| tristanC | zuul-jobs core, could you please review: https://review.openstack.org/#/c/632620/ . Without this, the add-build-sshkey doesn't work well for static nodes | 00:16 |
|---|---|---|
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: upgrade react-scripts to ^2.0.0 https://review.openstack.org/631902 | 00:35 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: upgrade react-scripts to ^2.0.0 https://review.openstack.org/631902 | 01:15 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: upgrade react and react-scripts to ^2.0.0 https://review.openstack.org/631902 | 01:57 |
| *** bhavikdbavishi has joined #zuul | 03:02 | |
| *** spsurya has joined #zuul | 04:28 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: switch jobs list to a tree view https://review.openstack.org/633437 | 06:04 |
| *** quiquell|off is now known as quiquell | 06:20 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: scheduler: add job's parent name to the rpc job_list method https://review.openstack.org/573473 | 06:25 |
| *** sshnaidm|off is now known as sshnaidm | 06:30 | |
| *** badboy has joined #zuul | 06:35 | |
| openstackgerrit | Simon Westphahl proposed openstack-infra/zuul master: List changed files for all commits between refs https://review.openstack.org/631797 | 07:01 |
| *** chandankumar has quit IRC | 07:07 | |
| *** panda|off is now known as panda | 07:20 | |
| *** chandan_kumar has joined #zuul | 07:21 | |
| *** avass has joined #zuul | 07:39 | |
| openstackgerrit | Rui Chen proposed openstack-infra/zuul master: Avoid using list branches with protected=1 in github driver https://review.openstack.org/630038 | 07:45 |
| openstackgerrit | Simon Westphahl proposed openstack-infra/zuul master: List changed files for all commits between refs https://review.openstack.org/631797 | 07:54 |
| *** pcaruana has joined #zuul | 08:01 | |
| *** chandan_kumar is now known as chandankumar | 08:02 | |
| *** themroc has joined #zuul | 08:08 | |
| *** gtema has joined #zuul | 08:44 | |
| *** electrofelix has joined #zuul | 08:53 | |
| *** jpena|off is now known as jpena | 08:54 | |
| *** luizbag has joined #zuul | 09:41 | |
| *** jesusaur has quit IRC | 09:45 | |
| *** strigazi has joined #zuul | 09:45 | |
| strigazi | hello, I'm trying to figure this out: http://logs.openstack.org/36/612336/12/check/container-build/a5ac779/ara-report/ Locally it works, but on the ci it fails to fetch the tarball. Any clues? | 09:47 |
| *** jesusaur has joined #zuul | 09:51 | |
| *** avass has quit IRC | 09:53 | |
| *** bhavikdbavishi has quit IRC | 09:53 | |
| *** AJaeger has quit IRC | 10:03 | |
| badboy | can anyone tell why running 'pip install .' once more in Zuul's dir removes web dashboard files? | 10:04 |
| badboy | \/usr/local/lib/python3.6/dist-packages/zuul/web/static is empty after that | 10:05 |
| *** AJaeger has joined #zuul | 10:07 | |
| jkt | badboy: is that a git checkout? | 10:33 |
| badboy | jkt: yes | 10:33 |
| jkt | badboy: dashboard builds need a ton of JS-specific tooling, including node.js and yarn. Do you have that? | 10:34 |
| jkt | badboy: installing from a prebuilt package at pypi does not require that | 10:35 |
| badboy | jkt: when I run 'pip install .' in Zuul git checkout everything works but running the same command once again removes the compiled js for dashboard | 10:36 |
| jkt | badboy: alright, then I cannot help, it's something else than I hit a week or two ago | 10:38 |
| *** SotK is now known as SotK__ | 10:39 | |
| *** SotK_ is now known as SotK | 10:39 | |
| badboy | I've created a dirty script for installing Zuul (no errorhandling whatsover): http://paste.openstack.org/show/5VTy8HrJclTXU3WJiT1p/ | 10:44 |
| *** rfolco has joined #zuul | 10:56 | |
| jkt | badboy: I was able to get it running with ansible using the openstack's ansible zuul playbooks, https://gerrit.cesnet.cz/plugins/gitiles/ci/ansible-cesnet/+/master/ | 11:03 |
| jkt | badboy: I was doing Puppet before, so I took this as an opportunity to get myself familiar with yet another tool, and my TL;DR impression is that it's mostly "okay" | 11:04 |
| badboy | jkt: thank you :) | 11:13 |
| badboy | jkt: regarding your '# TODO: this is buggy, `touch` always updates...' you can use: modification_time: "preserve" & access_time: "preserve" | 11:40 |
| badboy | jkt: https://docs.ansible.com/ansible/latest/modules/file_module.html | 11:41 |
| *** bhavikdbavishi has joined #zuul | 11:53 | |
| *** gtema has quit IRC | 11:53 | |
| *** gtema has joined #zuul | 11:53 | |
| *** avass has joined #zuul | 12:02 | |
| *** panda is now known as panda|lunch | 12:07 | |
| *** jesusaur has quit IRC | 12:07 | |
| *** jesusaur has joined #zuul | 12:15 | |
| jkt | badboy: nice, thanks | 12:26 |
| jkt | badboy: any clue why `pip install` always says that it modifies something when installing zuul? (it doesn't report anything for nodepool...) | 12:26 |
| *** jpena is now known as jpena|lunch | 12:31 | |
| *** panda|lunch is now known as panda | 12:33 | |
| *** pcaruana has quit IRC | 12:36 | |
| *** pcaruana has joined #zuul | 12:37 | |
| *** jesusaur has quit IRC | 12:46 | |
| *** jesusaur has joined #zuul | 12:51 | |
| *** rlandy has joined #zuul | 12:54 | |
| badboy | jkt: no clue, sorry | 13:10 |
| *** bhavikdbavishi has quit IRC | 13:10 | |
| *** bhavikdbavishi has joined #zuul | 13:18 | |
| sshnaidm | if I add required_jobs in job and in its child, is it overwritten or merged in child? | 13:23 |
| sshnaidm | by child I mean inherited | 13:23 |
| *** jpena|lunch is now known as jpena | 13:29 | |
| *** pcaruana has quit IRC | 13:32 | |
| *** bhavikdbavishi has quit IRC | 13:37 | |
| openstackgerrit | Benedikt Löffler proposed openstack-infra/zuul master: Report retried builds via sql reporter. https://review.openstack.org/633501 | 13:40 |
| *** pcaruana has joined #zuul | 13:50 | |
| *** badboy has quit IRC | 13:58 | |
| *** bhavikdbavishi has joined #zuul | 13:58 | |
| *** bhavikdbavishi has quit IRC | 14:07 | |
| pabelanger | jkt: you have leaked your private keys for your gerrit connection. You may want to rotate them. | 14:17 |
| *** gtema has quit IRC | 14:23 | |
| *** bhavikdbavishi has joined #zuul | 14:34 | |
| tobiash | yay, we had a scheduler wedge today because a user force-merged a change referencing a non-existing project-template and the post job reporting caused an exception loop blocking all tenants | 14:44 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Fix wedged scheduler on force-merge of non-existing template https://review.openstack.org/633514 | 14:44 |
| tobiash | this is the fix ^ | 14:44 |
| *** quiquell is now known as quiquell|off | 14:47 | |
| pabelanger | tobiash: we had a periodic job (with semaphore) wedge this weekend also. Think there was a network issue between gearman, scheduler tried to notifiy executor, but didn't reach it. Job looked to be started, but no executor servicing it | 14:51 |
| pabelanger | then post jobs were blocked because of semaphore | 14:51 |
| jkt | pabelanger: did I? Where? I would have sworn that I encrypted them | 14:51 |
| tobiash | pabelanger: maybe you want https://review.openstack.org/631721 | 14:51 |
| pabelanger | jkt: line 193 | 14:52 |
| pabelanger | tobiash: looking | 14:52 |
| tobiash | pabelanger: and https://review.openstack.org/599573 once we have a gear release | 14:52 |
| pabelanger | tobiash: https://softwarefactory-project.io/paste/show/1401/ was debug from scheduler, tristanC might have more thoughts. But we didn't actually need to restart scheduler, we could dequeue the change | 14:54 |
| pabelanger | sorry, wrong paste | 14:54 |
| pabelanger | https://softwarefactory-project.io/paste/show/1403/ | 14:54 |
| jkt | pabelanger: ah, that is badboy's script, actually | 14:55 |
| pabelanger | jkt: ah, sorry about that | 14:55 |
| tobiash | pabelanger: in your case the gear keepalive would have probably saved you | 14:58 |
| pabelanger | tobiash: cool! Agree, networking isn't the best here. Anything blocking the release? | 15:02 |
| tobiash | pabelanger: just doing afaik | 15:03 |
| pabelanger | ack, will ask on #openstack-infra now | 15:04 |
| *** chandankumar is now known as chkumar|out | 15:08 | |
| *** dkehn has joined #zuul | 15:19 | |
| *** swest has quit IRC | 15:22 | |
| openstackgerrit | Benedikt Löffler proposed openstack-infra/zuul master: Report retried builds via sql reporter. https://review.openstack.org/633501 | 15:42 |
| *** openstackgerrit has quit IRC | 15:51 | |
| avass | so, I deleted a change while it zuul started a job for it and now it's stuck | 16:02 |
| avass | and I can't dequeue it because the change doesn't exist | 16:02 |
| avass | any tips? | 16:02 |
| pabelanger | avass: how did you delete it, in gerrit? | 16:03 |
| avass | yep | 16:06 |
| avass | pabelanger: seems like it solved itself | 16:06 |
| pabelanger | avass: k, I don't think many people delete issues in gerrit, so might be use case not really tested. Good to hear now resolved | 16:07 |
| pabelanger | s/issues/changes | 16:07 |
| *** avass has quit IRC | 16:27 | |
| *** hashar has joined #zuul | 16:31 | |
| *** bhavikdbavishi has quit IRC | 16:34 | |
| *** bhavikdbavishi has joined #zuul | 16:37 | |
| *** dmsimard has quit IRC | 16:43 | |
| *** hashar has quit IRC | 16:44 | |
| *** pcaruana has quit IRC | 16:47 | |
| *** dmsimard has joined #zuul | 16:48 | |
| clarkb | is that via a gerrit plugin? | 16:49 |
| clarkb | or does newer gerrit than we run allow you to do that now? | 16:49 |
| *** dmsimard has quit IRC | 16:49 | |
| *** bhavikdbavishi has quit IRC | 16:50 | |
| *** bhavikdbavishi has joined #zuul | 16:51 | |
| *** dmsimard has joined #zuul | 16:52 | |
| corvus | tobiash: i was going to put out a last call for reviews on the multi-ansible spec, but the question raised by SpamapS of whether to run the ansible installation during pip install seems like a significant one; did you want to look into that for a little bit first? | 17:11 |
| *** bhavikdbavishi has quit IRC | 17:12 | |
| *** bhavikdbavishi has joined #zuul | 17:13 | |
| *** dkehn has quit IRC | 17:15 | |
| corvus | maybe we can try to land that next week | 17:15 |
| tobiash | corvus: I had a look and research on the weekend about this | 17:23 |
| tobiash | corvus: I managed to run the ansible installation during source based installations which works good in principle | 17:24 |
| tobiash | corvus: however wheel based installations have no real way of post install hooks (which are needed as the virtualenvs cannot be packaged inside the wheel as they are not relocatable) | 17:25 |
| tobiash | so for wheel based installations we need to call the ansible install script manually | 17:25 |
| corvus | tobiash: yeah. given that, it's probably not worth trying to do it as part of pip install | 17:25 |
| tobiash | but I'm fine with not supporting managing the installations during startup if you are | 17:25 |
| corvus | (tbh, my experience with doing things with pip/setuptools/etc, is that the less done, the better) | 17:26 |
| electrofelix | Finally getting a chance to look at nodepool (so may make some progress in upgrading to zuulv3 yay!) | 17:26 |
| corvus | tobiash: i still think that if we can have install-on-startup without too much effort, then it's worth it to make things easier for new users. | 17:27 |
| clarkb | ya wheels will pose a problem there. | 17:27 |
| electrofelix | we have a problem with our current set up around some limited resources that I'm hoping nodepool can help. I see it has static nodes, but we have a need for a static slave and a dynamic resource that jenkins won't be able to ssh into | 17:27 |
| corvus | tobiash: so i still like the idea of having both the external-script or install-on-startup [which does the same thing as external-script] options. | 17:27 |
| tobiash | corvus: I'm fine with supporting both as it's not really additional effort, but SpamapS point of view was contradicting yours so we need a decision on that ;) | 17:28 |
| tobiash | SpamapS: ^ | 17:28 |
| tobiash | corvus: I already have that external script in place and it works the same during startup | 17:28 |
| corvus | electrofelix: nodepool can mix and match static/dynamic nodes, but zuulv3 is going to expect to be able to ssh into all of them | 17:28 |
| tobiash | really supporting both is essentially a few lines of code | 17:29 |
| electrofelix | corvus: right, hence why I'm thinking that we need a custom driver that could grab a node from the static pool to go along with the limited resource and just return the static node with an additional label that we can use on the Jenkins side to know what data to inject into the job env | 17:30 |
| electrofelix | we have some resources in AWS and we need an agent attached to Jenkins to act as the launch point for connecting (note we're still Jenkins and I'm hoping this is the first step to allow moving away) | 17:31 |
| corvus | tobiash, SpamapS: yeah, I read SpamapS as mostly favoring install-time-install, which i agree would be nice, but if we accept that it's impractical, i think we're left with an external script being the main option that we're all happy with, and "run that script on startup" as being a simple secondary option. | 17:31 |
| tobiash | corvus, SpamapS: so if we can come up with an agreement on this point, I can update the spec tomorrow | 17:34 |
| electrofelix | Currently we're allowing for some fixed resources to be provided, e.g. a k8s env in AWS, but we only have a few of them, and still need to start the job from something that has access within the internal network | 17:34 |
| corvus | tobiash: might be worth adding your findings about pip in a comment | 17:34 |
| *** sdake has quit IRC | 17:34 | |
| tobiash | to summarize: install-time-install is possible during 'pip install .' (I have that working within tox), but not with wheels | 17:34 |
| tobiash | corvus: yeah, I'll add a comment about this | 17:35 |
| clarkb | electrofelix: nodepool can broker the k8s resources directly | 17:35 |
| clarkb | (though I don't know if there is quota management yet) | 17:35 |
| tobiash | for reference, there is an open issue about wheel compatible post install hooks: https://github.com/pypa/packaging-problems/issues/64 | 17:37 |
| electrofelix | clarkb: given the current env we still need a worker to be attached to jenkins that can run stuff against those k8s instances | 17:37 |
| electrofelix | but not sure how to describe those resources in nodepool so it can handle managing handing them to jenkins along with a worker to run the job without having to have a separate dedicated pool | 17:37 |
| *** sdake has joined #zuul | 17:37 | |
| electrofelix | of static machines | 17:37 |
| tobiash | corvus: but regarding the installation position I'd agree to install it by default in lib/zuul/... | 17:39 |
| electrofelix | you could say that both the k8s resources and the machines we have are static, but I'm thinking that the jenkins nodepool plugin cannot consume the k8s resources directly because it can't run an agent on it | 17:39 |
| tobiash | maybe we could change the default to /var/lib/zuul/... if we're managing ansible at runtime | 17:39 |
| electrofelix | but if I can work out a way to make this work, then the step to zuulv3 becomes much smaller internally | 17:40 |
| *** openstackgerrit has joined #zuul | 17:40 | |
| openstackgerrit | Paul Vinciguerra proposed openstack-infra/zuul master: Identify the gerrit source when logging stream-events errors. https://review.openstack.org/633567 | 17:40 |
| clarkb | electrofelix: I'm not familiar enough with how the jenkins agent + nodepool v3 stuff works to guide much there but on the zuulv3 side I think you'd simply request the k8s env then the executor would talk to that directly. Maybe you can use the jenkins master's agents as a stand in for the zuul executor there | 17:42 |
| electrofelix | clarkb: that's kind of what I'm hoping for, but I think jenkins will expect to be able to directly ssh to whatever nodepool provides, hence why I'm wondering if we need to go down a custom driver can a driver grab resources from another pool internally to nodepool? | 17:46 |
| electrofelix | maybe it might be better to put an email together and send it on | 17:48 |
| electrofelix | have to drop now, meant to start the questions earlier today | 17:48 |
| *** electrofelix has quit IRC | 17:48 | |
| *** sdake has quit IRC | 17:51 | |
| *** sdake_ has joined #zuul | 17:52 | |
| *** sdake_ has quit IRC | 17:53 | |
| *** bhavikdbavishi has quit IRC | 18:04 | |
| *** themroc has quit IRC | 18:04 | |
| *** jpena is now known as jpena|off | 18:13 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/nodepool master: Amazon EC2 driver https://review.openstack.org/535558 | 18:13 |
| *** sdake_ has joined #zuul | 18:13 | |
| *** sdake_ is now known as sdkae | 18:14 | |
| openstackgerrit | Merged openstack-infra/zuul master: Incoming connections over 19885/TCP are needed on nodes https://review.openstack.org/633242 | 18:18 |
| openstackgerrit | Paul Vinciguerra proposed openstack-infra/zuul master: Identify the gerrit source when logging stream-events errors. https://review.openstack.org/633567 | 18:37 |
| *** rfolco has quit IRC | 19:00 | |
| *** spsurya has quit IRC | 19:14 | |
| *** luizbag has quit IRC | 19:15 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Amazon EC2 driver https://review.openstack.org/535558 | 19:19 |
| tobiash | SpamapS: fixed test case ^ | 19:19 |
| *** rfolco has joined #zuul | 19:34 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/nodepool master: Amazon EC2 driver https://review.openstack.org/535558 | 20:08 |
| openstackgerrit | Merged openstack-infra/zuul master: Use gearman client keepalive https://review.openstack.org/599573 | 20:09 |
| tobiash | corvus: does gertty have the ability to preview the inline comments in the change view similar to the gerrit ui? | 20:09 |
| * tobiash is trying to use gertty more often | 20:10 | |
| corvus | tobiash: yes, if you create an inline comment, it's saved in draft form until you click 'review' on the change. more or less the same as in the gerrit ui. except in gertty, the drafts are saved locally in gertty's db; in gerrit they're saved on the server. | 20:13 |
| corvus | (so draft comments don't sync between gerrit<->gertty) | 20:14 |
| tobiash | corvus: I mean to see all the inline comments in the change overview | 20:15 |
| tobiash | In gerrit I can expand the inline comments without having to change to the diff screen | 20:17 |
| corvus | tobiash: oh, no that hasn't been added yet, but i'd love to have it | 20:18 |
| tobiash | Au, thx | 20:18 |
| tobiash | Currently I'm switching often between gerrit and gertty, mostly because of this feature ;) | 20:19 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP add provides/requires support https://review.openstack.org/633605 | 21:19 |
| openstackgerrit | Paul Vinciguerra proposed openstack-infra/zuul master: configloader.py: Not all jobs have an .updated attribute. https://review.openstack.org/633259 | 21:25 |
| *** panda is now known as panda|off | 21:35 | |
| openstackgerrit | Paul Vinciguerra proposed openstack-infra/zuul master: configloader.py: Not all jobs have an .updated attribute. https://review.openstack.org/633259 | 21:36 |
| *** panda|off has quit IRC | 21:54 | |
| *** panda has joined #zuul | 21:56 | |
| *** sdkae has quit IRC | 22:03 | |
| *** sdake has joined #zuul | 22:05 | |
| openstackgerrit | Paul Vinciguerra proposed openstack-infra/zuul master: configloader.py: Not all jobs have an .updated attribute. https://review.openstack.org/633259 | 22:15 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP add provides/requires support https://review.openstack.org/633605 | 22:34 |
| *** sdake has quit IRC | 22:45 | |
| *** sdake has joined #zuul | 22:45 | |
| *** sdake has quit IRC | 22:46 | |
| *** sdake has joined #zuul | 22:48 | |
| *** D0han has left #zuul | 22:50 | |
| *** sdake has quit IRC | 22:57 | |
| *** sdake has joined #zuul | 22:59 | |
| *** sdake has quit IRC | 23:01 | |
| *** sdake has joined #zuul | 23:03 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP add provides/requires support https://review.openstack.org/633605 | 23:35 |
| *** ianw is now known as ianw_pto | 23:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!