*** bhagyashris_ is now known as bhagyashris|ruck | 05:25 | |
opendevreview | Ian Wienand proposed zuul/zuul master: Pin sphinx < 4.1.0 https://review.opendev.org/c/zuul/zuul/+/800590 | 05:47 |
---|---|---|
opendevreview | Ian Wienand proposed zuul/zuul master: Pin sphinx < 4.1.0 https://review.opendev.org/c/zuul/zuul/+/800590 | 06:27 |
opendevreview | Dong Zhang proposed zuul/zuul master: test tox-docs build https://review.opendev.org/c/zuul/zuul/+/800595 | 06:39 |
opendevreview | Felix Edel proposed zuul/zuul master: Make reporting asynchronous https://review.opendev.org/c/zuul/zuul/+/691253 | 07:39 |
*** rpittau|afk is now known as rpittau | 07:39 | |
opendevreview | Dong Zhang proposed zuul/zuul master: test tox-docs build https://review.opendev.org/c/zuul/zuul/+/800595 | 08:57 |
opendevreview | Dong Zhang proposed zuul/zuul master: test tox-docs build https://review.opendev.org/c/zuul/zuul/+/800595 | 08:59 |
opendevreview | Dong Zhang proposed zuul/zuul master: test tox-docs build https://review.opendev.org/c/zuul/zuul/+/800595 | 09:24 |
*** dmellado_ is now known as dmellado | 09:25 | |
opendevreview | Simon Westphahl proposed zuul/zuul master: Fix wrong if condition in result event handler https://review.opendev.org/c/zuul/zuul/+/800610 | 09:32 |
swest | zuul-maint: ^ minor issue that slipped in with the merge result events in ZK | 09:34 |
* AndyLadjadj[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/NOOIwZottIlLKtfnUeXvVIwJ/message.txt > | 10:09 | |
* AndyLadjadj[m] < https://matrix.org/_matrix/media/r0/download/matrix.org/tpxsPEPfzBSLXKVmBmXMqCOV/message.txt > | 10:10 | |
tristanC | for a given project, is there a way to remove a job enabled by a template? e.g. somthing equivalent to setting voting to false and run to do-not-run.yaml | 11:14 |
opendevreview | Dong Zhang proposed zuul/zuul master: test tox-docs build https://review.opendev.org/c/zuul/zuul/+/800595 | 12:30 |
opendevreview | Ronelle Landy proposed zuul/zuul-jobs master: DNM - Testing rpm update from compose - iptables https://review.opendev.org/c/zuul/zuul-jobs/+/800629 | 12:39 |
opendevreview | Dong Zhang proposed zuul/zuul master: Display branch of queue in status page https://review.opendev.org/c/zuul/zuul/+/777613 | 12:46 |
*** ricolin_ is now known as ricolin | 12:48 | |
corvus | tristanC: you can't stop it from executing, but you can do what you propose in the project stanza and it should override the job definition. or -- you might try setting an impossible files matcher in the project stanza? | 13:25 |
corvus | tristanC: (or you could consider a different template hierarchy) | 13:26 |
tristanC | corvus: i see thanks, the context is https://pagure.io/fedora-zuul-jobs-config/pull-request/121 where it might be tricky to change the template hierarchy. The impossible files matcher should work, thanks | 13:32 |
opendevreview | Felix Edel proposed zuul/zuul master: Switch to ZooKeeper backed NodesProvisionedEvents https://review.opendev.org/c/zuul/zuul/+/799833 | 13:53 |
opendevreview | Simon Westphahl proposed zuul/zuul master: Create config cache ltime before requesting files https://review.opendev.org/c/zuul/zuul/+/800659 | 14:02 |
opendevreview | Simon Westphahl proposed zuul/zuul master: Ensure config cache stages are used correctly https://review.opendev.org/c/zuul/zuul/+/800660 | 14:02 |
opendevreview | Andy Ladjadj proposed zuul/zuul master: [DNM][merger] use git directly to retrieve tags instead of lazy method https://review.opendev.org/c/zuul/zuul/+/800664 | 14:13 |
clarkb | https://review.opendev.org/c/zuul/zuul/+/800590 should fix the issue that hit https://review.opendev.org/c/zuul/zuul/+/800610 ? | 15:01 |
clarkb | corvus: tristanC tobiash[m] ^ fyi may want to review that first one then recheck/reapprove the second? | 15:02 |
corvus | clarkb: done, thx | 15:04 |
opendevreview | Ronelle Landy proposed zuul/zuul-jobs master: DNM - Testing rpm update from compose - iptables https://review.opendev.org/c/zuul/zuul-jobs/+/800629 | 15:16 |
*** mgoddard- is now known as mgoddard | 15:27 | |
tosky | hi, something weird happend on a patch of mine: a change in branchless tempest is declared as dependency in a openstack stable/ussuri patch, but zuul doesn't show the depends-on patch | 16:01 |
tosky | this is the tempest patch: https://review.opendev.org/c/openstack/tempest/+/800614 | 16:01 |
*** marios is now known as marios|out | 16:02 | |
tosky | and this is the testing patch which doesn't show the tempest patch as dependency in the zuul view: https://review.opendev.org/c/openstack/cinder/+/800618 | 16:02 |
tosky | what did I do wrong? | 16:02 |
opendevreview | Merged zuul/zuul master: Pin sphinx < 4.1.0 https://review.opendev.org/c/zuul/zuul/+/800590 | 16:11 |
clarkb | tosky: if I had to guess there is some override forcing tempest master to be used | 16:13 |
tosky | uhm, I'm going to mess up with the terminology: tempest from the master branch should be used, but the variant which applies to ussuri should be used | 16:15 |
opendevreview | Andy Ladjadj proposed zuul/zuul master: [DNM][merger] use git directly to retrieve tags instead of lazy method https://review.opendev.org/c/zuul/zuul/+/800664 | 16:15 |
clarkb | tosky: yes I understand the intent there (and your intent is definitely a reasonable thing to do). I'm wondering if some config in the jobs says "always use tempest master" and that overrides your depends on somehow | 16:17 |
clarkb | looking at the jobs I'm not finding anything like that though | 16:21 |
tosky | I guess there is no other way to debug this apart from studying the source code of the jobs, right? | 16:22 |
clarkb | https://review.opendev.org/c/zuul/zuul/+/800595 also seemed to struggle with depends -n | 16:22 |
clarkb | tosky: and maybe checking zuul service logs | 16:22 |
clarkb | corvus: ^ is it possible some of the sos work has impacted depends on? | 16:23 |
*** rpittau is now known as rpittau|afk | 16:23 | |
opendevreview | Andy Ladjadj proposed zuul/zuul master: [DNM][merger] use git directly to retrieve tags instead of lazy method https://review.opendev.org/c/zuul/zuul/+/800664 | 16:26 |
opendevreview | Clark Boylan proposed zuul/zuul master: DNM testing depends on with change that should fail pep8 https://review.opendev.org/c/zuul/zuul/+/800683 | 16:29 |
opendevreview | Clark Boylan proposed zuul/zuul master: DNM testing depends on this one should fail https://review.opendev.org/c/zuul/zuul/+/800684 | 16:30 |
clarkb | tosky: ^ sanity check | 16:30 |
tosky | thanks | 16:31 |
tosky | I guess I should hope something else outside my patch is broken, but that's not a nice thing to hope for | 16:32 |
clarkb | corvus: ya 800684 passed testing when it should've failed. Also the status page doesn't show the dependency in check | 16:38 |
clarkb | I'm not in a great spot to debug further right this moment. I need to go do morning things like breakfast. But I can try and dig in more after that | 16:40 |
clarkb | not sure where to start though | 16:41 |
corvus | clarkb, tosky: i can triage that today. | 16:46 |
gmann | clarkb: corvus tosky I see depends-on not working in this too https://review.opendev.org/c/openstack/tempest/+/788346 | 16:47 |
opendevreview | Andy Ladjadj proposed zuul/zuul master: [DNM][merger] use git directly to retrieve tags instead of lazy method https://review.opendev.org/c/zuul/zuul/+/800664 | 16:48 |
tosky | thanks! | 16:51 |
opendevreview | Merged zuul/zuul master: Fix wrong if condition in result event handler https://review.opendev.org/c/zuul/zuul/+/800610 | 16:52 |
corvus | clarkb: i'm pretty sure it's the gerrit hostname change | 17:05 |
clarkb | oh fun | 17:05 |
clarkb | corvus: even though we set the canonical name? seems like that would be something zuul should handle? | 17:05 |
corvus | still digging | 17:06 |
corvus | clarkb: well, the canonical hostname is 'opendev.org', so that would be more wrong | 17:08 |
clarkb | hrm | 17:09 |
corvus | clarkb: i don't think there's a bug here, just a missing feature (we're expecting zuul to know that review.opendev.org is the same server as review01.opendev.org -- we have no mechanism for telling it that). | 17:09 |
clarkb | just thinking out loud here it seems reasonable that you might connect to one of many gerrits in a cluster and when you do that you'd need to identify the lgoical gerrit instance for things like depends on | 17:10 |
clarkb | corvus: yup that | 17:10 |
corvus | clarkb: yes, that's theoretically reasonable. however, it's also reasonable to suppose that any hypothetical person in that situation might also think it's reasonable for zuul and humans to both connect to the same such server. | 17:11 |
clarkb | (that said I don't actually know enough about gerrit clustering to say how you would most appropriately instruct zuul to talk to gerrt in that situation) | 17:11 |
clarkb | the reason for the config change is wanting to have control over what gerrit zuul talks to in order to coordinate dns updates without that impacting zuul directly | 17:12 |
corvus | clarkb: anyway, i see 2 options here: revert the hostname change, or add a new feature to zuul. | 17:12 |
corvus | oh, third option: start telling people to use review01.opendev.org | 17:13 |
clarkb | currently review01 is review. review02 will be review. To apply the change converting review02 to review we would ideally have zuul merge that code and the deploy jobs. But we should be able to manage that in another way (force merge then manually run playbooks?) | 17:13 |
clarkb | given ^ maybe reverting is the easiest thing for now? | 17:14 |
clarkb | I'm not sure how easy adding a feature to zuul would be, both will require zuul to be restarted. | 17:15 |
corvus | clarkb: tbh, i'm not current on the plan, so it's hard for me to say... also, maybe we should move to #opendev | 17:15 |
opendevreview | Andy Ladjadj proposed zuul/zuul master: [DNM][merger] use git directly to retrieve tags instead of lazy method https://review.opendev.org/c/zuul/zuul/+/800664 | 17:22 |
-opendevstatus- NOTICE: Depends-On using https://review.opendev.org URLs are currently not working. This was due to a config change in Zuul that we are reverting and will be restarting Zuul to pick up. | 17:43 | |
*** zbr is now known as Guest681 | 17:50 | |
JPEW | When using the k8s driver, is it possible to request resources other than cpu and memory (specifically, devices.kubevirt.io/kvm) ? | 19:50 |
corvus | JPEW: i believe those are the only currently supported, but i don't see why more couldn't be added | 19:56 |
clarkb | if you use the namespace type you should be able to provision any resources that you have access to from the job? that may be a reasonable workaround | 19:59 |
clarkb | corvus: does your comment on https://review.opendev.org/c/zuul/zuul/+/799833/4/zuul/scheduler.py imply that change needs an update to modify the js too? | 21:17 |
corvus | clarkb: oh yes, i thought it included that, sorry | 21:20 |
corvus | updated accordingly | 21:21 |
clarkb | cool I wanted to make sure before I dove in and accidentally approved somethign that might not be working | 21:22 |
clarkb | I'll still review it, but won't approve it | 21:22 |
corvus | i'll see if i can squeeze an update of that in in a bit | 21:25 |
opendevreview | James E. Blair proposed zuul/zuul master: DNM: Make reporting asynchronous https://review.opendev.org/c/zuul/zuul/+/800728 | 21:34 |
*** dviroel is now known as dviroel|out | 21:41 | |
corvus | tobiash: ^ that's a throw-away change that is a merge-conflict resolution for the async reporter and reporting intermediate buildsets/builds https://review.opendev.org/797182 -- the conflict resolution is straightforward, and the tests pass. i don't anticipate a significant problem. i'm happy to wait a bit longer if you want to do some more evaluation, but i think it at least passes a first-level evaluation. | 21:51 |
corvus | tobiash: i also tested it against the end of the stack, and that's fine too. no other conflicts so i didn't bother to do a second rebase | 21:53 |
opendevreview | James E. Blair proposed zuul/zuul master: Always report the build page https://review.opendev.org/c/zuul/zuul/+/800112 | 21:54 |
corvus | clarkb: i believe based on javascript's handling of missing fields, that result event change would not error, but we should still remove it. | 21:56 |
clarkb | corvus: noted | 21:56 |
opendevreview | James E. Blair proposed zuul/zuul master: Switch to ZooKeeper backed NodesProvisionedEvents https://review.opendev.org/c/zuul/zuul/+/799833 | 21:57 |
corvus | clarkb: ^ it's a pretty trivial update, so i went ahead and pushed it. if you have a review in progress i don't think it should disrupt you. | 21:58 |
clarkb | I haven't dug in to the point where I was leaving comments yet so this is fine | 21:58 |
clarkb | I've been finding that for bigger changes like that doing a high level review to understand all the large moving pieces then going back again and doing in depth read really helps me. | 21:58 |
corvus | it'll be worth one of us double checking the status page looks okay in the preview. | 21:58 |
corvus | clarkb: ++ | 21:59 |
clarkb | corvus: felixedel can you check my comments on https://review.opendev.org/c/zuul/zuul/+/799833 I feel like I'm missing something fundamental there | 23:10 |
gmann | clarkb: corvus depends-on is working fine now, thanks for fixing https://review.opendev.org/c/openstack/cinder/+/800618 | 23:25 |
clarkb | gmann: ya I did some independent verification too. Good to know it is doing what you expect on that change now | 23:26 |
corvus | clarkb: replied; lemme know if that doesn't make sense | 23:27 |
clarkb | corvus: yes that helps. The name isn't reconsituted from nodepool/zk though right? | 23:29 |
clarkb | I wonder if we should give it a more distinct name in that case like "invalid-x" | 23:29 |
corvus | clarkb: no objection to that; the only conceivable use at this point would be in a log line or something, so something is probably better than nothing, and distinct is probably better than indistinct. :) | 23:32 |
clarkb | I've left a +2 and a note about ^ | 23:32 |
corvus | clarkb: sounds good; i'm not in a rush to approve that today, so let's see what felixedel says tomorro | 23:37 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!