-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939201: wip: Don't invalidate zk context on stop event https://review.opendev.org/c/zuul/zuul/+/939201 | 07:10 | |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939208: Avoid lost events when aborting event processing https://review.opendev.org/c/zuul/zuul/+/939208 | 09:21 | |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939201: wip: Don't invalidate zk context on stop event https://review.opendev.org/c/zuul/zuul/+/939201 | 10:20 | |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939201: Don't invalidate zk context on stop event https://review.opendev.org/c/zuul/zuul/+/939201 | 10:20 | |
@dfajfer:fsfe.org | kinda out of blue question but how to debug queues? We had per-branch queues enabled but they seem not to work anymore after updating Zuul to 7.0 | 11:24 |
---|---|---|
@dfajfer:fsfe.org | ``` | 11:24 |
- queue: | ||
name: foo | ||
per-branch: true | ||
``` | ||
-@gerrit:opendev.org- Vladimir Kozhukalov proposed: [zuul/zuul-jobs] 939234: [remove-registry-tag] Improve usage experience https://review.opendev.org/c/zuul/zuul-jobs/+/939234 | 13:09 | |
@prime149:matrix.org | Hi team. Can someone please review this https://review.opendev.org/c/zuul/zuul-jobs/+/939234 ? This is to make it possible to use the role promote-container-image within a loop with multiple registries when one would like to publish images both to dockerhub and quay. Now it does not work due to one of variables that is set in the first iteration and then never redefined. | 13:51 |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939238: wip: Reproduce issue with match-on-config-updates disabled https://review.opendev.org/c/zuul/zuul/+/939238 | 14:00 | |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939238: wip: Reproduce issue with match-on-config-updates disabled https://review.opendev.org/c/zuul/zuul/+/939238 | 14:03 | |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939238: wip: Reproduce issue with job forcing a dependency to run https://review.opendev.org/c/zuul/zuul/+/939238 | 14:42 | |
@joao15130:matrix.org | Hello. | 14:47 |
I'm setting up CI system using Zuul and I'm currently stuck in the step where an ansible called prepare-workspace is trying to do an rsync to one nodepool VM and this is what I got from the zuul_executor. | ||
This is what I got from the logs | ||
2025-01-14 11:32:57,652 DEBUG zuul.AnsibleJob.output: [e: 004a9cfec4e0447dbb11779492ce273a] [build: b07822d90b3e4dd99318f0f6de4f8232] Ansible output: b'TASK [prepare-workspace : Synchronize src repos to workspace directory. delete=True, dest={{ zuul_workspace_root }}, recursive=True, src={{ zuul.executor.src_root }}, owner=False, group=False] ***' | ||
2025-01-14 11:32:58,413 DEBUG zuul.AnsibleJob.output: [e: 004a9cfec4e0447dbb11779492ce273a] [build: b07822d90b3e4dd99318f0f6de4f8232] Ansible result output: b'RESULT failure' | ||
2025-01-14 11:32:58,424 DEBUG zuul.AnsibleJob.output: [e: 004a9cfec4e0447dbb11779492ce273a] [build: b07822d90b3e4dd99318f0f6de4f8232] Ansible output: b'fatal: [ubuntu-jammy]: FAILED! => {"changed": false, "cmd": "/usr/bin/rsync --delay-updates -F --compress --delete-after --archive --no-owner --no-group --rsh=\'/usr/bin/ssh -S none -o Port=22 -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null\' --out-format=\'<<CHANGED>>%i %n%L\' /var/lib/zuul/builds/b07822d90b3e4dd99318f0f6de4f8232/work/src zuul@10.228.237.212:.", "msg": "Warning: Permanently added \'10.228.237.212\' (ED25519) to the list of known hosts.\\r\\nPermission denied, please try again.\\r\\nPermission denied, please try again.\\r\\nzuul@10.228.237.212: Permission denied (publickey,password).\\r\\nrsync: connection unexpectedly closed (0 bytes received so far) [sender]\\nrsync error: unexplained error (code 255) at io.c(231) [sender=3.2.7]\\n", "rc": 255}' | ||
@joao15130:matrix.org | I have already checked that SSH password less is configured correctly, and when I run rsync command from the executor manually, it works fine | 14:48 |
@joao15130:matrix.org | any help appreciated | 14:48 |
-@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 939238: wip: Reproduce issue with job forcing a dependency to run https://review.opendev.org/c/zuul/zuul/+/939238 | 14:52 | |
@fungicide:matrix.org | joao15130: did your manual test use the same ssh key? and you explicitly tried ssh'ing into the zuul account on the remote node? | 14:55 |
@joao15130:matrix.org | yes same key. | 14:59 |
This is what happens when I do it manually: | ||
root@6687c91eb78e:/# /usr/bin/rsync --delay-updates -F --compress --delete-after --archive --no-owner --no-group --rsh=ssh -avz /tmp zuul@10.228.237.212:/home/zuulbuilding file list ... donesent 118 bytes received 15 bytes 88.67 bytes/sectotal size is 0 speedup is 0.00 | ||
Or | ||
zuul@6687c91eb78e:~$ /usr/bin/rsync --delay-updates -F --compress --delete-after --archive --no-owner --no-group --rsh=ssh -avz /tmp zuul@10.228.237.212:/home/zuulbuilding file list ... donesent 118 bytes received 15 bytes 88.67 bytes/sectotal size is 0 speedup is 0.00 | ||
@joao15130:matrix.org | from the executor to the host | 14:59 |
@fungicide:matrix.org | is your executor process running from a container, and if so is your manual rsync being tested from withing the container context? | 15:01 |
@joao15130:matrix.org | yes and yes | 15:01 |
@joao15130:matrix.org | I'm following https://zuul-ci.org/docs/zuul/latest/tutorials/quick-start.html to spin up my env | 15:02 |
@dfajfer:fsfe.org | * ``` | 15:19 |
- queue: | ||
name: foo | ||
per-branch: true | ||
``` | ||
What worked for me was: | ||
``` | ||
- project: | ||
name: foo | ||
queue: foo | ||
default-branch: bar | ||
``` | ||
Question is why I had to do that change or why it worked previously | ||
@fungicide:matrix.org | dfajfer: you say it was when upgrading to 7.0.0? did you see the upgrade notes? https://zuul-ci.org/docs/zuul/latest/releasenotes.html#relnotes-7-0-0-upgrade-notes | 15:24 |
@fungicide:matrix.org | "The deprecated syntax of specifying project change queues on pipeline configurations has been removed. Specify queues using the project stanza now." | 15:24 |
@dfajfer:fsfe.org | yeah but I had just a standalone queue | 15:36 |
@dfajfer:fsfe.org | I just had to attach it via project | 15:37 |
@dfajfer:fsfe.org | I'd expect that if I had configuration like | 15:38 |
``` | ||
- project: | ||
name: foo | ||
gate: | ||
queue: foo | ||
``` | ||
@dfajfer:fsfe.org | I only had the queue.name, queue.per-branch config and it did the job before that | 15:38 |
@clarkb:matrix.org | Jean Pierre Roquesalane: I would try to simplify your debugging by checking on normal ssh connections through ansible before debugging rsync. That might help provide more clues? But in general zuul should add your ssh key to an ssh agent exposed to the ansible running in bwrap. Maybe check to see if that appears to be happening properly? | 15:44 |
@fungicide:matrix.org | ah, good point, the agent socket might be missing or unreachable by the process | 15:46 |
@joao15130:matrix.org | all other ansible actions are running fine, this is just happening for rsync or synchronize module | 16:28 |
@clarkb:matrix.org | I think rsync isn't able to use the control persistence setup that normal ssh uses. If this only affects rsync then I would guess something is preventing new ssh connections after the initial one is configured with control persistence | 17:05 |
-@gerrit:opendev.org- Dong Zhang proposed: [zuul/zuul] 936999: Fix auth redirect problem when root url is accessed https://review.opendev.org/c/zuul/zuul/+/936999 | 17:52 | |
@prime149:matrix.org | I restored this https://review.opendev.org/c/zuul/zuul-jobs/+/939234 because another PR https://review.opendev.org/c/zuul/zuul-jobs/+/881524 that seems to fix the issue looks not actual (updated 1 year ago). | 19:38 |
@clarkb:matrix.org | note you can update existing changes | 19:41 |
@clarkb:matrix.org | but also I think the original change stalled out because it wasn't backward compatible? | 19:41 |
-@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: | 19:42 | |
- [zuul/zuul] 931704: Fix handling large command output with split streams https://review.opendev.org/c/zuul/zuul/+/931704 | ||
- [zuul/zuul] 931706: Fix units for ansible max line length https://review.opendev.org/c/zuul/zuul/+/931706 | ||
@clarkb:matrix.org | looks like your change is mostly backward compatible and gives precedence to the inferred type rather than the explicit type. That is unlikely to be an issue unless someon sets an /etc/hosts entry for docker that poinst to quay or uses a proxy for that or similar | 19:46 |
@jim:acmegating.com | the other change has 2x +2s and just needs someone to make an announcement if necessary. | 19:49 |
@prime149:matrix.org | Yep, it must be backward compatible. Can you please approve it if you don't mind? | 19:51 |
@jim:acmegating.com | it kind of looks like everyone wanted the other change a year ago... what changed? | 19:52 |
@clarkb:matrix.org | corvus: I'm guessing it fizzled out on the need for backward compatibiliy or announcements and maybe other distractions> I think prime149's appraoch should work and then we can folluwup with making it explicit per image later? | 19:54 |
@jim:acmegating.com | why not just send the announcement and then approve 524 in 2 weeks? | 19:54 |
@clarkb:matrix.org | I guess that works too | 19:55 |
@clarkb:matrix.org | it does delay things by two weeks but probably not the end of the world if you just run two separate jobs for now ro something | 19:56 |
@jim:acmegating.com | doing so resolves the issue and makes the roles better and more consistent. | 19:57 |
@jim:acmegating.com | Clark: prime149 either of you are welcome to send a quick note to zuul-announce; that will start the 2 week clock and we can approve it then. leave a note on the change mentioning when the announcement was sent. | 19:58 |
@clarkb:matrix.org | I'll defer to prime149 since they seem more motivated to get this done than I was apparently | 19:59 |
@jim:acmegating.com | prime149: you can check the list archive for example messages | 19:59 |
@prime149:matrix.org | Yep, will do this. Thanks for your help. | 19:59 |
@prime149:matrix.org | I sent the announcement to the mailing list but I wasn't a member of this list. So waiting for approval from a mailing list moderator. The mail subject is ` [zuul-jobs] Update remove-registry-tag role to use explicit registry type` | 21:06 |
@jim:acmegating.com | prime149: thanks -- would you please make one change and re-send it? the message doesn't actually say that `remove_registry_tag_api_type` will be removed, and i think that's an important thing to highlight so people can tell if they are affected. maybe something like this: https://etherpad.opendev.org/p/aupQoOxJm0mB_TEdequN | 21:32 |
@jim:acmegating.com | once you do that, i'll discard the first message and approve the updated one | 21:33 |
@jim:acmegating.com | (everything else looks great, thanks) | 21:33 |
@prime149:matrix.org | Sure, give me a minute | 21:34 |
@prime149:matrix.org | sent the updated version. Also mentioned that we do the same in the ensure-quay-repo role. | 21:43 |
@jim:acmegating.com | prime149: approved, thanks! https://lists.zuul-ci.org/archives/list/zuul-announce@lists.zuul-ci.org/thread/FO5URKLKB6UY54IPCV6D6BPG4AAHOQJO/ | 21:45 |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 939291: Force hard transitive dependencies to run https://review.opendev.org/c/zuul/zuul/+/939291 | 22:51 | |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Simon Westphahl: [zuul/zuul] 939238: Fix issue with job forcing a dependency to run https://review.opendev.org/c/zuul/zuul/+/939238 | 22:51 | |
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 938685: Use opendevmirror for container images https://review.opendev.org/c/zuul/zuul/+/938685 | 22:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!