| *** odyssey4me has quit IRC | 00:02 | |
| *** odyssey4me has joined #zuul | 00:03 | |
| *** rlandy has quit IRC | 00:03 | |
| dmsimard | corvus: would it be appropriate to install the mysql-client package on the zuul scheduler node in puppet-zuul ? I wanted to check something but it's not installed :( | 00:19 |
|---|---|---|
| clarkb | if we want to do backups of that db we probably want those tools too | 00:19 |
| dmsimard | how are database backups done today ? is that automated by trove ? | 00:25 |
| clarkb | no, we use mysql dump to the local host then bup backs those files up to the bup server | 00:27 |
| dmsimard | so I guess that's broken right now then | 00:27 |
| * dmsimard looks | 00:27 | |
| clarkb | one of the goals there was to preserve append only backups like we have with the rest of our backups | 00:28 |
| dmsimard | I know the zuulv3 node had it because I remember using it once | 00:28 |
| dmsimard | so we lost it with the change to zuul01 | 00:28 |
| clarkb | I'm not sure we ever had backups for it though | 00:28 |
| corvus | dmsimard, clarkb: let's bmove to #openstack-infra | 00:39 |
| dmsimard | corvus: how about we send the puppet-zuul gerritbot notifications here ? | 00:52 |
| *** JasonCL_ has joined #zuul | 00:56 | |
| *** JasonCL has quit IRC | 00:58 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: [WIP] Test growroot in boot tests https://review.openstack.org/555103 | 00:59 |
| corvus | dmsimard: right now, i'd rather not. i'm unaware of anyone not openstack-related using it. | 01:05 |
| dmsimard | That's why I figured I should ask first, thanks for confirming :D | 01:05 |
| *** harlowja has quit IRC | 01:31 | |
| *** dtruong2 has joined #zuul | 02:09 | |
| *** dtruong2 has quit IRC | 02:29 | |
| *** harlowja has joined #zuul | 03:32 | |
| *** harlowja has quit IRC | 03:54 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: [WIP] Test growroot in boot tests https://review.openstack.org/555103 | 04:07 |
| *** persia has quit IRC | 04:13 | |
| *** persia has joined #zuul | 04:14 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: add a job build button https://review.openstack.org/554839 | 05:36 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: job-trigger: add initial driver and event https://review.openstack.org/555153 | 05:36 |
| *** JasonCL has joined #zuul | 05:40 | |
| *** JasonCL_ has quit IRC | 05:44 | |
| *** threestrands has quit IRC | 06:25 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: job-trigger: add initial driver and event https://review.openstack.org/555153 | 07:39 |
| *** Wei_Liu1 has joined #zuul | 07:46 | |
| *** Wei_Liu has quit IRC | 07:46 | |
| *** Wei_Liu1 is now known as Wei_Liu | 07:46 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Clarify position of depends-on footer https://review.openstack.org/555185 | 07:53 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Clarify location of depends-on footer https://review.openstack.org/555185 | 07:53 |
| tobiash | fixes doc bug ^ | 07:56 |
| *** flepied_ has joined #zuul | 08:22 | |
| *** snapiri has joined #zuul | 08:25 | |
| *** flepied_ has quit IRC | 08:31 | |
| *** electrofelix has joined #zuul | 08:35 | |
| *** yolanda_ is now known as yolanda | 08:35 | |
| *** JasonCL has quit IRC | 08:38 | |
| *** jpena|off is now known as jpena | 08:48 | |
| *** flepied_ has joined #zuul | 08:58 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: job-trigger: add initial driver and event https://review.openstack.org/555153 | 09:15 |
| tobiash | does someone already use nodepool with boot-from-volume? | 09:18 |
| tobiash | https://docs.openstack.org/infra/nodepool/configuration.html#pool-labels | 09:18 |
| *** jesusaur has quit IRC | 09:20 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: [WIP] Test growroot in boot tests https://review.openstack.org/555103 | 09:24 |
| *** jesusaur has joined #zuul | 09:28 | |
| openstackgerrit | Bernd Bausch proposed openstack-infra/zuul master: Add port to webhook URL in Zuul github driver docu https://review.openstack.org/555225 | 09:47 |
| openstackgerrit | Bernd Bausch proposed openstack-infra/zuul master: Add port to webhook URL in Zuul github driver docu https://review.openstack.org/555225 | 09:52 |
| *** hashar has joined #zuul | 09:58 | |
| tristanC | tobiash: we did use this option with nodepool-v2 on review.rdoproject.org | 09:59 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: add a job build button https://review.openstack.org/554839 | 10:00 |
| tobiash | tristanC: ah cool, so I hope it's reliable in deleting the volumes together with the nodes :) | 10:00 |
| tristanC | tobiash: it should as it set the terminate_volume create server option | 10:01 |
| tobiash | ah cool | 10:01 |
| *** sshnaidm has quit IRC | 11:00 | |
| *** jesusaur has quit IRC | 11:02 | |
| *** logan- has quit IRC | 11:03 | |
| *** logan- has joined #zuul | 11:03 | |
| *** jesusaur has joined #zuul | 11:12 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add port to webhook URL in Zuul github driver docu https://review.openstack.org/555225 | 11:12 |
| mordred | tristanC, tobiash: we actually just rolled a use of boot-from-volume in nodepool live in infra for vexxhost | 11:15 |
| tobiash | :) | 11:16 |
| mordred | they're a ceph cloud, so boot-from-volume + uploading raw images means _very_ inexpensive/quick boot times | 11:16 |
| tobiash | my use case is support for larger disks than the flavor | 11:16 |
| mordred | tobiash: good use case | 11:16 |
| tobiash | yah that's cool, our cloud is also ceph only | 11:16 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: [WIP] Test growroot in boot tests https://review.openstack.org/555103 | 11:21 |
| *** odyssey4me has quit IRC | 12:07 | |
| *** odyssey4me has joined #zuul | 12:08 | |
| *** sshnaidm has joined #zuul | 12:25 | |
| *** rlandy has joined #zuul | 12:29 | |
| kklimonda | hmm, any idea how to debug error "SSH Error: data could not be sent to remote host X. Make sure this host can be reached over ssh" when using add-fileserver and upload-logs ? The logserver only logs the following in auth.log: "Did not receive identification string from 192.168.0.5" | 12:34 |
| kklimonda | it's failing on "[upload-logs : Create log directories]" | 12:35 |
| tobiash | kklimonda: plain ssh with -vvv helps often | 12:49 |
| kklimonda | tobiash: can I somehow force ansible to use -vvv for ssh and print output? | 12:51 |
| tobiash | kklimonda: you can switch the executor into verbose mode | 12:52 |
| tobiash | 'zuul-executor verbose' | 12:52 |
| tobiash | then you find these logs in the debug logging | 12:52 |
| kklimonda | yes, I was running executor with verbose already - that did not display connection info, but that's because we don't add enough -v's | 12:56 |
| kklimonda | well, that doesn't actually print anything more :/ | 12:58 |
| *** electrofelix has quit IRC | 12:58 | |
| Shrews | protip: Do not attempt to use devstack with nodepool. Use the static driver b/c OMGSOMUCHEASIER | 13:01 |
| *** dkranz has joined #zuul | 13:02 | |
| kklimonda | all I get is this measly log http://paste.openstack.org/show/708949/ | 13:02 |
| *** JasonCL has joined #zuul | 13:18 | |
| *** sshnaidm has quit IRC | 13:19 | |
| *** sshnaidm has joined #zuul | 13:20 | |
| *** dmsimard has quit IRC | 13:21 | |
| *** JasonCL has quit IRC | 13:23 | |
| *** JasonCL has joined #zuul | 13:24 | |
| *** JasonCL has quit IRC | 13:28 | |
| *** JasonCL has joined #zuul | 13:32 | |
| *** dmsimard has joined #zuul | 13:34 | |
| kklimonda | looks like I've managed to run my zuul-executor as root and that broke things? I saw a weird error when trying to ssh to the logserver from an ansible task: http://paste.openstack.org/show/708955/ | 13:42 |
| *** JasonCL has quit IRC | 13:45 | |
| *** JasonCL has joined #zuul | 13:54 | |
| *** JasonCL has quit IRC | 14:08 | |
| *** JasonCL has joined #zuul | 14:09 | |
| *** JasonCL has quit IRC | 14:24 | |
| *** JasonCL has joined #zuul | 14:26 | |
| *** JasonCL has quit IRC | 14:35 | |
| *** JasonCL has joined #zuul | 14:37 | |
| *** JasonCL has quit IRC | 14:39 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add autohold debug info https://review.openstack.org/555327 | 14:53 |
| *** JasonCL has joined #zuul | 14:57 | |
| *** JasonCL has quit IRC | 14:58 | |
| *** JasonCL has joined #zuul | 14:58 | |
| *** swest has quit IRC | 15:10 | |
| *** hashar is now known as hasharAway | 15:13 | |
| Shrews | corvus: did you mean to output the autohold debug line when the key is None? | 15:21 |
| Shrews | i guess knowing if None is returned could be useful there | 15:22 |
| corvus | Shrews: yes, that's it exactly | 15:23 |
| *** JasonCL has quit IRC | 15:35 | |
| *** JasonCL has joined #zuul | 15:36 | |
| pabelanger | kklimonda: how did your executor start as root? Left over from the time when the executor would drop permissions to bind to finger port? | 15:37 |
| kklimonda | pabelanger: yeah, I have my own systemd scripts used in another deployment and that deployment hasn't been uploaded in a while | 15:39 |
| pabelanger | ack | 15:39 |
| *** JasonCL has quit IRC | 15:44 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add autohold debug info https://review.openstack.org/555327 | 15:44 |
| *** JasonCL has joined #zuul | 15:45 | |
| *** JasonCL has quit IRC | 15:46 | |
| *** JasonCL has joined #zuul | 15:46 | |
| *** harlowja has joined #zuul | 16:04 | |
| *** JasonCL has quit IRC | 16:06 | |
| *** JasonCL has joined #zuul | 16:07 | |
| *** harlowja has quit IRC | 16:09 | |
| *** rlandy is now known as rlandy|biab | 16:36 | |
| *** sshnaidm is now known as sshnaidm|off | 16:38 | |
| pabelanger | question about private clouds and nodepool / zuul. I know today zuul-executor would need to be able to access the private IP, but with nodepool-launcher assuming openstack APIs are public and don't want a public ip address for nodes, I still think nodepool-launcher needs to be on private network, since we ssh-keyscan for hostkeys. Thoughts on how we could maybe fix that and only have zuul-executor on private | 16:56 |
| pabelanger | network? | 16:56 |
| clarkb | pabelanger: we talked about this at the ptg as it is a need for rcarrillocruz and mrhillsman | 16:59 |
| clarkb | pabelanger: basically you'd run zuul-executor and nodepool launcher (possibly builder too) in the private portion of the cloud and punch holes for them to talk to control plane | 17:00 |
| pabelanger | yah, that's what I was hoping. Don't think I was in the room during that time | 17:00 |
| clarkb | then the major missing piece is executor affinity to cloud resources | 17:00 |
| clarkb | the PTG etherpad has a link to a story on how to do ^ | 17:00 |
| *** tosky has joined #zuul | 17:00 | |
| pabelanger | clarkb: right, in this case, the cloud is public, but doesn't have enough FIPs for ipv4 or ipv6. So, openstack APIs would be on public network, only traffic from zuul-executor would need to be private. | 17:01 |
| pabelanger | I think if we disabled ssh-keyscan in nodepool-launcher, that too could be public | 17:01 |
| mrhillsman | i started back testing this with the request to try out zuul from scratch | 17:01 |
| pabelanger | and / or move that to executor | 17:01 |
| mrhillsman | i have not got a job to run yet but will probably test it today | 17:01 |
| mrhillsman | right now i keep nodepool-launcher/builder and zuul-executor inside of the cloud via openstack VMs | 17:02 |
| mrhillsman | everything else is in the public space | 17:02 |
| mrhillsman | right now they can all talk | 17:02 |
| mrhillsman | just need to confirm job execution happens | 17:02 |
| pabelanger | yah, I'd like to keep nodepool bits public, since they could also work on another cloud, over installing additional services | 17:03 |
| clarkb | mrhillsman: ya then if you have a mixed env you just need to specify "these labels can only be reached by this executor" and there is a detailed story on making that change to zuul | 17:03 |
| mrhillsman | yep | 17:03 |
| pabelanger | but agree zuul-executor would need to be private | 17:03 |
| mrhillsman | how would that work though pabelanger | 17:03 |
| pabelanger | if only nova gave us hostkeys for nodes :) | 17:03 |
| mrhillsman | since nodepool has to talk to the cloud | 17:04 |
| pabelanger | mrhillsman: cloud is public | 17:04 |
| pabelanger | but doesn't have enough public ips for VMs | 17:04 |
| clarkb | right separation of cloud api and cloud resources | 17:04 |
| pabelanger | yah | 17:04 |
| clarkb | the cloud api is public, the resources are not | 17:04 |
| pabelanger | exactly | 17:04 |
| mrhillsman | oh ok | 17:04 |
| mrhillsman | i have the case of both being private :) | 17:04 |
| mrhillsman | so an executor in the private cloud does not facilitate that? | 17:05 |
| pabelanger | in fact, I think this could be tested in the gate via nodepool dsvm job, and a little hacking | 17:05 |
| clarkb | mrhillsman: it does, pabelanger is talking about trying to make the nodepool launcher not private | 17:06 |
| pabelanger | mrhillsman: no, I think that is the one item that needs to be in private | 17:06 |
| pabelanger | the only thing nodepool-launcher does to a node now is ssh-keyscan it for hostkeys, because we have no other way to get that info, so it would need routing access to private network | 17:06 |
| clarkb | pabelanger: my concern with moving that out of the nodepool launcher is that even though openstack can't do it today other compute resource platforms may be able to provide those host keys to nodepool in which case you would not want to set it up in the fashion you describe because you would be lowering your level of security | 17:06 |
| pabelanger | if we make that optional and move into executor, that would fix that | 17:06 |
| corvus | 17:05 < openstackgerrit> James E. Blair proposed openstack-infra/project-config master: Add publish-zuul-docs job https://review.openstack.org/555377 | 17:06 |
| corvus | ^ that is a step to get zuul docs published to zuul-ci.org/docs/zuul | 17:07 |
| clarkb | pabelanger: I'm somewhat inclined to say you should just run a private launcher for this reason too | 17:07 |
| clarkb | (I don't actually knwo if any other platforms support that though) | 17:07 |
| pabelanger | corvus: clarkb: right, I think that is completely reasonable to say. But, does mean more overhead on operator to support. | 17:08 |
| clarkb | pabelanger: ya but so is using complicated private resources | 17:08 |
| corvus | i believe rcarrillocruz has a patch in progress for executor affinity (we'll land it after 3.0) | 17:08 |
| clarkb | its complicated not because of nodepool but because of your operating environment | 17:08 |
| corvus | maybe we could make it a switch? | 17:09 |
| mrhillsman | launcher does not use much of anything so if one has executor inside private network how much more would it require to co-locate launcher | 17:09 |
| clarkb | corvus: ya secure (or in openstack case best effort) by default then opt in if people don't care? | 17:10 |
| pabelanger | keyscan a switch? | 17:10 |
| corvus | clarkb, pabelanger: yes | 17:10 |
| clarkb | if your executor is in the private env already it is likely that your chances for a mitm attack are lowered | 17:10 |
| clarkb | mrhillsman: I think its more about needing a one off launcher just for that cloud. If you are talking to many clouds that is annoying | 17:10 |
| clarkb | mrhillsman: if it is your only cloud it makes no difference | 17:10 |
| pabelanger | yah, I'd be okay with that | 17:10 |
| corvus | right, ^ this is a case where the right answer may be topology dependent, so a switch may not be unreasonable. | 17:11 |
| mrhillsman | true, makes sense | 17:11 |
| mrhillsman | but if you are using windmill :) | 17:11 |
| pabelanger | cool, I'll hack on it after 3.0 release :D | 17:12 |
| mrhillsman | just a little joke there | 17:12 |
| mrhillsman | i am still learning things but would be cool to have a switch | 17:12 |
| pabelanger | well, this is actually for testing with RDO cloud, they have limited public IPs ATM | 17:12 |
| clarkb | pabelanger: another important thing to keep in mind is that public vs private is really not the correct terminology | 17:13 |
| clarkb | pabelanger: the IPs need to be reachable from the executor and the launcher. In a corporate env or a single cloud that may mean everyone is "private" but its "public" to the serivces | 17:14 |
| mrhillsman | yeah i think i am getting confused as well by that and my lack of knowing how things work | 17:14 |
| clarkb | pabelanger: so if your launcher was running in the rdo cloud it could talk to rdo cloud resources and other public clouds just fine | 17:14 |
| clarkb | pabelanger: without changing anything (potentially depending on network topology, terms and conditions apply) | 17:14 |
| mrhillsman | no total lack of course | 17:14 |
| mrhillsman | s/no/not | 17:14 |
| pabelanger | clarkb: yah | 17:15 |
| mrhillsman | we do that right now but i do not think it scales as mentioned | 17:15 |
| clarkb | mrhillsman: it would scale as long as the other clouds were also accessible from there | 17:15 |
| clarkb | mrhillsman: possibly because they are true public clouds and the launcher can talk to them via NAT out of its private env or bceause its another cloud on the same private corp networking env | 17:15 |
| mrhillsman | we have limited public IPs so we do not attach public IPs to the launched VMs | 17:15 |
| pabelanger | clarkb: you summed it up right with: the cloud api is public, the resources are not | 17:16 |
| clarkb | mrhillsman: right but they don't have to be true public clouds, they just have to be "public" or rather reachable from the services | 17:16 |
| mrhillsman | yeah, all the services are running within that cloud | 17:16 |
| mrhillsman | and all the other clouds are public | 17:16 |
| mrhillsman | this conversation probably sheds some light on why our initial cloud behind VPN stuff was not working | 17:17 |
| mrhillsman | this was actually quite timely for me and gives me a little bit more to chew on with re to what we are doing | 17:19 |
| clarkb | for example if I was still at HP we had ssh access out of the corporate network. So I could run all my control plane in the "private" corp env against our internal cloud using privately IP'd instances. Then also talk to a true public cloud using public IPs from my control plane in private land | 17:19 |
| clarkb | (this particular use case helped inform a bit of the new zuulv3 and nodepool design aiui) | 17:19 |
| mrhillsman | was your control plane in the corp env running in the internal cloud? | 17:20 |
| clarkb | it was running on the precursor to the internal cloud, a vmware cluster | 17:21 |
| mrhillsman | ok, i am running everything in the internal cloud | 17:22 |
| mrhillsman | i hope things work but at least i have a bit more understanding now on this front and can probably ask better questions going forward | 17:22 |
| mrhillsman | it was a lot easier to get setup again and fix issues versus some months ago for sure | 17:23 |
| pabelanger | clarkb: I'm actually checking now to see if current setup of nodepool-launcher is on same private network as VMs today. Don't have direct access myself | 17:24 |
| clarkb | pabelanger: it doesn't have to be on the same network either, just be able to route to it | 17:24 |
| pabelanger | right | 17:25 |
| clarkb | zuulv2 required that instance be able to reach back to the control plane which is trickier in corporate envs if consuming public compute resources (or any other less trusted compute resource) | 17:26 |
| clarkb | which is why zuulv3 is so push heavy (git repos are pushed, it sshes to instances, etc) | 17:27 |
| *** rlandy|biab is now known as rlandy | 17:40 | |
| *** tosky has quit IRC | 17:44 | |
| *** tosky has joined #zuul | 17:48 | |
| Shrews | our zuul setup guides don't mention anything about gerrit setup. should it? any special steps one needs to take for that? | 17:53 |
| clarkb | Shrews: you need a gerrit user with the stream ssh events right in gerrit (it is an acl now) then you configure zuul to use that user and ssh key | 17:55 |
| clarkb | its probably something that should be included at some point | 17:55 |
| corvus | yeah, i'd like to branch the zuul-from-scratch guide so we have a github/gerrit path. choose your own adventure | 17:56 |
| *** flepied_ has quit IRC | 17:57 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Publish zuul docs to zuul-ci.org https://review.openstack.org/555395 | 17:58 |
| corvus | if folks could approve that, i can verify/fix the new docs publishing job, and then we can add it to the other repos | 17:59 |
| *** jpena is now known as jpena|off | 18:00 | |
| mordred | corvus: ++ | 18:02 |
| Shrews | corvus: ++ I'm attempting to setup things on my local machine and kind of hit that roadblock | 18:03 |
| corvus | Shrews: cool, if you end up going the gerrit route and record your process, we could incorporate it into the docs | 18:06 |
| Shrews | corvus: sure... assuming i can figure out the process :) | 18:06 |
| *** harlowja has joined #zuul | 18:15 | |
| *** harlowja_ has joined #zuul | 18:17 | |
| *** harlowja has quit IRC | 18:19 | |
| *** flepied_ has joined #zuul | 18:35 | |
| openstackgerrit | Merged openstack-infra/zuul master: Publish zuul docs to zuul-ci.org https://review.openstack.org/555395 | 18:41 |
| *** JasonCL has quit IRC | 18:44 | |
| *** JasonCL has joined #zuul | 18:50 | |
| *** JasonCL has quit IRC | 18:51 | |
| *** JasonCL has joined #zuul | 19:01 | |
| *** JasonCL has quit IRC | 19:05 | |
| Shrews | corvus: standard gerrit setup doesn't seem to match our normal workflow (no +/-W buttons present). Do we want to recommend a particular configuration? | 19:21 |
| Shrews | oh, maybe that's b/c i'm the admin user | 19:22 |
| clarkb | oh right they only use code review by default now | 19:22 |
| Shrews | yeah, i'm running with the latest | 19:22 |
| Shrews | 2.14.6 iirc | 19:23 |
| Shrews | 2.14.7 | 19:23 |
| Shrews | i guess we aren't version restricted, right? | 19:23 |
| tobiash | Shrews: yes, default is without workflow and verified only between -1 and 1 | 19:24 |
| tobiash | but I guess if the target is gating then at least verified needs to be -2 to 2 | 19:24 |
| tobiash | workflow might be optional | 19:24 |
| Shrews | i need to figure out the proper config settings then :/ | 19:25 |
| *** trishnag has joined #zuul | 19:25 | |
| clarkb | its pretty straightforward we probably have examples in the spec repo configs | 19:26 |
| trishnag | Hi, is there a way that I can get an offline version of zuul doc https://docs.openstack.org/infra/zuul? | 19:27 |
| Shrews | clarkb: yeah, trying to parse http://git.openstack.org/cgit/openstack-infra/puppet-gerrit/tree/templates/gerrit.config.erb atm | 19:27 |
| Shrews | but if they've recently changed the default, we probably aren't using the config value yet | 19:28 |
| tobiash | Shrews: we have something like this in the project config: http://paste.openstack.org/show/708994/ | 19:29 |
| clarkb | Shrews: we override the default iirc | 19:30 |
| clarkb | Shrews: so that we have the extra stuff | 19:30 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add missing localhost delegation checks to some modules https://review.openstack.org/555435 | 19:30 |
| corvus | Shrews: we actually don't go into gating for github either yet; it might be worth just sticking with 'check' for starters, then adding 'gate' later | 19:31 |
| *** robled has quit IRC | 19:31 | |
| tobiash | corvus, mordred, clarkb: ^ | 19:31 |
| Shrews | corvus: good point | 19:32 |
| tobiash | I think without gating the default settings (except permissions) should be ok | 19:32 |
| clarkb | tobiash: corvus I realize the commit message about the whitelist is a little stale now | 19:35 |
| clarkb | tobiash: corvus but I don't know that that is important enough to change at this point? | 19:36 |
| tobiash | ups | 19:36 |
| tobiash | overlooked that | 19:36 |
| tobiash | just a sec | 19:36 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add missing localhost delegation checks to some modules https://review.openstack.org/555435 | 19:37 |
| clarkb | I guess its good for breadcrumbs in the future to get that right | 19:37 |
| corvus | trishnag: you can locally generate an html version (just run 'tox -e docs'). or it should be possible to build a pdf, though i haven't tried it and am not sure how well it will turn out. | 19:37 |
| *** patriciadomin has quit IRC | 19:38 | |
| clarkb | tobiash: noticed somethign else linting may care about but if it doesn't then lets just merge it as is and clean it up later | 19:39 |
| tobiash | clarkb: I thought I ran pep8 locally | 19:40 |
| clarkb | tobiash: its possible top level unused vars like that are treated differently because they are global | 19:40 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add missing localhost delegation checks to some modules https://review.openstack.org/555435 | 19:42 |
| clarkb | sorry, its just things like that are easier to notice in gerrit :) | 19:44 |
| tobiash | no problem, it's easy to fix :) | 19:44 |
| *** rlandy is now known as rlandy|brb | 19:54 | |
| *** robled has joined #zuul | 20:07 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add missing localhost delegation checks to some modules https://review.openstack.org/555435 | 20:20 |
| tobiash | \o/ | 20:20 |
| tobiash | that should be the last of the current sec fixes | 20:21 |
| tobiash | corvus, clarkb, fungi: https://etherpad.openstack.org/p/IMSdAnUVuN | 20:21 |
| pabelanger | tobiash: awesome work increasing our testing around ansible! Great to see. | 20:23 |
| tobiash | thanks :) | 20:24 |
| clarkb | tobiash: lgtm | 20:26 |
| pabelanger | wfm, added trivial update | 20:27 |
| trishnag | corvus: thanks | 20:27 |
| *** rlandy|brb is now known as rlandy | 20:30 | |
| *** flepied_ has quit IRC | 20:33 | |
| *** flepied has joined #zuul | 20:33 | |
| fungi | tobiash: advisory looks great, thanks! | 20:39 |
| *** JasonCL has joined #zuul | 21:04 | |
| clarkb | http://zuul.openstack.org/stream.html?uuid=a363be3162ab4ce0b4f0df527a004a78&logfile=console.log says Failure from finger client: host and port was not specified and no sock specified should i be worriedabout that post executor restart? | 21:09 |
| *** flepied has quit IRC | 21:10 | |
| clarkb | hrm now it shows up | 21:12 |
| tobiash | clarkb: I also see this in my deployment | 21:13 |
| clarkb | tobiash: some kind of delay in starting the socket maybe? | 21:14 |
| tobiash | It looks like the streamink link is shown too early | 21:14 |
| pabelanger | clarkb: yah, think there is a race | 21:14 |
| pabelanger | I think if merger on executor is delayed, it happens | 21:14 |
| pabelanger | maybe we should only update webui after local repos are setup | 21:15 |
| tobiash | At some point the executor reports a url back to the scheduler an I think there was a change a while ago to report that later | 21:17 |
| tobiash | But it still might be too early | 21:18 |
| tobiash | pabelanger, clarkb: that updates the link: http://git.openstack.org/cgit/openstack-infra/zuul/tree/zuul/executor/server.py#n757 | 21:24 |
| clarkb | and that happens before run playbooks is called | 21:25 |
| tobiash | Ya | 21:25 |
| clarkb | http://git.openstack.org/cgit/openstack-infra/zuul/tree/zuul/executor/server.py#n880 is maybe a better place to call it? | 21:25 |
| tobiash | Probably | 21:27 |
| *** JasonCL has quit IRC | 21:33 | |
| *** JasonCL has joined #zuul | 21:34 | |
| *** JasonCL has quit IRC | 21:36 | |
| *** JasonCL has joined #zuul | 21:43 | |
| *** JasonCL has quit IRC | 21:47 | |
| pabelanger | ++ | 21:58 |
| *** JasonCL has joined #zuul | 22:01 | |
| *** rlandy is now known as rlandy|bbl | 22:04 | |
| *** JasonCL has quit IRC | 22:08 | |
| *** JasonCL has joined #zuul | 22:41 | |
| *** JasonCL has quit IRC | 22:42 | |
| corvus | https://zuul-ci.org/docs/zuul/ | 22:46 |
| corvus | wow that worked the first time :) | 22:46 |
| Diabelko | hi, I'm having an issue where Zuul suddenly started saying that change is not ready to be merged and ignoring the review | 22:50 |
| Diabelko | parent of the commit is merged, there's everything in place | 22:50 |
| Diabelko | I can give links to gerrit/code/logs, anything really | 22:51 |
| Diabelko | but I have no idea what the issue might be | 22:51 |
| Diabelko | parent is merged, this is a revert commit | 22:51 |
| corvus | Diabelko: a link to the change would help. and do you have zuul debug-level logs? | 22:53 |
| Diabelko | yes, I do | 22:54 |
| Diabelko | corvus: https://review.opencontrail.org/#/c/40958/ this is a review that refused to run, https://review.opencontrail.org/#/c/40570/2 this one otoh, for the same repo, runs just fine | 22:55 |
| Diabelko | http://zuulv3.opencontrail.org | 22:55 |
| Diabelko | contrail-project-config is our repo with zuul configuration and ansible code | 22:56 |
| Diabelko | https://review.opencontrail.org/#/c/40970/ and this is duplicated 40958, because I wanted to see what's happening | 22:57 |
| corvus | Diabelko: can you paste the logs starting at "Considering adding change .* 40958" | 22:57 |
| corvus | example: 2018-03-22 06:28:18,247 DEBUG zuul.Pipeline.openstack.gate: Considering adding change <Change 0x7f79792ecef0 555152,1> | 22:58 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Test growroot in boot tests https://review.openstack.org/555103 | 23:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool master: Publish docs to zuul-ci.org https://review.openstack.org/555497 | 23:06 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-sphinx master: Publish docs to zuul-ci.org https://review.openstack.org/555501 | 23:09 |
| Diabelko | mhm, sure | 23:09 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-sphinx master: Publish docs to zuul-ci.org https://review.openstack.org/555501 | 23:11 |
| Diabelko | corvus: http://paste.openstack.org/show/709119/ | 23:11 |
| corvus | Diabelko: ok it added it to the queue. but it must have later removed it without reporting. | 23:14 |
| corvus | Diabelko: what other log lines are there for 40958,2 after that? | 23:14 |
| Diabelko | corvus: I don't see anything apart from that, let me run recheck again for that | 23:21 |
| Diabelko | corvus: http://paste.openstack.org/show/709140/ saw something new | 23:23 |
| Diabelko | "No jobs for change <Change 0x7f0704a27898 40958,3>" | 23:23 |
| clarkb | I'm assuming that is a trusted config project? which would mean that noop jobs addition wouldn't apply until it merges | 23:24 |
| clarkb | that leaves you with the lint jobs that we would expect to run right? | 23:24 |
| Diabelko | yes | 23:24 |
| Diabelko | to both yes | 23:25 |
| Diabelko | I can open another review with different content, but with the same commit id as parent | 23:25 |
| Diabelko | and we can see if it happens | 23:25 |
| corvus | Diabelko: zuul-jobs-linters has a files restriction: https://github.com/Juniper/contrail-zuul-jobs/blob/master/zuul.d/infra-jobs.yaml#L86 | 23:26 |
| corvus | Diabelko: and since it doesn't match, there are in fact no jobs to run on that change | 23:26 |
| Diabelko | https://review.opencontrail.org/#/c/40946/ | 23:29 |
| Diabelko | while I can see what you mean, I have no idea why it worked before then | 23:30 |
| clarkb | could be a newer restriction? | 23:33 |
| corvus | yeah, was added 3 days ago: https://github.com/Juniper/contrail-zuul-jobs/commit/0cdbfdd4ac067ea037056aed892567d8c920215d | 23:33 |
| *** JasonCL has joined #zuul | 23:33 | |
| corvus | er, actually, it looks like that commit merged today | 23:34 |
| corvus | Diabelko: https://review.opencontrail.org/#/c/40820/ | 23:34 |
| *** JasonCL has quit IRC | 23:34 | |
| Diabelko | ah! | 23:34 |
| Diabelko | I was sure it was added 3 days ago | 23:35 |
| Diabelko | thank you for that guys... MVPs! | 23:35 |
| Diabelko | :) | 23:35 |
| corvus | Diabelko: no prob! | 23:35 |
| Diabelko | and I am so sorry for bothering you with stuff like that :( | 23:35 |
| clarkb | I think there is a bug about logging this case right? | 23:36 |
| Diabelko | kklimonda: ^ FYI it's your fault :> | 23:38 |
| Diabelko | thank you guys again | 23:38 |
| *** tosky has quit IRC | 23:43 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Add a documentation index page https://review.openstack.org/555505 | 23:51 |
| *** xinliang has quit IRC | 23:58 | |
| *** JasonCL has joined #zuul | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!