| SpamapS | ugh | 00:53 |
|---|---|---|
| SpamapS | need to spend some time debugging ec2+nodepool | 00:53 |
| SpamapS | just found 7 leaked instances | 00:53 |
| SpamapS | http://paste.openstack.org/show/734843/ | 00:58 |
| SpamapS | I think that's the culprit | 00:58 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/nodepool master: WIP: Implement an Amazon EC2 driver https://review.openstack.org/535558 | 01:02 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/nodepool master: AWS driver fix to not explode on common images https://review.openstack.org/601361 | 01:02 |
| SpamapS | something very broken in that driver too | 01:16 |
| SpamapS | I say max-servers: 5, but it just keeps adding servers until the cloud refuses. | 01:16 |
| SpamapS | dmsimard: tristanC ^^ have you all been playing with that driver much since we last talked about it? | 01:19 |
| * SpamapS will try and whip it in to shape | 01:19 | |
| dmsimard | SpamapS: I have not | 03:43 |
| tristanC | SpamapS: me neither, and my 12 month period amazon account is now over... | 03:46 |
| *** bjackman has quit IRC | 03:58 | |
| *** bjackman has joined #zuul | 04:17 | |
| SpamapS | hah ok | 04:50 |
| *** chandankumar has joined #zuul | 04:59 | |
| *** chandankumar is now known as chkumar|ruck | 05:03 | |
| *** bhavikdbavishi has joined #zuul | 05:05 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/nodepool master: Amazon EC2 driver https://review.openstack.org/535558 | 05:09 |
| SpamapS | ^^ tests fixed, schema normalized closer to the openstack one. | 05:10 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul-jobs master: Add a markdownlint job and role https://review.openstack.org/607691 | 05:16 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul-jobs master: Add a markdownlint job and role https://review.openstack.org/607691 | 05:35 |
| *** rlandy has quit IRC | 05:48 | |
| *** gundalow has quit IRC | 05:58 | |
| *** samccann has quit IRC | 06:00 | |
| *** corvus has quit IRC | 06:00 | |
| *** gundalow has joined #zuul | 06:00 | |
| *** samccann has joined #zuul | 06:00 | |
| *** threestrands has quit IRC | 06:01 | |
| *** pcaruana has joined #zuul | 07:19 | |
| *** goern has joined #zuul | 07:24 | |
| *** bhavikdbavishi has quit IRC | 07:53 | |
| *** quiquell is now known as quiquell|bbl | 08:10 | |
| *** sshnaidm|pto is now known as sshnaidm | 08:15 | |
| *** jpena|off is now known as jpena | 08:35 | |
| *** bhavikdbavishi has joined #zuul | 08:39 | |
| *** bhavikdbavishi has quit IRC | 08:40 | |
| *** bhavikdbavishi has joined #zuul | 08:40 | |
| *** ParsectiX has joined #zuul | 08:47 | |
| *** ParsectiX has quit IRC | 08:48 | |
| *** ParsectiX has joined #zuul | 08:48 | |
| *** bhavikdbavishi1 has joined #zuul | 08:59 | |
| *** bhavikdbavishi has quit IRC | 08:59 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 08:59 | |
| *** bjackman has quit IRC | 09:05 | |
| *** bhavikdbavishi has quit IRC | 09:15 | |
| *** hashar has joined #zuul | 09:22 | |
| *** sshnaidm is now known as sshnaidm|afk | 09:26 | |
| *** corvus has joined #zuul | 09:56 | |
| *** sshnaidm|afk is now known as sshnaidm | 10:02 | |
| *** bjackman has joined #zuul | 10:09 | |
| *** sdake has quit IRC | 10:11 | |
| *** sdake has joined #zuul | 10:14 | |
| *** kklimonda has joined #zuul | 10:17 | |
| *** ParsectiX has quit IRC | 10:19 | |
| *** ParsectiX has joined #zuul | 10:38 | |
| *** ParsectiX has quit IRC | 11:00 | |
| *** ParsectiX has joined #zuul | 11:04 | |
| *** panda|rover|off is now known as panda|rover | 11:08 | |
| *** dkehn has quit IRC | 11:32 | |
| *** bhavikdbavishi has joined #zuul | 11:39 | |
| *** ParsectiX has quit IRC | 11:47 | |
| *** hashar has quit IRC | 11:50 | |
| *** ParsectiX has joined #zuul | 11:55 | |
| *** ParsectiX has quit IRC | 11:56 | |
| *** electrofelix has joined #zuul | 12:00 | |
| *** jpena is now known as jpena|lunch | 12:02 | |
| tobiash | @zuulers here in Berlin, have to leave now. It was great to see you again and hopefully cya next time in Denver :) | 12:08 |
| *** quiquell|bbl is now known as quiquell | 12:09 | |
| openstackgerrit | Quique Llorente proposed openstack-infra/zuul master: Prevent exception at executor https://review.openstack.org/617984 | 12:14 |
| tristanC | tobiash: have a safe trip home! | 12:17 |
| quiquell | fungi: ^ | 12:17 |
| quiquell | tristanC: Hello there, found small issue at cloning on executors | 12:17 |
| quiquell | tristanC: Maybe you can take a look at the review | 12:17 |
| tristanC | quiquell: i think the parameter is exist_ok | 12:20 |
| tobiash | quiquell, corvus: I think there was a fix race, you both have identical patches in review ^ ;) | 12:20 |
| tobiash | tristanC: thanks, u2 | 12:21 |
| quiquell | tristanC: Damn, fixing... | 12:23 |
| quiquell | tristanC: I test locally exist_ok then write it wrong :-) | 12:23 |
| openstackgerrit | Quique Llorente proposed openstack-infra/zuul master: Prevent exception at executor https://review.openstack.org/617984 | 12:23 |
| quiquell | tristanC, tobiash: Done | 12:23 |
| *** bjackman has quit IRC | 12:28 | |
| openstackgerrit | Quique Llorente proposed openstack-infra/zuul master: Prevent exception at executor https://review.openstack.org/617984 | 12:33 |
| quiquell | Refactored another makedirs in the same file | 12:34 |
| *** hashar has joined #zuul | 12:34 | |
| AJaeger | tobiash: great seeing you here! Safe travels home! | 12:41 |
| quiquell | tristanC: Do you know if there are any option to disable host key verificatoin at executors ? | 12:43 |
| pabelanger | tobiash: o/ always great to meet up | 12:55 |
| *** rlandy has joined #zuul | 12:56 | |
| *** ParsectiX has joined #zuul | 12:59 | |
| *** bhavikdbavishi has quit IRC | 13:08 | |
| *** chkumar|ruck is now known as chkumar|ruck|brb | 13:15 | |
| *** chkumar|ruck|brb is now known as chkumar|ruck | 13:18 | |
| *** ianychoi has joined #zuul | 13:24 | |
| *** dkehn has joined #zuul | 13:35 | |
| *** jpena|lunch is now known as jpena | 13:42 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: WIP Add support for zones in executors https://review.openstack.org/549197 | 13:47 |
| pabelanger | rcarrillocruz: corvus: just a rebase, going to start adding tests now ^ | 13:48 |
| pabelanger | and docs | 13:48 |
| quiquell | Hi there | 14:00 |
| quiquell | I am facing this Timeout exception waiting for the logger. Please check connectivity to [38.145.34.145:19885] | 14:00 |
| quiquell | Playing with zuul at my machine | 14:00 |
| quiquell | It means that I have to open some kind of port at the security group at my cloud provider ? | 14:01 |
| *** ParsectiX has quit IRC | 14:01 | |
| clarkb | quiquell: the way console logging works is the job ansible runs a plugin hook thing (I forget actual term) on port 19885 that the executor reads from | 14:01 |
| clarkb | quiquell: so ya need to have that open to the executor | 14:01 |
| quiquell | clarkb: executor or node ? | 14:02 |
| *** chkumar|ruck has quit IRC | 14:02 | |
| clarkb | executor opens tcp connection to port 19885 on the test node | 14:02 |
| quiquell | clarkb: So the port has to be accesible from the node targeted by ansible, is that it ? | 14:02 |
| quiquell | clarkb: I mean nodes configured at nodepool | 14:03 |
| clarkb | it has to be opened on the node targetted by ansible | 14:03 |
| *** ParsectiX has joined #zuul | 14:03 | |
| quiquell | clarkb: Cool thanks so much, do you know any mort port needed there ? | 14:03 |
| clarkb | quiquell: I think ssh and 19885 are it and thinking about this more it might make sense to hvae zuul tunnel that connection over ssh since ssh is already assumed to be open | 14:04 |
| clarkb | mordred: ^ fyi | 14:04 |
| quiquell | clarkb: Humm that's a good one, It's not done because performance ? | 14:04 |
| clarkb | quiquell: or just an oversight? I'm not sure | 14:04 |
| quiquell | clarkb: Me neither I am noob here :-), but like it | 14:05 |
| Diabelko | pabelanger: here | 14:15 |
| pabelanger | Diabelko: o/ | 14:15 |
| *** hashar has quit IRC | 14:20 | |
| tobiash | clarkb: I just saw a picture of zuul socks... Nice! | 14:23 |
| AJaeger | tobiash: please send a photo ;) | 14:26 |
| corvus | clarkb, quiquell: we should totally tunnel the port. mordred's new idea for how to do logginng will obsolete that (he actually wants to tunnel a unix domain socket), but i think we can implement tunelling the tcp port now fairly easily. | 14:34 |
| quiquell | corvus: It's difficult to implement, I would like to put some reviews for that | 14:35 |
| quiquell | ? | 14:35 |
| quiquell | Was a question forgot the question mark ? | 14:35 |
| corvus | quiquell: i think the tunnel needs to be set up when the ssh agent is started; i think mordred's unix domain work may do something similar; we may be able to copy that. | 14:36 |
| quiquell | corvus: ack, thanks for info though | 14:37 |
| corvus | i would characterize this as moderate difficulty -- the code involved can be a little hard to follow, but it's not a complex problem | 14:37 |
| corvus | mostly requires looking into the log callback plugin and zuul_console | 14:37 |
| quiquell | corvus: 80 read 20 write relation | 14:40 |
| *** ParsectiX has quit IRC | 15:08 | |
| quiquell | corvus, clarkb: Do you know when do we create /etc/ci/mirror_info.sh ? | 15:08 |
| quiquell | Humm this is more infra than zuul isue | 15:09 |
| *** ParsectiX has joined #zuul | 15:37 | |
| *** ParsectiX has quit IRC | 16:04 | |
| *** ParsectiX has joined #zuul | 16:04 | |
| *** ParsectiX has quit IRC | 16:11 | |
| pabelanger | quiquell: yah, it is in openstack-zuul-jobs mirror-info role | 16:18 |
| pabelanger | IIRC we do it in base job for openstack | 16:18 |
| quiquell | pabelanger: Yap found it thanks | 16:18 |
| *** bjackman has joined #zuul | 16:26 | |
| *** chandankumar has joined #zuul | 16:26 | |
| *** chandankumar has quit IRC | 16:30 | |
| *** bjackman has quit IRC | 16:36 | |
| *** bhavikdbavishi has joined #zuul | 16:42 | |
| *** quiquell has quit IRC | 16:46 | |
| *** bhavikdbavishi has quit IRC | 16:46 | |
| *** bjackman has joined #zuul | 16:51 | |
| *** bjackman has quit IRC | 16:58 | |
| *** j^2 has joined #zuul | 17:14 | |
| *** j^2 has quit IRC | 17:16 | |
| *** j^2 has joined #zuul | 17:16 | |
| *** sshnaidm is now known as sshnaidm|afk | 17:37 | |
| *** chandankumar has joined #zuul | 17:44 | |
| dmsimard | Would there be a way for Zuul to automatically enqueue a change that depended on a change that merged ? | 17:48 |
| dmsimard | Pretend there's changes A and B, B has been approved and depends on A. A hasn't been merged yet -- if I approve A, it doesn't kick the jobs for B | 17:49 |
| dmsimard | (hopefully that makes sense) | 17:49 |
| dmsimard | A concrete example would be https://review.openstack.org/#/c/618018/ which I had to re-approve | 17:51 |
| pabelanger | dmsimard: do they share the same change queue for gate? | 17:59 |
| pabelanger | sounds like they don't | 17:59 |
| *** panda|rover is now known as panda|rover|off | 18:05 | |
| *** zigo has quit IRC | 18:06 | |
| dmsimard | pabelanger: I didn't set one so I guess they're in their own queues -- I'll check that out, thanks | 18:15 |
| *** chandankumar has quit IRC | 18:17 | |
| *** jpena is now known as jpena|off | 18:34 | |
| *** j^2 has quit IRC | 18:59 | |
| *** hashar has joined #zuul | 19:19 | |
| *** electrofelix has quit IRC | 19:31 | |
| SpamapS | Seems like free CI services are popping up every day | 19:43 |
| SpamapS | https://meta.sr.ht/ | 19:43 |
| SpamapS | dmsimard: isn't what you're asking handled by this: https://zuul-ci.org/docs/zuul/admin/drivers/zuul.html?highlight=parent%20change%20enqueued ? | 19:46 |
| *** sshnaidm|afk is now known as sshnaidm|off | 20:00 | |
| *** pcaruana has quit IRC | 20:14 | |
| *** pcaruana has joined #zuul | 20:27 | |
| manjeets | corvus, Hi ping quick question regarding nodepool | 20:46 |
| manjeets | https://github.com/openstack-infra/zuul/blob/master/doc/source/admin/examples/etc_nodepool/nodepool.yaml | 20:46 |
| manjeets | instead of using some provider cloud, would I be able to use a pool of baremetal machines (ip list) | 20:47 |
| *** j^2 has joined #zuul | 21:35 | |
| *** hashar has quit IRC | 21:45 | |
| *** pcaruana has quit IRC | 21:55 | |
| ianw | manjeets: i think what you want is the static driver -> https://zuul-ci.org/docs/zuul/admin/nodepool_static.html | 22:38 |
| manjeets | cool so using ip as name + ssh keys might work | 23:09 |
| manjeets | thanks ianw | 23:09 |
| *** j^2 has quit IRC | 23:22 | |
| ianw | manjeets: yes, although do you plan to re-use the nodes? | 23:39 |
| manjeets | ianw, so idea is to have a cleanup script for cleanning the node after test run as part of job itself ? | 23:40 |
| manjeets | would that work ? | 23:40 |
| ianw | manjeets: i think it depends on what you're doing; i think yes that can work, but you can also envisage scenarios where jobs go wrong a destroy machines or do things the cleanup doesn't handle | 23:43 |
| ianw | for example, well before secret support, we used static nodes with secret keys deployed on them for things like signing | 23:44 |
| ianw | if it's complete general purpose workloads, however, maybe you want to investigate something like using ironic to manage your baremetal and working via that | 23:45 |
| manjeets | ianw we need baremetal nodes for hardware testing, other thing that could be done is run cleanup before starting a job and after job, but something crazy can happen in b/w | 23:45 |
| manjeets | in that case I think using cloud with ironic could be a option too ? | 23:46 |
| ianw | tbh, it sounds like that's probably where you want to end up. it seems like it would be nice for your environment to be managing your hardware like that | 23:47 |
| ianw | but, certainly no reason not to start with static nodes and work from there | 23:47 |
| manjeets | ianw I'm kind of learning the whole pipeline, the initial goal is get at run of job atleast once on baremetal and then if that works fine, we just need to change the nodepool config to replace static nodes with ironic cloud ? if I understand it right | 23:49 |
| ianw | well yeah, i mean there's a lot going on so minimising your moving parts to start is probably helpful :) | 23:50 |
| ianw | if you get the zuul part up and running, taking the changes and running the job as you want on a static node that seems like a good start | 23:51 |
| ianw | then really it shouldn't be too different to get that plugged into a cloud that's managing your hardware via ironic | 23:52 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!