| *** tosky has quit IRC | 00:06 | |
| *** rfolco has quit IRC | 00:12 | |
| *** rlandy|rover|bbl is now known as rlandy|rover | 00:28 | |
| *** Goneri has quit IRC | 01:01 | |
| *** iurygregory has quit IRC | 01:32 | |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 01:46 |
|---|---|---|
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 01:47 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 02:00 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 02:05 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 02:09 |
| *** zenkuro has quit IRC | 02:31 | |
| *** bhavikdbavishi has joined #zuul | 02:40 | |
| *** bhavikdbavishi1 has joined #zuul | 02:42 | |
| *** bhavikdbavishi has quit IRC | 02:44 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:44 | |
| *** rlandy|rover has quit IRC | 02:52 | |
| *** bhavikdbavishi has quit IRC | 03:21 | |
| *** bhavikdbavishi has joined #zuul | 03:22 | |
| *** bhavikdbavishi has quit IRC | 03:29 | |
| *** bhavikdbavishi has joined #zuul | 03:31 | |
| *** vishalmanchanda has joined #zuul | 03:52 | |
| *** donnyd has quit IRC | 04:07 | |
| *** donnyd has joined #zuul | 04:08 | |
| *** ChrisShort has quit IRC | 04:16 | |
| *** ChrisShort has joined #zuul | 04:16 | |
| *** bhavikdbavishi has quit IRC | 04:19 | |
| *** bhavikdbavishi has joined #zuul | 04:19 | |
| *** bhavikdbavishi has quit IRC | 04:34 | |
| *** bhavikdbavishi has joined #zuul | 04:37 | |
| *** johnsom has quit IRC | 04:46 | |
| *** johnsom has joined #zuul | 04:49 | |
| *** bhavikdbavishi has quit IRC | 04:52 | |
| *** johnsom has quit IRC | 05:18 | |
| *** johnsom has joined #zuul | 05:19 | |
| *** aprice has quit IRC | 05:27 | |
| *** aprice has joined #zuul | 05:27 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #zuul | 05:33 | |
| *** johnsom has quit IRC | 06:27 | |
| *** johnsom has joined #zuul | 06:27 | |
| *** bhavikdbavishi has joined #zuul | 06:53 | |
| *** bhavikdbavishi1 has joined #zuul | 06:56 | |
| *** bhavikdbavishi has quit IRC | 06:58 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 06:58 | |
| *** bhavikdbavishi has quit IRC | 07:24 | |
| *** bhavikdbavishi has joined #zuul | 07:25 | |
| *** mach1na has joined #zuul | 07:25 | |
| *** CraigR has joined #zuul | 07:29 | |
| *** bhavikdbavishi has quit IRC | 07:29 | |
| *** mach1na has quit IRC | 07:33 | |
| *** jcapitao has joined #zuul | 08:03 | |
| *** rpittau|afk is now known as rpittau | 08:05 | |
| *** bhavikdbavishi has joined #zuul | 08:08 | |
| *** jpena|off is now known as jpena | 08:09 | |
| *** hashar has joined #zuul | 08:11 | |
| *** jfoufas1 has joined #zuul | 08:11 | |
| *** sugaar has quit IRC | 08:12 | |
| *** sugaar has joined #zuul | 08:12 | |
| *** guillaumec has quit IRC | 08:14 | |
| *** guillaumec has joined #zuul | 08:15 | |
| *** tobiash has quit IRC | 08:15 | |
| *** bhavikdbavishi1 has joined #zuul | 08:15 | |
| *** bhavikdbavishi has quit IRC | 08:16 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 08:16 | |
| *** tobiash has joined #zuul | 08:16 | |
| *** guillaumec has quit IRC | 08:23 | |
| *** guillaumec has joined #zuul | 08:24 | |
| *** slaweq has joined #zuul | 08:51 | |
| slaweq | hi zuul experts :) | 08:51 |
| slaweq | I'm trying to fix Neutron ovn multinode jobs | 08:52 |
| *** iurygregory has joined #zuul | 08:52 | |
| slaweq | the problem is that in those jobs we are installing ovn and ovs from source during devstack installation | 08:52 |
| slaweq | and that breaks tunnel created by multi-node-bridge role | 08:52 |
| slaweq | so I tried to do something like https://review.opendev.org/#/c/762650/ to workaround it | 08:53 |
| slaweq | but it seems for me that this jinja2 template isn't rendered correctly | 08:53 |
| slaweq | see https://b18b498c87a70760171a-55be5aa52ff9f3c19ea7817c824a7dd8.ssl.cf5.rackcdn.com/762654/5/check/neutron-ovn-tempest-slow/d4cc10a/job-output.txt | 08:53 |
| slaweq | can You help me with that maybe? | 08:53 |
| *** tosky has joined #zuul | 08:56 | |
| *** sean-k-mooney has quit IRC | 09:27 | |
| *** sean-k-mooney has joined #zuul | 09:27 | |
| *** zenkuro has joined #zuul | 09:46 | |
| *** odyssey4me has quit IRC | 09:50 | |
| *** odyssey4me has joined #zuul | 09:50 | |
| *** felixedel has quit IRC | 09:56 | |
| *** felixedel has joined #zuul | 09:57 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: WIP: Report max 50 file comments to github https://review.opendev.org/762869 | 09:57 |
| *** nils has joined #zuul | 10:01 | |
| *** bhavikdbavishi has quit IRC | 10:15 | |
| *** wuchunyang has joined #zuul | 10:18 | |
| *** CraigR has quit IRC | 10:20 | |
| *** mach1na has joined #zuul | 10:20 | |
| *** wuchunyang has quit IRC | 10:22 | |
| *** mach1na has quit IRC | 10:54 | |
| *** wuchunyang has joined #zuul | 11:05 | |
| *** bhavikdbavishi has joined #zuul | 11:13 | |
| *** migi has joined #zuul | 11:29 | |
| *** mach1na has joined #zuul | 11:33 | |
| *** mach1na has quit IRC | 11:36 | |
| *** mach1na has joined #zuul | 11:36 | |
| *** jcapitao is now known as jcapitao_lunch | 11:55 | |
| *** slaweq has quit IRC | 11:56 | |
| *** slaweq has joined #zuul | 11:57 | |
| *** mach1na has quit IRC | 12:09 | |
| *** ianychoi_ has quit IRC | 12:10 | |
| *** wuchunyang has quit IRC | 12:11 | |
| *** rfolco has joined #zuul | 12:26 | |
| *** rfolco has quit IRC | 12:28 | |
| *** jpena is now known as jpena|lunch | 12:36 | |
| *** bhavikdbavishi has quit IRC | 12:48 | |
| *** bhavikdbavishi has joined #zuul | 12:48 | |
| *** jcapitao_lunch is now known as jcapitao | 12:55 | |
| *** rlandy has joined #zuul | 12:55 | |
| *** rlandy is now known as rlandy|rover | 12:56 | |
| *** mach1na has joined #zuul | 12:57 | |
| *** mach1na has quit IRC | 12:58 | |
| *** mach1na has joined #zuul | 12:58 | |
| *** jpena|lunch is now known as jpena | 13:24 | |
| *** Goneri has joined #zuul | 13:40 | |
| tobiash | slaweq: could the delegate_to be an issue there? | 13:43 |
| slaweq | tobiash: tbh I don't know | 13:43 |
| slaweq | I'm not zuul and ansible expert | 13:43 |
| slaweq | but should delegate_to simply create this file on the delegated node? | 13:44 |
| tobiash | slaweq: I think the delegated task uses the hostvars of the original node | 13:44 |
| tobiash | I don't know if that's an issue there since I don't know the details of that role | 13:44 |
| slaweq | tobiash: my modification is "just" doing script using variables which were already used in that job, also in task which was delegate_to | 13:45 |
| slaweq | IMO there is some issue with rendering jinja template - probably I'm doing something wrong in that template :/ | 13:46 |
| tobiash | at least I don't see any obvious typo at first glance | 13:47 |
| slaweq | tobiash: I try to use this new script here https://review.opendev.org/#/c/762654/ | 13:50 |
| tobiash | slaweq: http://paste.openstack.org/show/800094/ | 13:50 |
| slaweq | tobiash: exactly | 13:50 |
| tobiash | looking at that it looks like there is an unintended line break in there | 13:50 |
| slaweq | that's my understanding too | 13:50 |
| slaweq | but there is no this line break in the template :/ | 13:50 |
| slaweq | and I don't know why it's like that | 13:51 |
| tobiash | maybe the {{ switch_ip }} variable has one | 13:51 |
| tobiash | yes, after each of that ip address there is a line break in that file | 13:51 |
| slaweq | tobiash: ha | 13:52 |
| slaweq | You are genius :) | 13:52 |
| slaweq | thx a lot | 13:52 |
| tobiash | no problem :) | 13:52 |
| openstackgerrit | Slawek Kaplonski proposed zuul/zuul-jobs master: [multi-node-bridge] Add script to configure connectivity https://review.opendev.org/762650 | 14:00 |
| slaweq | tobiash: ^^ lets see if that will help :) | 14:01 |
| *** zenkuro has quit IRC | 14:04 | |
| *** zenkuro has joined #zuul | 14:05 | |
| *** bhavikdbavishi has quit IRC | 14:08 | |
| *** bhavikdbavishi has joined #zuul | 14:56 | |
| *** lliu82 has joined #zuul | 14:58 | |
| *** CraigR has joined #zuul | 14:58 | |
| lliu82 | Hi, I find it a little hard to understand zuul.child_jobs What kind of jobs will show in the list? | 14:59 |
| *** mach1na has quit IRC | 15:01 | |
| *** lliu82 has quit IRC | 15:02 | |
| *** LLIU82 has joined #zuul | 15:04 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: replace react-ansi component with re-ansi https://review.opendev.org/762759 | 15:08 |
| *** CraigR has quit IRC | 15:15 | |
| *** mach1na has joined #zuul | 15:16 | |
| *** jfoufas1 has quit IRC | 16:14 | |
| *** mach1na has quit IRC | 16:26 | |
| *** iurygregory has quit IRC | 16:29 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: replace react-ansi component with re-ansi https://review.opendev.org/762759 | 16:37 |
| corvus | LLIU82: jobs that set 'dependencies' on the parent job | 16:45 |
| *** LLIU82 has quit IRC | 16:48 | |
| *** hamalq has quit IRC | 16:56 | |
| *** hamalq has joined #zuul | 16:57 | |
| *** jpena is now known as jpena|off | 17:01 | |
| *** jcapitao has quit IRC | 17:19 | |
| *** iurygregory has joined #zuul | 17:19 | |
| *** rpittau is now known as rpittau|afk | 17:27 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-jobs master: DNM: generate ansi log for benchmark purpose https://review.opendev.org/763054 | 17:40 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: validate-host: Options to require v4 and v6 routes https://review.opendev.org/763065 | 18:01 |
| fungi | clarkb: frickler: ^ i should probably work out how to include a test for that | 18:02 |
| clarkb | fungi: we might rely on base-test, not sure | 18:02 |
| fungi | well, i meant a test for those new vars in the battery of zuul-jobs testing | 18:03 |
| clarkb | fungi: gotcha. Also I think continuing to support the default behavior would be good. Maybe we can use a single var: require_network_connectivty then default to 'either' but handle ipv6, ipv4, and both as values? | 18:04 |
| fungi | isn't it still supporting the prior behavior? i merely inverted how the passed flag was being toggled | 18:06 |
| fungi | oh... i see | 18:06 |
| *** y2kenny has joined #zuul | 18:06 | |
| corvus | fungi: also see inline comments | 18:06 |
| fungi | i just made it a no-op when neither is required, yup | 18:06 |
| clarkb | ya | 18:06 |
| fungi | that was not intentional ;) | 18:06 |
| fungi | i started with separate flags and then thought i could condense them to simplify the logic, but nope. fixing | 18:07 |
| *** nirmoy has joined #zuul | 18:07 | |
| tristanC | zbr: corvus: i shared ansi code rendering benchmark in https://review.opendev.org/762759 | 18:10 |
| nirmoy | Hi did any one faced: | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478390 | buildpod | MODULE FAILURE: | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478603 | buildpod | Traceback (most recent call last): | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478636 | buildpod | File "<stdin>", line 102, in <module> | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478661 | buildpod | File "<stdin>", line 94, in _ansiballz_main | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478683 | buildpod | File "<stdin>", line 40, in invoke_module | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478706 | buildpod | File "/usr/lib64/python3.9/runpy.py", line 210, in run_module | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478728 | buildpod | return _run_module_code(code, init_globals, run_name, mod_spec) | 18:10 |
| nirmoy | 2020-11-17 17:41:05.478750 | buildpod | File "/usr/lib64/python3.9/runpy.py", line 97, in _run_module_code | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478771 | buildpod | _run_code(code, mod_globals, init_globals, | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478793 | buildpod | File "/usr/lib64/python3.9/runpy.py", line 87, in _run_code | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478815 | buildpod | exec(code, run_globals) | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478836 | buildpod | File "/tmp/ansible_command_payload_tajemifl/ansible_command_payload.zip/ansible/modules/command.py", line 675, in <module> | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478860 | buildpod | File "/tmp/ansible_command_payload_tajemifl/ansible_command_payload.zip/ansible/modules/command.py", line 651, in main | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478882 | buildpod | File "/tmp/ansible_command_payload_tajemifl/ansible_command_payload.zip/ansible/modules/command.py", line 498, in zuul_run_command | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478904 | buildpod | AttributeError: 'Thread' object has no attribute 'isAlive' | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478925 | buildpod | command terminated with exit code 1 | 18:11 |
| nirmoy | 2020-11-17 17:41:05.478959 | buildpod | changed: All items complete | 18:11 |
| clarkb | nirmoy: please use a paste service in the future as it is more readable and doesn't flood the channel wit hmessages. And yes people have hit that, there is a discussion on the mialing list | 18:11 |
| clarkb | let me get a link | 18:11 |
| clarkb | http://lists.zuul-ci.org/pipermail/zuul-discuss/2020-November/001389.html the discussion starts there | 18:12 |
| corvus | tristanC: thanks; my other question is whether we can use it in the logs tab | 18:12 |
| nirmoy | clarkb: sorry about that | 18:12 |
| tristanC | corvus: for sure, i designed re-ansi so that it integrates transparently in zuul ui component | 18:13 |
| corvus | tristanC: iirc, the logs tab is a little unusual in that it operates line-by-line to support line numbers and syntax highlighting of error messages | 18:13 |
| clarkb | nirmoy: if you run the head of master the issue is fixed, but the latest reelase doesn't have it yet. YOu can also have zuul run under an older python version on the remote (while still testing under 3.9) | 18:14 |
| y2kenny | clarkb: head of master for the executor? | 18:15 |
| clarkb | y2kenny: yes | 18:15 |
| tristanC | corvus: well i'd be happy to improve the logs tab too, but i think the main advantage of re-ansi is provided by bucklescript, and i'd like to benefit from the compiler optimizations | 18:15 |
| y2kenny | clarkb: (nirmoy is my colleague) | 18:16 |
| clarkb | I broght up the idea of doing a 3.9.2 release in that mailing list thread but not sure how much interest there is in that currently | 18:16 |
| clarkb | (the existing goal was to release 4.0 with the new zk required stuff) | 18:16 |
| corvus | tristanC: i don't understand your answer | 18:16 |
| clarkb | corvus: ^ not sure if you saw that thread | 18:16 |
| nirmoy | clarkb: thanks for your quick reply :) | 18:16 |
| y2kenny | clarkb: are the the dockerhub images are always tip of master | 18:16 |
| corvus | clarkb: it took about a week of work last time to do a point release; is someone willing to do that? (i don't have time) | 18:17 |
| clarkb | y2kenny: oh wait sorry I misread your question, no it is the python version that ansible runs under on the remote | 18:17 |
| corvus | clarkb: i'm happy to help push the buttons if someone else wants to do the prep. and make sure we don't end up with crazy release notes again. | 18:17 |
| clarkb | y2kenny: you can avoid that by running the latest executor code | 18:17 |
| clarkb | corvus: I don't have time this week at least | 18:17 |
| y2kenny | clarkb: ok. | 18:17 |
| clarkb | y2kenny: and yes the latest docker image should be up to date | 18:18 |
| clarkb | y2kenny: basically 3.9.1 executor sends code to the remote node that can't run under 3.9. Latest sends code that can. | 18:18 |
| tristanC | corvus: we could use re-ansi here: https://opendev.org/zuul/zuul/src/branch/master/web/src/containers/logfile/LogFile.jsx#L221 | 18:19 |
| corvus | tristanC: want to try that out and see if it scales to 10k lines? | 18:19 |
| tristanC | corvus: but i think we could also benefit from bucklescript optimization if the whole logfile container was written in reason, similarly to re-ansi | 18:19 |
| zbr | tristanC: i am not sure if a benchmark done between incomplete ansi implementations would be correct. | 18:20 |
| y2kenny | clarkb: understood. thanks | 18:20 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul-jobs master: validate-host: Options to require v4 and v6 routes https://review.opendev.org/763065 | 18:20 |
| tristanC | zbr: well react-ansi is also incomplete, as far as i understand it only adds color | 18:22 |
| zbr | tristanC: i am not saying is bad but I would first collect some real ANSI examples of stuff we naturally find on build logs. combine them and benchmark on them. also checking that it works fine. | 18:23 |
| zbr | imho, performance matters when feature level is relatively equal. it did not observe any performance issues myself with current one. | 18:24 |
| zbr | on the other hand, I have the impression that I will get more attention if I make a PR towards your re-ansi repo instead of of current react-ansi | 18:25 |
| zbr | the author does not seem to be very active | 18:25 |
| zbr | tristanC: did you cc the patterfly feature request? it would be awesome to get this into pf itself. | 18:27 |
| tristanC | zbr: i would be happy to review your changes. We are using ReasonML to write react component, which i find a better fit for web (and native) development than es6, thus i'd be also happy to do some onboarding too. | 18:32 |
| tristanC | zbr: i'm not sure why patternfly would provide such specific component | 18:33 |
| *** cloudnull is now known as kecarter | 18:36 | |
| *** kecarter is now known as cloudnull | 18:36 | |
| *** nirmoy has quit IRC | 18:39 | |
| tobiash | clarkb, corvus: what do you think about getting the 4.0 release done instead? I think we'd need mostly the first few changes of the zk preparation stack plus the mandatory sql? | 19:03 |
| tobiash | Mandatory sql got broken due to rebases but I could fix that by tomorrow probably | 19:05 |
| clarkb | I'm not opposed, though currently distracted by all the gerrit upgrade planning and next week is a big holiday | 19:06 |
| tobiash | We could target 4.0 maybe the week after the gerrit upgrade? | 19:08 |
| tobiash | I guess that work would collide with 3.9.2 as well? | 19:08 |
| clarkb | it conflicts with USA thanksgiving so mayn of us likely won't be around tohelp | 19:08 |
| clarkb | week after I expect things to be more normal for me | 19:08 |
| tobiash | The other question is can/should we extend reno such that it supports our use case of branching and releasing from older releases | 19:10 |
| clarkb | I want to say that openstack already does that, but I can't seem to keep the related issues paged in | 19:11 |
| tobiash | At least the 3.9.1 caused a lot of trouble | 19:11 |
| tobiash | The problem I think was we have one page im the docs with all releases which was not possible with simple branching and tagging | 19:12 |
| pabelanger | anyone up for a review: https://review.opendev.org/762939/ that switches to --password-stdin for container-upload-role, in an effort to remove no_log: true task | 19:13 |
| tobiash | pabelanger: won't the copy task have the pw in the json log? | 19:14 |
| pabelanger | shouldn't | 19:15 |
| pabelanger | I don't believe content is rendered | 19:15 |
| pabelanger | https://6fe79164c1899b3f89c6-6113b1a0d951f28e43a544861377cc1f.ssl.cf1.rackcdn.com/762939/5/check/zuul-jobs-test-build-container-image-release/93189f0/job-output.json | 19:16 |
| pabelanger | testpassword only shows up, because we use set_fact | 19:17 |
| *** y2kenny has quit IRC | 19:17 | |
| pabelanger | over security stanza | 19:17 |
| fungi | clarkb: the only openstack reno examples i could find was branched release notes (the releases on stable branches aren't mixed back into the master branch document) | 19:17 |
| clarkb | fungi: gotcha the issue is the shred doc which is what tobiash thoughti t was | 19:17 |
| clarkb | seems like that use case woudl be one reno should support since release notes tend to be a single large doc | 19:18 |
| corvus | i think 4.0.0 would be preferable to 3.9.2; i don't think we should target it until 1st week of december though due to holidays/upgrade | 19:19 |
| fungi | yeah, i can't say for certain the problem is solveable... right now the way it works out how to include notes in a release goes by the point at which the tag appears in the branch history, so if we merge 3.9.2 back into master and delete the temporary stable branch for it, basically all release notes merged in master before 3.9.2 got merged in would get listed as belonging to 3.9.2 | 19:20 |
| fungi | one workaround we talked about was merging a change to delete all new release notes, merging the tag in, then reverting the deletion, but that's exceptionally messy and not great for bisecting/pickaxe | 19:21 |
| clarkb | fungi: reno could create a single combined doc from scans of ever branch? | 19:21 |
| fungi | then the branch would need to be kept indefinitely | 19:22 |
| clarkb | basically aggregate all the release notes found in /tmp or wherever and build from there rather than only building off of a branch | 19:22 |
| tobiash | something like a --all-releases option might make sense | 19:22 |
| tobiash | aka current branch plus all tags | 19:22 |
| fungi | oh, mmm maybe iterating over tags yeah | 19:22 |
| fungi | and then doing a version sort on the tag names, and deduplicating release notes associating them with the first release (according to version sort order) they're contained in | 19:23 |
| tobiash | corvus: so we could target in 3-4 weeks? | 19:23 |
| fungi | that would be a pretty big modification to reno, but it might work | 19:23 |
| corvus | tobiash: sounds reasonable | 19:24 |
| tobiash | great :) | 19:24 |
| corvus | i honestly don't see us being able to do 3.9.2 any earlier | 19:24 |
| *** bhavikdbavishi has quit IRC | 19:41 | |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Fix gerrit merge commit change with zuul configuration https://review.opendev.org/762886 | 19:47 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Fix gerrit amended merge commit change with zuul configuration https://review.opendev.org/762887 | 19:47 |
| avass | I got a quick review if anyone has time: https://review.opendev.org/#/c/762767/ | 20:17 |
| avass | it adds the job node to the graph over the components, we've had a lot of users that thinks the 'executors' are the nodes the tests run on | 20:18 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul master: web: replace react-ansi component with re-ansi https://review.opendev.org/762759 | 20:18 |
| avass | I think it would be a simple way to make that a bit clearer | 20:18 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Fix gerrit merge commit change with zuul configuration https://review.opendev.org/762886 | 20:22 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Fix gerrit amended merge commit change with zuul configuration https://review.opendev.org/762887 | 20:22 |
| avass | pabelanger: re 762939 | 20:25 |
| avass | pabelanger: oh you already got that question and answered it :) | 20:26 |
| avass | tobiash, pabelanger: content shows up as 'content: null' so I'm unsure if that's intentional or not. it seems like it's an unecessary risk to not use no_log there | 20:28 |
| pabelanger | yah, using copy / content will be safe, and it is a pattern we already use in roles | 20:28 |
| pabelanger | sign-artifact for example | 20:28 |
| pabelanger | add-sshkey | 20:29 |
| avass | there's a no log for it so I guess it's fine: https://github.com/ansible/ansible/blob/devel/lib/ansible/modules/copy.py#L508 | 20:32 |
| pabelanger | right | 20:33 |
| fungi | corvus: are you still -1 on 761621 if combined with 762588? | 20:45 |
| corvus | fungi: i think so; i think the disappearing search icon is just plain not the way to do a hyperlink. | 20:47 |
| fungi | yeah, i'm not terribly keen on the icon business either. clickable text is preferable, so long as making it clickable doesn't also cause it to no longer be highlightable so the text can be copied to the clipboard | 20:51 |
| openstackgerrit | Merged zuul/zuul-jobs master: Use --password-stdin for upload-container-image https://review.opendev.org/762939 | 20:53 |
| fungi | not sure if this is a case of fighting the framework, and whether designers assume ~nobody copies page content from their browsers these days | 20:53 |
| pabelanger | corvus: do you think we can tag zuul-registry for pypi release? | 20:54 |
| corvus | fungi: no, i don't think so. i think the original state of having the result be the link is perfectly feasible. i believe this was an attempted improvement and it hasn't worked out. | 20:54 |
| pabelanger | corvus: actually, we'd need pypi jobs it looks like | 20:55 |
| pabelanger | I can work on that now | 20:55 |
| corvus | i believe it's okay to try things like this out, but when they don't work we should be willing to roll back. so i'd like to see us just go back to the result being a link. but if that was hard to discover, i have no objection to making the job also be a link. | 20:57 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-registry master: Begin publishing to pypi https://review.opendev.org/763086 | 20:59 |
| *** hashar has quit IRC | 21:06 | |
| *** hashar_ has joined #zuul | 21:06 | |
| *** hashar_ is now known as hashar | 21:20 | |
| pabelanger | is anyone aware of build-python-release job being broken? https://zuul.opendev.org/t/zuul/build/225b52058c8e454ea4152f9008a0b5a3 | 21:24 |
| pabelanger | or a change to setuptools missing on nodes? | 21:24 |
| pabelanger | before I start down the path of fixing | 21:25 |
| pabelanger | actually, let me move that into #opendev | 21:25 |
| clarkb | I think fungi just updated those jobs? | 21:25 |
| clarkb | that could be fallout | 21:25 |
| clarkb | ? | 21:25 |
| fungi | pabelanger: we merged https://review.opendev.org/762699 earlier today, could it be fallout from that? we stopped running ensure-tox and bindep roles | 21:27 |
| fungi | we do still include ensure-pip, so i would expect we'd still have setuptools | 21:28 |
| clarkb | fungi: I just checked and I think these jobs are in zuul-jobs and not affected by openstack | 21:28 |
| clarkb | I think I know what it is | 21:28 |
| clarkb | python2 is being used and not python3? | 21:28 |
| *** nils has quit IRC | 21:29 | |
| clarkb | pabelanger: ^ check zuul and nodepool release jobs they may set python3 explicitly and that may be the fix you need | 21:29 |
| pabelanger | oh, you might be right | 21:29 |
| fungi | yeah, so probably unrelated as i only altered pti-python-tarball and python-branch-tarball playbooks in openstack/project-config | 21:29 |
| pabelanger | yah, that looks to be it | 21:29 |
| openstackgerrit | Paul Belanger proposed zuul/zuul-registry master: Begin publishing to pypi https://review.opendev.org/763086 | 21:30 |
| pabelanger | clarkb: thanks for the pointer | 21:30 |
| *** holser has quit IRC | 21:52 | |
| pabelanger | clarkb: corvus: https://review.opendev.org/763086 is passing now (release jobs) | 21:57 |
| pabelanger | for zuul-registry | 21:57 |
| *** y2kenny has joined #zuul | 22:03 | |
| y2kenny | has anyone run into "Executing local code is prohibited" when using prepare-workspace-openshift recently? | 22:03 |
| y2kenny | http://paste.openstack.org/show/800125/ | 22:04 |
| clarkb | y2kenny: the raeson for that is going to be that the role does command or shell on localhost and zuul restricts that to trusted playbooks only | 22:04 |
| clarkb | y2kenny: the likely fix in your case is to move that role into a bse job in a trusted repo or similar | 22:05 |
| *** y2kenny58 has joined #zuul | 22:05 | |
| *** y2kenny58 has quit IRC | 22:05 | |
| *** y2kenny1 has joined #zuul | 22:06 | |
| *** y2kenny1 has quit IRC | 22:06 | |
| *** y2kenny64 has joined #zuul | 22:07 | |
| *** y2kenny has quit IRC | 22:08 | |
| *** y2kenny64 has quit IRC | 22:09 | |
| *** y2kenny has joined #zuul | 22:09 | |
| y2kenny | it has been working for me without problem. The only thing I just changed is upgrading the executor (which I didn't think it's an issue.) | 22:10 |
| clarkb | y2kenny: depending on which version you upgraded from that may be it. There was a security hole plugged around this | 22:10 |
| y2kenny | clarkb: ok... I am guessing prepare-workspace-openshift needs to be run by trusted project all the time? (like it can't be included by untrusted project?) | 22:12 |
| fungi | yes, older versions of the executor didn't realize they were supposed to forbid that | 22:12 |
| fungi | it can be in a parent job inherited by jobs in an untrusted project | 22:12 |
| fungi | but the playbook calling prepare-workspace-openshift should be in a job defined in a trusted repo | 22:13 |
| y2kenny | um... Ok... it needs to be inherited at the job level... ok | 22:13 |
| clarkb | the reason for that is the trusted repo's code will only be used in a merged state not a pre merge state | 22:13 |
| clarkb | having this check in place ensures that people have checked any code changes and merged them before they are used when they can be used in sensitive areas | 22:14 |
| y2kenny | right... I probably wasn't quite clear about the security model at the time and thought I understood the model since things work | 22:14 |
| y2kenny | this actually answered a question I have for awhile... when to use job inheritance and when to just include_role | 22:15 |
| y2kenny | thanks folks. | 22:15 |
| y2kenny | clarkb: although this bring up another issue I ran into awhile back (I am not sure if it's fixed now.) Is it still true that I cannot modify the inventory across play? | 22:16 |
| clarkb | y2kenny: you cannot across playbooks, but each playbook can load in extra inventory stuff iirc | 22:18 |
| fungi | y2kenny: announcement was here: http://lists.zuul-ci.org/pipermail/zuul-announce/2020-July/000078.html | 22:18 |
| fungi | for the executor security fix | 22:19 |
| y2kenny | fungi: thanks | 22:19 |
| y2kenny | clarkb: um... is there a recommended way to pass information across playbook? | 22:20 |
| clarkb | I think zuul return maybe available between playbooks. corvus is that true? | 22:21 |
| y2kenny | actually... may be this still work... I just need to be more creative I guess. What I have is a role that launches a pod in a received k8s namespace resources and add the pod into the inventory. | 22:21 |
| y2kenny | and then run prepare-workspace-openshift | 22:21 |
| clarkb | tristanC and tobiash may have ideas too as they use k8s a bit more than I do | 22:22 |
| y2kenny | this role was included in the playbook such that the pod in the inventory is available within the same playbook. But now that the security hole is fixed, the prepare-workspace-openshift no longer works. | 22:23 |
| y2kenny | Ideally I should have the pod launching in a pre- playbook in some base job but then I need to figure out how to pass the pod info across the playbook | 22:24 |
| clarkb | I think another option is you can write into the build dir then read that info back in later | 22:24 |
| y2kenny | I think tristanC or someone else may have a patch that allow inventory modification but I don't think that was merged. | 22:25 |
| clarkb | so make pod, prepare-workspace, write details to the build dir, in next playbook read those details and add host to inventory | 22:25 |
| *** holser has joined #zuul | 22:25 | |
| y2kenny | clarkb: ah... I can give that a try | 22:25 |
| *** hashar has quit IRC | 22:26 | |
| tristanC | clarkb: iiuc, zuul return passes variables accross jobs, and using the special set_facts `cacheable: true` may makes a fact variable available for the next playbook | 22:30 |
| tristanC | y2kenny: the change to enable a playbook to update the inventory for the other playbook is https://review.opendev.org/590092 , and that can be emulated by dumping host info the executor work_dir, and then reloading in the next playbook | 22:31 |
| clarkb | oh right you can set facts and cache them | 22:32 |
| clarkb | that is probably the easiest option here | 22:32 |
| y2kenny | awesome. I will play around with this. Thanks clarkb, tristanC | 22:33 |
| tristanC | y2kenny: and this is how software-factory setup container native job, using the emulation trick here: https://softwarefactory-project.io/cgit/software-factory/sf-config/tree/ansible/roles/sf-repos/files/config/playbooks/openshift/deploy-project.yaml#n63 | 22:33 |
| tristanC | y2kenny: in the run phase, you can load the pods.yaml file and use add_host | 22:33 |
| y2kenny | nice. | 22:34 |
| tristanC | i think that's what `cacheable: true` does, but i find the explicit dump less magic | 22:35 |
| tristanC | y2kenny: here is how the load works from the integration job: https://softwarefactory-project.io/cgit/software-factory/sf-ci/tree/roles/health-check/openshift/tasks/demo_project_zuul_configuration.yml#n40 | 22:37 |
| y2kenny | ok, got it. | 22:38 |
| corvus | there are a number of jobs in zuul-jobs that use cacheable:true to pass facts from one playbook to the next | 23:07 |
| corvus | y2kenny: it's a good pattern for passing information | 23:07 |
| *** armstrongs has joined #zuul | 23:12 | |
| *** slaweq has quit IRC | 23:21 | |
| *** armstrongs has quit IRC | 23:22 | |
| pabelanger | I'm seeing missing dependency when running nodepool config-validate | 23:26 |
| pabelanger | msrestazure | 23:26 |
| pabelanger | http://paste.openstack.org/show/800137/ | 23:27 |
| pabelanger | looks like azure driver? | 23:27 |
| pabelanger | ah | 23:31 |
| pabelanger | https://review.opendev.org/752724/ | 23:31 |
| pabelanger | clarkb: corvus: do you think we can have a new nodepool release to pull in^ | 23:32 |
| *** zbr has quit IRC | 23:57 | |
| *** zbr4 has joined #zuul | 23:57 | |
| *** tosky has quit IRC | 23:58 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!