easy, but you get the point | 21:44 |
| *** doddstack has quit IRC | 21:44 | |
| bodepd | it really needs to be fixed in Puppet | 21:44 |
| bodepd | it's annoying | 21:44 |
| bodepd | s/Puppet/Facter/ | 21:44 |
| clarkb | thanks | 21:44 |
| jeblair | clarkb: istr some changes involving etc/hosts on nodepool hosts... | 21:46 |
| clarkb | jeblair: I think they only happened on the image builds side | 21:46 |
| clarkb | jeblair: so by the time the actual slave boots we need to do it again | 21:46 |
| jeblair | https://review.openstack.org/#/c/98561/ | 21:47 |
| clarkb | and the ready node part doesn't get hostanme as $1 from nodepool | 21:47 |
| clarkb | I think the old d-g scripst did that and nodepool hasn't grown the feature. so yes we should clean that up and get it in | 21:47 |
| clarkb | except for dib? maybe mordred can comment on that | 21:47 |
| *** ociuhandu has joined #openstack-infra | 21:48 | |
| jeblair | bodepd: okay, what exactly needs to be set? | 21:50 |
| jeblair | because "hostname" returns devstack-trusty-rax-iad-726268.slave.openstack.org | 21:50 |
| jeblair | and $HOSTNAME is devstack-trusty-rax-iad-726268.slave.openstack.org | 21:50 |
| jeblair | 104.130.7.44devstack-trusty-rax-iad-726268.slave.openstack.org | 21:51 |
| jeblair | 10.208.227.2devstack-trusty-rax-iad-726268.slave.openstack.org | 21:51 |
| jeblair | and those are in /etc/hosts | 21:51 |
| clarkb | jeblair: can you do `facter fqdn` ? | 21:51 |
| jeblair | facter fqdn | 21:51 |
| jeblair | devstack-trusty-rax-iad-726268.slave.openstack.org | 21:51 |
| clarkb | huh | 21:51 |
| jeblair | let me try precise on the provider that failed in case there's a diff | 21:51 |
| notmyname | clarkb: just got out of a meeting. looks like the -infra patch landed? | 21:52 |
| fungi | notmyname: yes | 21:52 |
| clarkb | notmyname: yup and I rechecked your change | 21:52 |
| clarkb | it has ~11 minutes left according to zuul | 21:52 |
| jeblair | okay, yeah it's different on hpcloud | 21:52 |
| jeblair | apparently hpcloud doesn't set hostnames, but rackspace does | 21:53 |
| clarkb | yay clouds | 21:53 |
| jeblair | mordred: ^ for the litany | 21:53 |
| *** harlowja_at_home has quit IRC | 21:53 | |
| notmyname | clarkb: awesome, thanks | 21:53 |
| jeblair | on hpcloud: | 21:53 |
| jeblair | hostname: bare-precise-hpcloud-b3-726277 | 21:53 |
| jeblair | $HOSTNAME: bare-precise-hpcloud-b3-726277 | 21:53 |
| jeblair | 127.0.1.1 bare-precise-1404396872.template.openstack.org | 21:53 |
| *** mrda-away is now known as mrda | 21:53 | |
| notmyname | clarkb: and it looks like it's way past the point that would have broken. so I'll recheck the other two too | 21:54 |
| clarkb | notmyname: yup | 21:54 |
| notmyname | clarkb: done. thanks | 21:54 |
| bodepd | jeblair: start by running facter fqdn | 21:55 |
| *** nelsnelson has joined #openstack-infra | 21:55 | |
| jeblair | bodepd: i think i got it now | 21:55 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 21:58 |
| anteaya | adalbas: great | 21:59 |
| anteaya | adalbas: what happens when you try that option? | 21:59 |
| clarkb | https://jenkins05.openstack.org/job/gate-config-puppet-apply-precise/235/console can follow along there | 22:00 |
| jeblair | that last sudo warning should be the "sudo fix etc/hosts" command | 22:00 |
| jeblair | so a good sign there's nothing after | 22:01 |
| *** gondoi is now known as zz_gondoi | 22:01 | |
| *** denis_makogon has quit IRC | 22:01 | |
| *** devkulkarni has quit IRC | 22:02 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 22:02 |
| *** signed8bit_ZZZzz is now known as signed8bit | 22:02 | |
| *** andreykurilin_ has quit IRC | 22:02 | |
| *** devkulkarni has joined #openstack-infra | 22:02 | |
| *** dangers is now known as dangers_away | 22:02 | |
| jeblair | wow that's really laggy | 22:02 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 22:03 |
| fungi | i have seen gerritbot lag for up to a minute at times | 22:04 |
| fungi | not sure what triggers it | 22:04 |
| jeblair | fungi: maybe somebody slinging up a 20 patchset series? | 22:04 |
| fungi | oh, you mean dansmith | 22:04 |
| dansmith | not me :) | 22:04 |
| clarkb | that somebody | 22:04 |
| jeblair | or lifeless | 22:04 |
| clarkb | :P | 22:05 |
| fungi | heh | 22:05 |
| clarkb | you know what the problem with having 15 gallons of homebrew fermenting is? I have non carbonated and ready to drink | 22:06 |
| jeblair | clarkb: i like how at least the job is pretending to think about it longer | 22:06 |
| openstackgerrit | pritesh proposed a change to openstack-dev/hacking: Add hacking check for editor configuration in source files https://review.openstack.org/101969 | 22:06 |
| clarkb | jeblair: in parallel even :) | 22:06 |
| jeblair | success! | 22:06 |
| jeblair | land it! make it gate! :) | 22:07 |
| clarkb | its starting too feel like that time of day though. where is jhesketh we should do this thing | 22:07 |
| clarkb | jeblair: ok reviewing it properly now | 22:07 |
| clarkb | jeblair: can you push a new patchset that makes it gate too? | 22:07 |
| lifeless | speaking of me, sorry. | 22:07 |
| jeblair | clarkb: sure | 22:07 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Hide fully reviewed projects by default https://review.openstack.org/99591 | 22:07 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Don't crash on comments on unchanged files https://review.openstack.org/99563 | 22:07 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Support comments in commits https://review.openstack.org/99272 | 22:07 |
| clarkb | jeblair: I think we can do all of that together | 22:07 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Only show verified reviews by default. https://review.openstack.org/100138 | 22:07 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Filter out WIP patches by default in changes list. https://review.openstack.org/100121 | 22:08 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Attempt to fetch all new refs for a change at once https://review.openstack.org/104704 | 22:08 |
| fungi | lifeless: you should totally apologize for being a coding machine and getting work done | 22:08 |
| *** EntropyWorks has left #openstack-infra | 22:08 | |
| *** zehicle_at_dell has quit IRC | 22:08 | |
| jhesketh | Morning | 22:08 |
| jhesketh | clarkb: ^ that thing? | 22:08 |
| lifeless | fungi: :) I was apologising for flooding your screens | 22:08 |
| clarkb | jhesketh: oh hi | 22:08 |
| clarkb | jhesketh: we want to approve your wsgi change for os loganalyze | 22:09 |
| *** zehicle_at_dell has joined #openstack-infra | 22:09 | |
| clarkb | jhesketh: but ifgured you should be around if it breaks everything | 22:09 |
| clarkb | jhesketh: you ok with that? I can approve if so | 22:09 |
| anteaya | morning jhesketh | 22:09 |
| *** dimtruck is now known as zz_dimtruck | 22:10 | |
| jhesketh | clarkb: sounds good to me | 22:10 |
| mattoliverau | Morning | 22:10 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 22:10 |
| clarkb | jhesketh: done lets see how that does | 22:11 |
| clarkb | jhesketh: zuul change was merged too but we haven't done a restart. I figure that is slightly less important as we can mostly verify everything without that change | 22:11 |
| clarkb | once the script update merges I can fire off image rebuilds tooo | 22:12 |
| *** julim has quit IRC | 22:12 | |
| jhesketh | clarkb: cool, thanks | 22:12 |
| jhesketh | yep | 22:12 |
| anteaya | morning mattoliverau | 22:13 |
| *** dims__ has joined #openstack-infra | 22:13 | |
| clarkb | jeblair: thanks. +2 | 22:13 |
| jhesketh | clarkb: will we need to reload the wsgi app? | 22:14 |
| clarkb | jhesketh: maybe? | 22:15 |
| clarkb | actually probably. I can restart apache too | 22:15 |
| *** dims_ has quit IRC | 22:15 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add in wsgi.conf for os-loganalyze https://review.openstack.org/104437 | 22:15 |
| clarkb | I had to recheck the / change | 22:17 |
| clarkb | because of weirdness in it being cancelled in the gateafter it couldn't merge when fungi removed his +A | 22:17 |
| *** signed8bit is now known as signed8bit_ZZZzz | 22:17 | |
| *** Sukhdev has quit IRC | 22:18 | |
| *** mestery__ has quit IRC | 22:18 | |
| jhesketh | clarkb: looks like serving static logs still works so that's good | 22:18 |
| *** gfa has left #openstack-infra | 22:18 | |
| fungi | clarkb: yeah, i think it probably hit the same zuul bug i'm currently trying to fix | 22:19 |
| mordred | clarkb, jeblair: what's the deal with hostnames | 22:19 |
| mordred | ? | 22:19 |
| clarkb | mordred: hpcloud doesn't set them proeprly or some such | 22:19 |
| clarkb | mordred: so puppet fails | 22:19 |
| openstackgerrit | pritesh proposed a change to openstack-dev/hacking: Add hacking check for editor configuration in source files https://review.openstack.org/101969 | 22:19 |
| clarkb | jhesketh: I think we have to wait for puppet to run on static.o.o with the new stuff | 22:20 |
| clarkb | jhesketh: should happen in ~15 minutes | 22:20 |
| jhesketh | oh of course | 22:20 |
| mordred | clarkb: lovely | 22:21 |
| mordred | clarkb: perhaps we're intended to do that with cloud-init or something? | 22:21 |
| clarkb | mordred: to be clear it sets a hostname but not an fqdn as facter expects it | 22:21 |
| mordred | clarkb: also, wait - it _fails_? | 22:21 |
| clarkb | mordred: maybe? | 22:21 |
| mordred | ah. yes | 22:21 |
| mordred | that | 22:22 |
| mordred | that's a nova behavior | 22:22 |
| mordred | there are people who consider it "correct" | 22:22 |
| openstackgerrit | A change was merged to openstack-infra/config: Ensure there is a slash in the object prefix https://review.openstack.org/104451 | 22:22 |
| mordred | and have told me I'm wrong when I tell them that people running real things want it to do what we ask | 22:22 |
| clarkb | clearly the users are never right | 22:22 |
| clarkb | and we should just hack in a vacuum | 22:22 |
| *** terryw has joined #openstack-infra | 22:23 | |
| mordred | nope. certainly don't listen to anyone who is actually using nova in anger at massive scale | 22:23 |
| *** otherwiseguy has quit IRC | 22:23 | |
| clarkb | elasticsearch sitrep. ssds have made 07 happy this is good | 22:23 |
| clarkb | however logstash daemons still get cranky after a while then everything sucks | 22:23 |
| clarkb | I will have to brave strace on jvm at some point probably | 22:24 |
| mordred | eek | 22:24 |
| clarkb | or run things in jdb | 22:24 |
| clarkb | not sure which is worse | 22:24 |
| mordred | clarkb: aren't we going to move to not logstash? | 22:24 |
| clarkb | mordred: we are but thats a long ways out | 22:24 |
| mordred | ok | 22:24 |
| clarkb | mostly beacuse have to do things in projects that I have had zero time to do | 22:24 |
| clarkb | and need a way to present data to humans | 22:24 |
| clarkb | so that sdague doesn't kill me | 22:24 |
| mordred | and me | 22:25 |
| mordred | I would also kill you | 22:25 |
| mordred | if I was a user | 22:25 |
| clarkb | good to know | 22:25 |
| clarkb | (note to self: kill mordred first) | 22:25 |
| mordred | :) | 22:25 |
| mordred | clarkb: ok - next question - what do we do about pbr integration test? | 22:25 |
| clarkb | mordred: what about it? | 22:25 |
| clarkb | sorry I think I may be missing context | 22:25 |
| mordred | clarkb: it uses pypi-mirror | 22:25 |
| clarkb | oh | 22:26 |
| mordred | clarkb: which we're about to stop using anywhere else | 22:26 |
| clarkb | just point it at our mirror now | 22:26 |
| clarkb | I think thats ok | 22:26 |
| clarkb | beacuse its everything including new reqs | 22:26 |
| mordred | nod | 22:26 |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Bring some uniformity to olso acl files https://review.openstack.org/101361 | 22:26 |
| mordred | lemme try that | 22:26 |
| fungi | it uses pypi-mirror to test installing things | 22:26 |
| clarkb | and we don't need to test pypi-mirror that way | 22:26 |
| clarkb | oh right it does it over and over and over | 22:26 |
| fungi | i suspect we need to reinvent part of pypi-mirror in the pbr testing | 22:26 |
| clarkb | maybe we can use our new mirror + pip cache? iirc pip cache does the wrong thing though | 22:26 |
| mordred | oh - well, also - it wants to test requirements changes ... | 22:27 |
| mordred | we could use devpi | 22:27 |
| clarkb | mordred: ya it will do that though | 22:27 |
| clarkb | because the new mirror has all the possible requirements | 22:27 |
| mordred | you can inject things into devpi | 22:27 |
| clarkb | even the ones we don't set now | 22:27 |
| fungi | yeah, basically to step back, its reason for existing is to try installing all the things and make sure nothing breaks on installation (together) | 22:27 |
| mordred | hrm. | 22:27 |
| *** devkulkarni has quit IRC | 22:27 | |
| mordred | the only think that _won't_ be in the mirror is pbr | 22:27 |
| mordred | so it won't test that the current commmit of pbr can be installed if it's in the mirror | 22:28 |
| *** eharney has quit IRC | 22:28 | |
| mordred | s/if it's in/if it were to be put in/ | 22:28 |
| mordred | but ... | 22:28 |
| mattoliverau | clarkb, mordred: does ansible's facts for for FQDN? Cause your using ansible to call puppet now right? If it does you can pass the --certname in the puppet command to specifcy the puppet node name which you could populate from ansible. As a work around. | 22:28 |
| mordred | pbr is WAY more stable now | 22:28 |
| mattoliverau | s/for/wrok/ | 22:29 |
| mordred | mattoliverau: lemme look | 22:29 |
| mattoliverau | work even, typing is good this morning | 22:29 |
| mordred | mattoliverau: oh - but we don't use it during nodepool | 22:29 |
| mattoliverau | ahh, yeah the inital build | 22:30 |
| jeblair | clarkb: "not logstash" ? | 22:30 |
| lifeless | hmm | 22:30 |
| clarkb | jhesketh: image builds started | 22:30 |
| mattoliverau | unless nodepool puts the FQDN in hosts or exports it and uses it in --certname | 22:30 |
| lifeless | anyone else noticed gertty creates thousands of https connections | 22:30 |
| clarkb | jeblair: I think openstack should log in json. And enable the use of simpler tools like fluentd | 22:30 |
| *** d34dh0r53 is now known as mostly_d34dh0r53 | 22:31 | |
| clarkb | jeblair: we are spending a ton of resources on the "try to udnerstand this particular log line" problem when the tools doing the logging could just give us structured data from the start | 22:31 |
| clarkb | jeblair: they will always be more accurate too | 22:31 |
| openstackgerrit | K Jonathan Harker proposed a change to openstack-infra/nodepool: Enable debugging output https://review.openstack.org/104708 | 22:31 |
| clarkb | then we can spend lots of cpu on machine learning and actually interseting stuff | 22:31 |
| jeblair | clarkb: and we decided to do that? | 22:31 |
| clarkb | jeblair: ish. oslo logging apparently supports it | 22:32 |
| jeblair | mordred: approve this please: https://review.openstack.org/#/c/104690/ | 22:32 |
| clarkb | jeblair: so that isn't an issue | 22:32 |
| jeblair | clarkb: when did we decide this? | 22:32 |
| clarkb | jeblair: we haven't really decided anything | 22:32 |
| clarkb | jeblair: it has been discussed (you were around for that first discussion iirc0 | 22:33 |
| clarkb | jeblair: but once openstack supports it via oslo logging we can actually play with it and see if it makes sense | 22:33 |
| jhesketh | clarkb: shouldn't need new images for the wsgi change | 22:33 |
| clarkb | jhesketh: we do for the / change | 22:33 |
| clarkb | jeblair: I think it would be premature to decide now before oslo is synced everywhere | 22:34 |
| jeblair | clarkb: ok. this all sounds new to me. it's possible i have forgotten it, of course. however, mordred made it sound much more settled than it apparently is. | 22:34 |
| jhesketh | right | 22:34 |
| mordred | jeblair: yah - sorry, what I meant was "hadn't we talked about" | 22:34 |
| clarkb | jeblair: do you have concerns with that? | 22:34 |
| clarkb | jeblair: sdague's concern is human readbility | 22:34 |
| asselin | HI, do any of you have time for a zuul gerrit question? 2014-07-03 15:32:33,792 ERROR zuul.Gerrit: Cannot get references from https://repomaster.usa.hp.com:8081/p/ ... Unauthorized | 22:35 |
| clarkb | with tools like journalctl I don't think it is crazy to filter structured logs through somethign that makes them readable | 22:35 |
| jeblair | clarkb: any chance you remember where we all talked about it? | 22:35 |
| clarkb | jeblair: it was here in this channel. I don't remember dates though | 22:35 |
| asselin | I'm not sure where to set authorization..? I'm able to stream gerrit events. This is an internal gerrit server. | 22:35 |
| clarkb | sometime around june 11th? | 22:35 |
| * clarkb starts looking at logs | 22:35 | |
| jeblair | clarkb: i was in the wilderness | 22:36 |
| clarkb | asselin: you appear to be trying to get refs via https | 22:36 |
| clarkb | jeblair: huh I thought you caught the discussion | 22:36 |
| clarkb | jeblair: I will find it in logs | 22:36 |
| jeblair | clarkb: http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2014-06-12.log | 22:36 |
| jeblair | clarkb: i'll try to catch up on that next week | 22:37 |
| *** andreaf has quit IRC | 22:37 | |
| jeblair | i have to run now | 22:37 |
| clarkb | jeblair: http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2014-06-11.log starts at 2014-06-11T21:55:47 | 22:37 |
| jeblair | clarkb: 2 days in a row :) | 22:38 |
| clarkb | :) | 22:38 |
| clarkb | jeblair: long story short I think the best place to structure the data is at the sourc | 22:38 |
| clarkb | logstash is a tool for wrangling bad logs into shape | 22:38 |
| asselin | clarkb, https://github.com/openstack-infra/zuul/blob/master/zuul/trigger/gerrit.py#L121 | 22:38 |
| clarkb | we are getting better at producing good logs and can hopefully level up into a better place with structured logs | 22:38 |
| clarkb | asselin: right just pointing out that ssh stream events is different than https | 22:38 |
| *** MarkAtwood has joined #openstack-infra | 22:39 | |
| clarkb | asselin: does your gerrit allow read access via https without being logged in? | 22:39 |
| asselin | ok oh I see. Maybe not | 22:39 |
| asselin | ok...another dead end. I guess I should go on vacation | 22:39 |
| jeblair | jhesketh: can you please review/approve https://review.openstack.org/#/c/104690/ ? | 22:40 |
| jeblair | i'd like to latch that before it slips any further | 22:40 |
| jeblair | now i have to go afk for a while | 22:41 |
| jhesketh | jeblair: will do | 22:41 |
| jhesketh | shall I wait for you to return before merging? | 22:41 |
| *** harlowja_at_home has joined #openstack-infra | 22:41 | |
| jeblair | nope | 22:41 |
| *** yamahata has quit IRC | 22:41 | |
| *** harlowja_at_home has quit IRC | 22:42 | |
| *** harlowja_at_home has joined #openstack-infra | 22:42 | |
| clarkb | gah puppet is hard | 22:43 |
| clarkb | fix for wsgi thing incoming | 22:43 |
| *** melwitt has quit IRC | 22:44 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: File resources use content not template. https://review.openstack.org/104710 | 22:45 |
| clarkb | jeblair: fungi jhesketh mordred ^ | 22:45 |
| openstackgerrit | A change was merged to openstack-infra/config: Only run relevant puppet apply tests https://review.openstack.org/104690 | 22:45 |
| *** rlandy has quit IRC | 22:45 | |
| *** MarkAtwood has quit IRC | 22:45 | |
| jhesketh | clarkb: sigh, thanks | 22:48 |
| *** zzelle has quit IRC | 22:48 | |
| anteaya | I need a had here: https://review.openstack.org/#/c/101361/9 | 22:49 |
| anteaya | hand | 22:49 |
| anteaya | I rebased between patchsets 7 and 8 and patchset 8 passed | 22:50 |
| anteaya | I changed the commit messge (in the gui) in patchest 9 | 22:50 |
| anteaya | and patchset 9 is calling for a rebase | 22:50 |
| clarkb | anteaya: have you tried t orebase it yet to see what conflicts there are? | 22:50 |
| *** terryw has quit IRC | 22:50 | |
| anteaya | not for patchset 9 | 22:51 |
| anteaya | I can do that | 22:51 |
| clarkb | that is the best place to start | 22:51 |
| anteaya | I'll pull down the patch and do git rebase -i master, with a updated master | 22:51 |
| anteaya | yes? | 22:51 |
| clarkb | yup | 22:51 |
| *** ihrachyshka has quit IRC | 22:51 | |
| anteaya | k | 22:51 |
| *** timrc-afk is now known as timrc | 22:51 | |
| openstackgerrit | A change was merged to openstack-infra/config: File resources use content not template. https://review.openstack.org/104710 | 22:54 |
| *** asselin has quit IRC | 22:55 | |
| *** msabramo has quit IRC | 22:56 | |
| *** david-lyle has quit IRC | 22:57 | |
| *** david-ly_ has joined #openstack-infra | 22:58 | |
| *** Sukhdev has joined #openstack-infra | 23:00 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 23:02 | |
| *** HACKING-FACEBOOK has quit IRC | 23:02 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 23:02 | |
| *** HACKING-FACEBOOK has joined #openstack-infra | 23:02 | |
| *** HACKING-FACEBOOK has quit IRC | 23:02 | |
| anteaya | I am just digging a bigger hole: http://paste.openstack.org/show/85460/ | 23:02 |
| anteaya | I added my files and did git commit --amend | 23:03 |
| *** HACKING-FACEBOOK has joined #openstack-infra | 23:03 | |
| *** HACKING-FACEBOOK has quit IRC | 23:03 | |
| harlowja_at_home | anyway i could get a recheck on https://review.openstack.org/#/c/96631/ (its just been rebase failing a bunch, been signed off and +2'd previously) | 23:03 |
| anteaya | this is not my commit message | 23:03 |
| *** david-ly_ has quit IRC | 23:03 | |
| anteaya | what did I do and how do I stop doing it? | 23:03 |
| *** shashankhegde has quit IRC | 23:04 | |
| anteaya | I should have done git rebase --continue, shouldn't I? | 23:04 |
| *** pballand has quit IRC | 23:04 | |
| anteaya | sigh | 23:04 |
| clarkb | anteaya you do not git commit amend in a rebase. you want rebase --continue | 23:04 |
| anteaya | right | 23:04 |
| *** jamielennox|away is now known as jamielennox | 23:04 | |
| clarkb | so do a rebase --abort | 23:04 |
| anteaya | that was my mistake | 23:04 |
| anteaya | so quit the editor? | 23:05 |
| clarkb | that will get you back to where you started | 23:05 |
| anteaya | okay thanks | 23:05 |
| clarkb | yes | 23:05 |
| anteaya | okay | 23:05 |
| *** ildikov_ has joined #openstack-infra | 23:08 | |
| *** msabramo has joined #openstack-infra | 23:09 | |
| mordred | clarkb: fwiw, old review is still running | 23:09 |
| clarkb | oh fun | 23:09 |
| * mordred was looking for a count of how many machines we run in not-nodepool | 23:10 | |
| mordred | the answer is 79, if anyone is interested | 23:10 |
| pleia2 | wow | 23:10 |
| clarkb | I am restarting apache on static.o.o now | 23:10 |
| *** ildikov has quit IRC | 23:10 | |
| clarkb | jhesketh: ^ | 23:10 |
| clarkb | jhesketh: image builds are almost done too | 23:10 |
| jhesketh | clarkb: :-) | 23:10 |
| clarkb | I seem to be able to pull up logs still | 23:11 |
| *** michchap has quit IRC | 23:11 | |
| clarkb | can others indepedently verify to make sure it isn't my caches or anything | 23:11 |
| openstackgerrit | Anita Kuno proposed a change to openstack-infra/config: Bring some uniformity to olso acl files https://review.openstack.org/101361 | 23:11 |
| *** michchap has joined #openstack-infra | 23:11 | |
| *** nithyag__ has quit IRC | 23:11 | |
| anteaya | I can only hope it gets merged before I have to rebase again | 23:11 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Use a requests.Session object to enable pooling. https://review.openstack.org/104715 | 23:12 |
| *** nithyag__ has joined #openstack-infra | 23:12 | |
| anteaya | mordred: thanks I find that interesting | 23:12 |
| *** gokrokve has quit IRC | 23:12 | |
| *** Ryan_Lane has quit IRC | 23:12 | |
| lifeless | tchaypo: ^ should help with your gertty sync performance (but be sure to grab git GitPython 0.3, to get my multi-ref patch, which I think works] | 23:12 |
| devananda | how can gertty have Sync: 750 when my subscribed projects have ~100 reviews open, total. | 23:12 |
| *** nijaba has quit IRC | 23:12 | |
| *** msabramo1 has joined #openstack-infra | 23:13 | |
| *** nijaba has joined #openstack-infra | 23:13 | |
| *** nijaba has joined #openstack-infra | 23:13 | |
| mordred | if you count each pool of things. like jenkins0*, git0* and elasticseaerch0* as one - we have 43 discrete things | 23:13 |
| *** msabramo has quit IRC | 23:13 | |
| lifeless | devananda: the sync count is tasks to sync, not reviews. | 23:13 |
| lifeless | devananda: so outstanding http/git requests basically | 23:13 |
| lifeless | devananda: IIRC | 23:13 |
| *** gmurphy has quit IRC | 23:13 | |
| clarkb | correct | 23:14 |
| *** camunoz has joined #openstack-infra | 23:14 | |
| *** alexandra_ has joined #openstack-infra | 23:14 | |
| devananda | huh | 23:14 |
| *** alexandra_ is now known as asettle | 23:14 | |
| clarkb | so fetching refs for intermediate patchsets, getting lists of changes since last update, all that goes into the queue | 23:14 |
| lifeless | devananda: which I should, as I was just improving that code. | 23:14 |
| jhesketh | clarkb: is the restart done? | 23:14 |
| clarkb | jhesketh: apache restart is | 23:14 |
| *** bookwar has quit IRC | 23:14 | |
| *** Ryan_Lane has joined #openstack-infra | 23:14 | |
| *** gmurphy has joined #openstack-infra | 23:14 | |
| clarkb | jhesketh: I guess we can try opening that one from last night? | 23:14 |
| jhesketh | clarkb: http://logs.openstack.org/16/104416/1/experimental/experimental-swift-logs-config-pep8/19640d0e1a0d4dbb99b568e6f0d6fa48consoleText | 23:15 |
| jhesketh | :-( | 23:15 |
| lifeless | hmm, I'm going to push the requests fix to the bottom of my stack, its clearly a win | 23:15 |
| jhesketh | static logs are working though so at least we haven't broken the world | 23:15 |
| jhesketh | clarkb: mind taking a poke at the access/error logs to see if there are any hints please? | 23:15 |
| * devananda notices lifeless' gertty patch series is fixing the same things | 23:15 | |
| lifeless | incoming, sorry. | 23:15 |
| clarkb | jhesketh: sure | 23:15 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Use a requests.Session object to enable pooling. https://review.openstack.org/104715 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Hide fully reviewed projects by default https://review.openstack.org/99591 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Don't crash on comments on unchanged files https://review.openstack.org/99563 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Attempt to fetch all new refs for a change at once https://review.openstack.org/104704 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Support comments in commits https://review.openstack.org/99272 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Only show verified reviews by default. https://review.openstack.org/100138 | 23:16 |
| openstackgerrit | lifeless proposed a change to stackforge/gertty: Filter out WIP patches by default in changes list. https://review.openstack.org/100121 | 23:16 |
| *** pballand has joined #openstack-infra | 23:16 | |
| lifeless | devananda: same things? | 23:16 |
| devananda | lifeless: same things that have been bugging me to the point i was considering finding time to fix them | 23:16 |
| lifeless | oh right. yeah, done. | 23:17 |
| notmyname | hey I have a requirements question to ponder for the afternoon. any takers? | 23:17 |
| *** bookwar has joined #openstack-infra | 23:17 | |
| lifeless | its hellishly faster to sync with the connection pooling in place | 23:17 |
| lifeless | the git multiple refspec thing isn't quite right | 23:18 |
| dstufft | connection pooling ftw | 23:18 |
| notmyname | suppose there is a feature that is quite nice to have, but optional because all systems may not support it. does the library that enables that feature get added to requirements.txt? if not, then that feature isn't tested int he gate. but if it is, it's not quite true | 23:18 |
| mordred | notmyname: I mean ... maybe? | 23:18 |
| notmyname | mordred: there ^ | 23:18 |
| *** signed8bit has quit IRC | 23:18 | |
| mordred | notmyname: it kinda depends ... | 23:19 |
| clarkb | jhesketh: I get File does not exist: /srv/static/logs/16/104416/1/experimental/experimental-swift-logs-config-pep8 in the error log | 23:19 |
| mordred | let me give a couple of examples | 23:19 |
| *** alexpilotti has quit IRC | 23:19 | |
| clarkb | jhesketh: and access log shows a 404 for the full path | 23:19 |
| clarkb | jhesketh: I don't see any tracebacks or anything though | 23:19 |
| mordred | notmyname: for nova, mysql and postgres are both things that are optional but are clearly a good idea to test with on a regular basis | 23:19 |
| mordred | notmyname: so we put both in test-requirements.txt | 23:19 |
| clarkb | jhesketh: I am not seeing much else | 23:20 |
| mordred | notmyname: for cinder, for a while, hp3parclient was in the requirements.txt, because although you might not have a 3par, there was nothing about the python lib that prevented you from isntalling it | 23:20 |
| mordred | notmyname: although I believe they've recently removed it | 23:20 |
| notmyname | mordred: interesting. thanks | 23:20 |
| mordred | notmyname: it's definitely a judgement call case | 23:20 |
| anteaya | okay morganfainberg and dhellmann: https://review.openstack.org/#/c/101361/ can I get some reviews before I have to rebase this again? | 23:21 |
| clarkb | personally I like things to be in requirements.txt so I don't have to think about it | 23:21 |
| mordred | notmyname: s/prevented you from installing it/was hard about installing it/ | 23:21 |
| clarkb | everything is in a venv anyways | 23:21 |
| mordred | like, it was a pure-python library, so who cares | 23:21 |
| notmyname | mordred: specifically in this case, we've got someone working on some really cool zero-copy stuff in swift. but it's linux-only (and kernels >=2.6.38) | 23:21 |
| mordred | for thigns that require c libs... | 23:21 |
| jhesketh | clarkb: okay thanks | 23:21 |
| jhesketh | clarkb: so, what's interesting is that my browser has a "the requested url ... was not found..." etc whereas it looks like the wsgi app should just say "File not found" http://git.openstack.org/cgit/openstack-infra/os-loganalyze/tree/os_loganalyze/wsgi.py#n472 | 23:22 |
| dstufft | mordred: Working on PEP 440 again, there aren't any symbols in here which are super confusing with other versions systems you're aware of right -> 2!2.4.5a5.post8.dev9+foo.bar.123 (this is a horrible version but techincally it's about the most complex that is supported) | 23:22 |
| mordred | notmyname: interesting. WELL... it is possible to add a pbr setup filter that will only adda thing to install_requires on a certain platform | 23:22 |
| *** james_li has quit IRC | 23:22 | |
| mordred | dstufft: holy crap. that's the craziest version ever | 23:22 |
| notmyname | mordred: test-requirements may be sufficient | 23:22 |
| dstufft | mordred: yea lol, it's using absolutely everything from PPE 440 | 23:22 |
| dstufft | mordred: which is super flexible so we can capture as many different use cases as possible to minimize the arguments about why people can't use PEP 440, minor syntax changes are easier to convince people of than telling them how they version is wrong | 23:23 |
| mordred | notmyname: cool. if you want to see platform conditional example, neutron/hooks.py in neutron | 23:23 |
| notmyname | mordred: next question is if the python bindings to the c kernel stuff for this are GPL, what problems does that cause? | 23:23 |
| mordred | notmyname: which is referenced in setup.py as neutron.hooks.setup_hooks | 23:23 |
| clarkb | jhesketh: hrm. | 23:23 |
| clarkb | jhesketh: there si clearly a 404 coming out according to access log | 23:24 |
| mordred | notmyname: that is an interesting and fascinating question | 23:24 |
| *** torgomatic has joined #openstack-infra | 23:24 | |
| notmyname | mordred: ok, that's a good enough answer for now :-) | 23:24 |
| *** alexandra_ has joined #openstack-infra | 23:25 | |
| *** melwitt has joined #openstack-infra | 23:25 | |
| *** asettle has quit IRC | 23:26 | |
| *** alexandra_ is now known as asettle | 23:26 | |
| clarkb | jhesketh: is it possile apache is doing munging of the 404? | 23:26 |
| jhesketh | yeah, that's a good thought | 23:27 |
| *** shashankhegde has joined #openstack-infra | 23:27 | |
| jhesketh | clarkb: does the user have read permissions? | 23:27 |
| clarkb | jhesketh: oh gah maybe I got that wrong | 23:28 |
| clarkb | www-data can read it | 23:29 |
| clarkb | is that not the right user? | 23:29 |
| clarkb | apache is running as that use | 23:29 |
| jhesketh | clarkb: I meant the swift user | 23:29 |
| jhesketh | how is that configured? | 23:29 |
| clarkb | oh maybe, let me login as that user | 23:30 |
| jhesketh | clarkb: sigh, I think I've screwed up | 23:30 |
| jhesketh | clarkb: I set it up as infra-files-ro apparently | 23:31 |
| *** mtreinish has quit IRC | 23:31 | |
| clarkb | no I did that | 23:31 |
| clarkb | and that user can see the file and the container | 23:31 |
| jhesketh | oh okay, I must have missed that change | 23:31 |
| clarkb | I wonder | 23:31 |
| clarkb | if the tenant is different for that user | 23:32 |
| clarkb | which it probably is and that may explain it | 23:32 |
| clarkb | I wish this info was easier to get | 23:32 |
| clarkb | well maybe tenant isn't different because these are users agains tthe same thing | 23:32 |
| lifeless | devananda: try my stack, its working well for me | 23:32 |
| clarkb | jhesketh: there is no swift in my path on that box | 23:33 |
| lifeless | devananda: be sure to pip install the 0.3 branch of https://github.com/gitpython-developers/GitPython | 23:33 |
| *** zz_dimtruck is now known as dimtruck | 23:33 | |
| clarkb | jhesketh: is that it? do we need to install swift client? | 23:33 |
| mordred | clarkb: what was the puppet sequencing thing we uncovered the other day? | 23:33 |
| jhesketh | that's quite possible | 23:33 |
| lifeless | devananda: there's one more obvious fix to do, which is to coalesce all the git fetches into one big batch so we get much better compression | 23:34 |
| mordred | clarkb: it's taht we need to run pupept on jenkins masters then on zuul, right? | 23:34 |
| jhesketh | clarkb: just the python swiftclient | 23:34 |
| clarkb | I am going to manually pip install swift client then resetart apache | 23:34 |
| clarkb | jhesketh: ya | 23:34 |
| lifeless | rather than one per review push | 23:34 |
| lifeless | I have a change to do one per review | 23:34 |
| jhesketh | clarkb: good catch | 23:34 |
| clarkb | mordred: yes, so that the jobs are registered before zuul runs them | 23:34 |
| lifeless | but if you have 100 open reviews thats still v inefficient | 23:34 |
| mordred | clarkb: thanks | 23:34 |
| clarkb | jhesketh: should that be added to the requirements for os-loganalyze? | 23:35 |
| devananda | lifeless: GitPython==0.3.2.RC1 ? | 23:35 |
| jhesketh | clarkb: yep, will fix that up | 23:35 |
| clarkb | jhesketh: I think that is the simplest way to make it work in a managed way | 23:35 |
| lifeless | devananda: no, from git - I had a PR accepted last night | 23:35 |
| clarkb | apache is restarting again | 23:35 |
| devananda | ack | 23:35 |
| clarkb | still 404ing | 23:35 |
| clarkb | I can still get static hosted logs though | 23:36 |
| lifeless | devananda: its referenced in my stack, and there's backwards compat code too, I think :) | 23:36 |
| lifeless | ok offline | 23:36 |
| clarkb | now to read more logs | 23:36 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/os-loganalyze: Add swiftclient into requirements.txt https://review.openstack.org/104718 | 23:36 |
| clarkb | jhesketh: same error and access log entries | 23:37 |
| jhesketh | clarkb: hmm, so maybe we should figure out how to catch less rather than all exceptions | 23:37 |
| clarkb | jhesketh: or at least log them | 23:37 |
| clarkb | then they will end up in the error log and we can debug | 23:37 |
| jhesketh | yep | 23:37 |
| clarkb | jhesketh: image builds are done we should do a check experimental | 23:38 |
| jhesketh | clarkb: do I just write to stderr for the wsgi app to log to the apache error log or do I need to do something fancy | 23:38 |
| clarkb | I think that is what you do | 23:38 |
| clarkb | jhesketh: though it is curious that the error I have is that the dir does not exist | 23:39 |
| clarkb | jhesketh: it isn't complaining about the full path | 23:39 |
| clarkb | jhesketh: File does not exist: /srv/static/logs/16/104416/1/experimental/experimental-swift-logs-config-pep8 | 23:39 |
| clarkb | jhesketh: I wonder if something is checking the dirname? | 23:39 |
| *** mtreinish has joined #openstack-infra | 23:39 | |
| jhesketh | maybe | 23:40 |
| clarkb | actualy | 23:40 |
| * clarkb reads chost | 23:40 | |
| clarkb | *vhost | 23:40 |
| clarkb | jhesketh: yup its the vhost | 23:41 |
| clarkb | jhesketh: it only rewrites file extensions it knows about | 23:41 |
| clarkb | jhesketh: everything else is served direct | 23:41 |
| jhesketh | ah cool | 23:41 |
| jhesketh | that explains the apache error | 23:41 |
| clarkb | yup | 23:41 |
| clarkb | so that will need sorting out... | 23:42 |
| clarkb | and my day is just about done + holidays | 23:42 |
| clarkb | but still progress | 23:42 |
| jhesketh | yeah | 23:42 |
| jhesketh | thanks clarkb :-) | 23:42 |
| jhesketh | clarkb: actually do you have a minute to check the error log again? | 23:44 |
| clarkb | jhesketh: what are we looking for? | 23:44 |
| jhesketh | it looks like the vhost just rewrites to htmilfy/$1 which gives you this: http://logs.openstack.org/htmlify/16/104416/1/experimental/experimental-swift-logs-config-pep8/19640d0e1a0d4dbb99b568e6f0d6fa48consoleText | 23:44 |
| clarkb | jhesketh: but only if it is a .html or .gz | 23:45 |
| jhesketh | right, but I just put htmlify into the url | 23:45 |
| jhesketh | and now I get a 500 error | 23:45 |
| clarkb | oh | 23:45 |
| clarkb | gotcha ya give me a minute I suspended the box | 23:45 |
| *** bdpayne has quit IRC | 23:45 | |
| jhesketh | sorry ! | 23:45 |
| *** msabramo has joined #openstack-infra | 23:45 | |
| *** msabramo1 has quit IRC | 23:46 | |
| clarkb | Auth version 2.0 requires python-keystoneclient, install it or use Auth | 23:46 |
| clarkb | I geuss I am installing python-keystoneclient too? | 23:47 |
| * clarkb does that and restarts apache | 23:47 | |
| clarkb | notmyname: ^ is a thing that should just be installed imo | 23:47 |
| clarkb | notmyname: but this has been argued before so meh | 23:47 |
| notmyname | clarkb: ie you mean in the hard requirements of python-swiftclient? | 23:47 |
| clarkb | notmyname: yes | 23:48 |
| clarkb | notmyname: now its file not found | 23:48 |
| clarkb | er jhesketh ^ | 23:48 |
| clarkb | jhesketh: but nothing in error log because we don't log anything in that case | 23:48 |
| jhesketh | clarkb: oh yeah (I also added that to the requirements.txt) | 23:49 |
| jhesketh | well that's progress too! :-) | 23:49 |
| jhesketh | I'll get this os-loganalyze debug stuff up | 23:49 |
| clarkb | cool | 23:49 |
| clarkb | notmyname: having the ability to auth v2 doesn't hurt anyone that is using v1 | 23:50 |
| clarkb | notmyname: but not having it breaks lots of people | 23:50 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/os-loganalyze: Log swift errors to stderr https://review.openstack.org/104719 | 23:52 |
| jhesketh | clarkb: ^ | 23:52 |
| clarkb | jhesketh: you might want to use traceback.print_exc | 23:53 |
| clarkb | jhesketh: as that will give you a traceback too | 23:53 |
| clarkb | and you should be able to pass sys.stderr as the file | 23:53 |
| jhesketh | clarkb: good idea | 23:54 |
| jhesketh | looks like sys.stderr is the default file | 23:54 |
| clarkb | even better | 23:55 |
| clarkb | I -1'd it and jenkins was unhappy too | 23:55 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/os-loganalyze: Log swift errors to stderr https://review.openstack.org/104719 | 23:55 |
| clarkb | huh double requirement? | 23:55 |
| clarkb | I don't understand why the tests are failing | 23:55 |
| clarkb | but they are not happy | 23:55 |
| clarkb | oh its in test requirements too | 23:56 |
| clarkb | you need to remove it from there | 23:56 |
| *** homeless has quit IRC | 23:57 | |
| jhesketh | clarkb: right, except I copied the requirements from zuul | 23:57 |
| jhesketh | so not sure why they aren't being installed | 23:57 |
| *** derekh_ has joined #openstack-infra | 23:57 | |
| clarkb | jhesketh: test requirements are not installed by default | 23:58 |
| clarkb | so we need to put them in requirements.txt and remove from test-requirements.txt | 23:58 |
| jhesketh | oh I see | 23:59 |
| *** oomichi has joined #openstack-infra | 23:59 | |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/os-loganalyze: Add swiftclient into requirements.txt https://review.openstack.org/104718 | 23:59 |
| openstackgerrit | Joshua Hesketh proposed a change to openstack-infra/os-loganalyze: Log swift errors to stderr https://review.openstack.org/104719 | 23:59 |
| jhesketh | clarkb: ^ | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!