-@gerrit:opendev.org- Clark Boylan proposed on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 938685: Use opendevmirror for container images https://review.opendev.org/c/zuul/zuul/+/938685 | 00:51 | |
@clarkb:matrix.org | corvus: ^ that was a small typo and rather than post a review and wait for a new ps I just fixed it | 00:52 |
---|---|---|
@clarkb:matrix.org | the error message that you get is interesting too | 00:52 |
@clarkb:matrix.org | `2 [internal] load metadata for qay.io/opendevmirror/node:16-bookworm | 00:52 |
#2 ERROR: encountered unknown type text/html; children may not be fetched | ||
` | ||
@clarkb:matrix.org | unittests also failed but I assume that is an independent problem | 00:53 |
@mordred:waterwanders.com | Clark: did something edit the error message to remove a u from quay.io ? | 00:57 |
@jim:acmegating.com | no that was the error :) | 00:57 |
@mordred:waterwanders.com | ah. hahahaha | 00:57 |
@mordred:waterwanders.com | well - it was easy to spot :) | 00:57 |
@jim:acmegating.com | not for me! :) | 00:58 |
@jim:acmegating.com | Clark: thanks approved | 00:58 |
@joao15130:matrix.org | ok, any alternative to rsync available ? | 08:00 |
@fungicide:matrix.org | joao15130: as a backend for ansible's sync module, do you mean? (more context needed) | 14:06 |
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 937895: Add "autohold_if_failure" var in job inventory https://review.opendev.org/c/zuul/zuul/+/937895 | 16:18 | |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 16:35 | |
- [zuul/zuul] 937946: Add image/upload delete lifecycle https://review.opendev.org/c/zuul/zuul/+/937946 | ||
- [zuul/zuul] 937947: Add web API image delete endpoints https://review.opendev.org/c/zuul/zuul/+/937947 | ||
- [zuul/zuul] 938022: Allow deleting images through web UI https://review.opendev.org/c/zuul/zuul/+/938022 | ||
- [zuul/zuul] 938023: Add REST API method to trigger image build https://review.opendev.org/c/zuul/zuul/+/938023 | ||
- [zuul/zuul] 938024: Add a web UI button to build an image https://review.opendev.org/c/zuul/zuul/+/938024 | ||
- [zuul/zuul] 938087: Add labels and flavors to web https://review.opendev.org/c/zuul/zuul/+/938087 | ||
- [zuul/zuul] 938088: Add niz nodes to rest api nodes list endpoint https://review.opendev.org/c/zuul/zuul/+/938088 | ||
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 17:06 | |
- [zuul/zuul] 937210: Add an option to skip workspace checkouts https://review.opendev.org/c/zuul/zuul/+/937210 | ||
- [zuul/zuul] 937211: Further reduce workspace checkouts https://review.opendev.org/c/zuul/zuul/+/937211 | ||
- [zuul/zuul] 937212: Use empty sparse checkouts in mergers https://review.opendev.org/c/zuul/zuul/+/937212 | ||
- [zuul/zuul] 937571: Parallelize workspace repo restoration https://review.opendev.org/c/zuul/zuul/+/937571 | ||
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 938463: sqlreporter: ensure build end data is stored even if log URL is too long https://review.opendev.org/c/zuul/zuul/+/938463 | 17:10 | |
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 938128: autohold REST API: add ref filter validation https://review.opendev.org/c/zuul/zuul/+/938128 | 17:11 | |
-@gerrit:opendev.org- Matthieu Huin https://matrix.to/#/@mhuin:matrix.org proposed: [zuul/zuul] 938463: sqlreporter: ensure build end data is stored even if log URL is too long https://review.opendev.org/c/zuul/zuul/+/938463 | 19:19 | |
@joao15130:matrix.org | fungi: If the module doesn't work yes. | 20:22 |
Other question I'm configuring two zuul jobs, devstack-minimal and devstack-base by following this example:https://opendev.org/openstack/devstack/src/commit/96de42981360c874dd36e2e447b63ecdde046732/.zuul.yaml#L438 | ||
But it looks like all parameters from the parent job are not inherited and I end up with a local.conf file with just the parameters from devstack-minimal job. Any idea? | ||
@clarkb:matrix.org | the zuul variable dict includes an _inheritance_path variable showing you the inheritance of the job. I wuold start there to make sure it reflects what you expect | 20:30 |
@clarkb:matrix.org | we log that here: https://zuul.opendev.org/t/openstack/build/98f117b2a1c4491eab5a893e769aff5d/log/zuul-info/inventory.yaml#33-36 in our jobs using the log-inventory role | 20:31 |
@joao15130:matrix.org | ok how to display this zuul variable? | 20:31 |
@joao15130:matrix.org | ok | 20:31 |
@clarkb:matrix.org | youcan also use the ansible debug module to print the contents out and records that in the console stream/log | 20:35 |
@joao15130:matrix.org | I don't have anything to display in the Logs tab under Zuul UI | 20:40 |
@joao15130:matrix.org | can you show how I can display it using ansible? | 20:42 |
@joao15130:matrix.org | debug with var = "{{ _inheritance_path }}" ? | 20:42 |
@clarkb:matrix.org | I think you would do: | 20:44 |
``` | ||
debug: | ||
var: zuul._inheritance_path | ||
``` | ||
@clarkb:matrix.org | you could also just do `var: zuul` and get the entire contents of that dictionary | 20:45 |
@joao15130:matrix.org | ok let me try | 20:49 |
@joao15130:matrix.org | any reason why I don't get anything in log in the UI? | 20:49 |
@clarkb:matrix.org | that requires recording the job-output.json log file here: https://zuul.opendev.org/t/openstack/build/98f117b2a1c4491eab5a893e769aff5d/logs | 20:50 |
@clarkb:matrix.org | which the upload-logs role should do for you | 20:51 |
@clarkb:matrix.org | the web ui fetches that json manifest to determine what and how to render the logs in the ui iirc | 20:51 |
@joao15130:matrix.org | ok will look after | 20:51 |
@joao15130:matrix.org | "_inheritance_path": [' | 20:51 |
"<Job dell-base branches: None source: zuul-config/zuul.d/jobs.yaml@master#1>",' | ||
"<Job devstack-base branches: None source: zuul-config/zuul.d/jobs.yaml@master#12>",' | ||
"<Job devstack-minimal branches: None source: zuul-config/zuul.d/jobs.yaml@master#115>",' | ||
"<Job devstack-minimal branches: None source: zuul-config/zuul.d/projects.yaml@master#14>"' | ||
@joao15130:matrix.org | this is what I have in the inheritance variable | 20:51 |
@clarkb:matrix.org | the bottom of that list is the childmost job and the top is the parent most. It shows you repos, branches and line numbers so now you can work backwards and see if things are tied together wronly | 20:52 |
@clarkb:matrix.org | * the bottom of that list is the childmost job and the top is the parent most. It shows you repos, branches and line numbers so now you can work backwards and see if things are tied together incorrectly | 20:52 |
@joao15130:matrix.org | it looks ok from an inheritance standpoint | 20:53 |
@joao15130:matrix.org | Why do I have two devstack-minimal? one in jobs and one in projects? | 20:54 |
@clarkb:matrix.org | you've likely created a job variant in your pipeline definition. | 20:54 |
@clarkb:matrix.org | by setting an attribute like nodeset or branch or voting etc on the job in the pipeline definition | 20:54 |
@clarkb:matrix.org | that implicitly inherits from the base definition of the job | 20:55 |
@joao15130:matrix.org | in projects.yaml: | 20:55 |
- project: | ||
name: opendev/ci-sandbox | ||
check: | ||
jobs: | ||
- devstack-minimal | ||
@clarkb:matrix.org | maybe it always creates an implicit child when definied in a pipeline. Looks like the example I linked to above does similar | 20:56 |
@joao15130:matrix.org | and this job is defined in jobs.yaml | 20:56 |
@joao15130:matrix.org | I don't understand why it's not inherited from devstack-base | 20:57 |
@clarkb:matrix.org | it is according to your list above | 20:57 |
@joao15130:matrix.org | yes | 20:58 |
@clarkb:matrix.org | devstack-minimal pipeline -> devstack-minimal job -> devstack-base job -> dell-base job | 20:58 |
@joao15130:matrix.org | dell-base -> devstack-base -> evstack-minimal | 20:58 |
@clarkb:matrix.org | is your zuul-config repo a trusted repo in your zuul configuration? if so then changes to it won't be executed speculatively | 20:58 |
@joao15130:matrix.org | from top to bottom | 20:58 |
@clarkb:matrix.org | is it possible that the zuul-config repo is trusted and you're trying to test an unmerged update to the devstack-minimal job in that repo? | 20:59 |
@joao15130:matrix.org | source: | 21:00 |
gerrit: | ||
config-projects: | ||
- zuul-config | ||
@joao15130:matrix.org | openstack.gerrit: | 21:00 |
untrusted-projects: | ||
- opendev/ci-sandbox | ||
@clarkb:matrix.org | ok yes config project is trusted (I probably used the wrong term) | 21:00 |
@clarkb:matrix.org | that means changes to zuul-config jobs won't be applied until you have merged them essentially (to avoid speculative execution being abused to expose sensitive info) | 21:00 |
@joao15130:matrix.org | that's correct | 21:01 |
@joao15130:matrix.org | bottom line is I don't see why it's not inheriting from devstack-base | 21:02 |
@joao15130:matrix.org | I'm dropping for today, it's getting late in FR | 21:02 |
@joao15130:matrix.org | In case you have any idea/tip/recommendation, be my guest | 21:02 |
@clarkb:matrix.org | did you merge your changes? | 21:02 |
@joao15130:matrix.org | no change is pending | 21:02 |
@clarkb:matrix.org | that is mostly likely why beacuse the repo is a config project those changes don't apply to the job runs until you merge them | 21:03 |
@joao15130:matrix.org | and I see the inheritance which confirmed the configuration change was merged | 21:03 |
@clarkb:matrix.org | ok if all changes have merged then I don't have any good ideas at the moment. I would need to see the configuration probably | 21:04 |
@joao15130:matrix.org | do you have access? | 21:06 |
https://github.com/dell-esg/cinder-ci-zuul | ||
@joao15130:matrix.org | https://github.com/dell-esg/cinder-ci-zuul/blob/main/zuul.d/jobs.yaml | 21:07 |
@clarkb:matrix.org | I can see that | 21:07 |
@joao15130:matrix.org | ok cool | 21:08 |
@joao15130:matrix.org | hoping that you'll find something, I'm stuck | 21:08 |
@joao15130:matrix.org | unless I want to repeat all the variables on each job | 21:08 |
@joao15130:matrix.org | which I don't obvioulsy | 21:08 |
@clarkb:matrix.org | why do you think the minimal job variables are not being used? | 21:08 |
@joao15130:matrix.org | because the file which contains the variables in devstack only has the one that goes along with devstack-minimal | 21:09 |
@clarkb:matrix.org | ok it is possible that you are hitting a devstack bug give me a minute to dig it up | 21:10 |
@clarkb:matrix.org | Jean Pierre Roquesalane: https://bugs.launchpad.net/devstack/+bug/2082617 may be the problem specifically check my comment in there | 21:11 |
@joao15130:matrix.org | sure, I'm dropping for today | 21:11 |
@clarkb:matrix.org | if you have invalid variables in the devstack job definition then it defaults to the default essentially | 21:11 |
@clarkb:matrix.org | looking at your jobs external_bridge_mtu may not be set? | 21:12 |
@clarkb:matrix.org | or maybe you don't have a controller node defined in your nodeset so the host var lookup for controller fails | 21:15 |
@clarkb:matrix.org | but thats my next best hunch. That zuul inheritance is fine and you're hitting the bug above | 21:15 |
@mnaser:matrix.org | I assume there hasn't been much progress on nodepool letting/allowing you to fulfill node requests across providers.. | 22:14 |
@mnaser:matrix.org | I have a job that needs both an arm64 and amd64 builder but they are on different providers and the node request is getting declined. | 22:15 |
@jim:acmegating.com | that is not possible with nodepool but may be possible in the future with zuul. for now, best approach is two jobs in a graph. | 22:29 |
@mnaser:matrix.org | hmm, okay, i'll have to think about how i approach this, i've been succesfully using buildx + another node as a worker since i have keys and all because they're in the same nodeset, but just need more speed :) | 22:38 |
@mnaser:matrix.org | Has anyone ran into an issue where the zuul-merger fails to merge a change due to a conflict, but is otherwise not reproducible locally nor Gerrit sees it as a merge conflict? | 22:52 |
@mnaser:matrix.org | Getting logs from debug really does show Git indeed blowing up at a merge failure | 22:53 |
@clarkb:matrix.org | mnaser: Gerrit uses jgit which isn't quite the same as C git and you can get inconsistencies like that. Usually it is the other way around and jgit can't merge something C git can. Also note that the Gerrit merge conflict determination relies on background checks with timeouts and is best effort aiui. It may simply not have been able to determine one way or another | 23:25 |
@mnaser:matrix.org | Clark: It's odd because if i open a branch and cherry-pick both into it, they apply cleanly :\ | 23:42 |
@clarkb:matrix.org | Oh also depending on your version of Gerrit you need to explicitly enable merge conflict checking | 23:46 |
@clarkb:matrix.org | Sometimes just resolving it locally even if it looks like a noop is all you need. Git does occasionally get smarter with merge algorithms too so newer git can sometimes do better | 23:47 |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 939406: Fix whitelabel authz https://review.opendev.org/c/zuul/zuul/+/939406 | 23:58 | |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Dong Zhang: [zuul/zuul] 936999: Fix auth redirect problem when root url is accessed https://review.opendev.org/c/zuul/zuul/+/936999 | 23:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!