| fungi | as an aside, i notice by default hyperkitty is using gravatar for user icons | 00:01 |
|---|---|---|
| clarkb | looks like the colibri fix may have worked | 00:01 |
| clarkb | that is probably good for revie wnow if we want to assume that we just need a random keystore and try that | 00:01 |
| clarkb | otherwise maybe a hold and try to test? | 00:02 |
| fungi | i won't really be able to test tomorrow, except possibly in the evening, but sounds worth rolling out | 00:03 |
| fungi | apparently we should be able to set HYPERKITTY_ENABLE_GRAVATAR to false | 00:05 |
| *** ysandeep|out is now known as ysandeep | 00:08 | |
| fungi | i've got a revision prepped for that, just want to wait for the imports to complete before i test adding it on the held server | 00:17 |
| fungi | openstack-stable-maint import is taking a long time at the "Computing thread structure" step | 00:54 |
| fungi | real 68m42.988s | 00:56 |
| fungi | for just that one ml | 00:56 |
| fungi | just the import step | 00:57 |
| fungi | Indexing 99549 emails | 00:57 |
| fungi | it will probably hit 100k messages next week at the current rate | 00:57 |
| fungi | by comparison, openstack-discuss less than a third of that, at 30358 messages | 00:59 |
| fungi | still indexing | 01:09 |
| fungi | 18m23.747s to index it | 01:14 |
| fungi | so roughly 90 minutes to import and index that ml | 01:15 |
| fungi | release-announce is not exactly small either | 01:25 |
| fungi | 13321 messages in it | 01:31 |
| fungi | full site import required 2h24m36s wall clock time | 01:39 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Add a mailman3 list server https://review.opendev.org/c/opendev/system-config/+/851248 | 01:40 |
| fungi | clarkb: ^ that worked to turn off gravatar integration | 01:40 |
| *** rlandy|bbl is now known as rlandy | 02:03 | |
| *** rlandy is now known as rlandy|out | 02:06 | |
| opendevreview | Ian Wienand proposed opendev/system-config master: bootstrap-bridge: drop pip3 role https://review.opendev.org/c/opendev/system-config/+/856593 | 02:47 |
| opendevreview | Ian Wienand proposed opendev/system-config master: bootstrap-bridge: drop pip3 role https://review.opendev.org/c/opendev/system-config/+/856593 | 03:40 |
| *** ysandeep is now known as ysandeep|out | 03:42 | |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] bootstrap-bridge: drop pip3 role/add venv https://review.opendev.org/c/opendev/system-config/+/856593 | 04:51 |
| *** ykarel|away is now known as ykarel | 05:01 | |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] bootstrap-bridge: drop pip3 role/add venv https://review.opendev.org/c/opendev/system-config/+/856593 | 05:06 |
| opendevreview | Ian Wienand proposed opendev/system-config master: install-ansible: remove testinfra version install workaround https://review.opendev.org/c/opendev/system-config/+/852475 | 05:16 |
| opendevreview | Ian Wienand proposed opendev/system-config master: testinfra: install with ansible extras https://review.opendev.org/c/opendev/system-config/+/852476 | 05:16 |
| opendevreview | Ian Wienand proposed opendev/system-config master: install-ansible: remove stevedore workaround https://review.opendev.org/c/opendev/system-config/+/852477 | 05:16 |
| opendevreview | Ian Wienand proposed opendev/system-config master: install-ansible: remove stub install for ARA https://review.opendev.org/c/opendev/system-config/+/852478 | 05:16 |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] bootstrap-bridge: drop pip3 role/add venv https://review.opendev.org/c/opendev/system-config/+/856593 | 05:16 |
| frickler | clarkb: thx for the reminder, cleaned those up | 05:48 |
| frickler | https://zuul.opendev.org/t/openstack/nodes still shows node 0027999262 being held since 8 months now without a matching autohold, seems it got leaked somehow? | 05:49 |
| *** pojadhav|out is now known as pojadhav | 06:00 | |
| *** chkumar|ruck is now known as chandankumar | 06:29 | |
| *** ysandeep|out is now known as ysandeep | 06:49 | |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] bootstrap-bridge: drop pip3 role/add venv https://review.opendev.org/c/opendev/system-config/+/856593 | 06:56 |
| opendevreview | Brin Zhang proposed openstack/project-config master: Add api-ref-job for Venus https://review.opendev.org/c/openstack/project-config/+/856382 | 07:19 |
| *** jpena|off is now known as jpena | 07:35 | |
| elodilles | fungi clarkb : yes, unfortunately, anytime something is fixed and periodic-stable fail count goes down comes another issue :/ for now we have ~44 new failing job on ussuri+train because of new virtualenv release (with new setuptools version bundled in for py37 as well)... I've pushed fix for them already, but needs stable maintainers to review ( https://review.opendev.org/q/topic:cap-virtualenv-py37 | 07:37 |
| elodilles | )... | 07:37 |
| elodilles | (and of course those cover up another gate issues with test jobs that are not covered with periodic-stable...) | 07:39 |
| opendevreview | Brin Zhang proposed openstack/project-config master: Add api-ref-job for Venus https://review.opendev.org/c/openstack/project-config/+/856382 | 07:40 |
| *** bhagyashris is now known as bhagyashris|ruck | 07:57 | |
| opendevreview | Merged openstack/project-config master: Add api-ref-job for Venus https://review.opendev.org/c/openstack/project-config/+/856382 | 07:58 |
| frickler | maybe we should rename the pipeline to "periodic-unstable" ;) | 08:02 |
| opendevreview | Ian Wienand proposed opendev/system-config master: [wip] bootstrap-bridge: drop pip3 role/add venv https://review.opendev.org/c/opendev/system-config/+/856593 | 08:03 |
| *** arxcruz|rover is now known as arxcruz | 08:10 | |
| opendevreview | Frikin Evgenii proposed openstack/diskimage-builder master: Add variable for check installing python3 in yum element https://review.opendev.org/c/openstack/diskimage-builder/+/856577 | 08:16 |
| opendevreview | Frikin Evgenii proposed openstack/diskimage-builder master: Add variable for check installing python3 in yum element https://review.opendev.org/c/openstack/diskimage-builder/+/856577 | 08:20 |
| elodilles | frickler: well, it is stable, as it breaks on a stable way :) | 08:27 |
| *** ysandeep is now known as ysandeep|lunch | 08:29 | |
| *** ysandeep|lunch is now known as ysandeep | 09:48 | |
| gthiemonge | Hi, we have a failure in the octavia tips jobs on ussuri, I don't understand what happens: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_2a7/856399/1/check/octavia-tox-py37-tips/2a7226f/tox/py37-siblings.txt | 09:53 |
| gthiemonge | any ideas? | 09:53 |
| frickler | gthiemonge: I think that is what elodilles mentioned earlier? | 09:55 |
| gthiemonge | oops i should have read the backlog | 09:56 |
| gthiemonge | frickler: thanks | 09:57 |
| gthiemonge | I don't see a patch for octavia, I'll propose one, thanks | 09:57 |
| fungi | gthiemonge: did you notice that because you've been paying attention to the periodic stable job failures, or did it not come to your attention until someone tried to merge a backport? i'm curious how much the various project teams pay attention to periodic job runs, if at all | 10:22 |
| frickler | fungi: that log at least came from a dedicated test patch, so kind of a mixture of both I guess https://review.opendev.org/c/openstack/octavia/+/856399/ | 10:23 |
| fungi | i guess triggered by a failure someone observed in check and then pushed in order to figure out if it was caused by the patch or not | 10:24 |
| fungi | perhaps without knowing they could just check the periodic job results which were also failing | 10:24 |
| frickler | they didn't fail, because only the -tips variant seems to be affected https://zuul.opendev.org/t/openstack/builds?project=openstack%2Foctavia&pipeline=periodic-stable&skip=0 | 10:25 |
| fungi | interesting | 10:30 |
| fungi | and yeah, they're not running any -tips jobs in periodic-stable | 10:32 |
| fungi | so i guess my question is a bit more academic. would they have noticed if it were being run periodically? maybe elodilles would have noticed and pushed a patch for them at least | 10:33 |
| fungi | but if nothing else, that probably explains why they didn't get one | 10:34 |
| *** rlandy|out is now known as rlandy | 10:34 | |
| gthiemonge | fungi: I noticed it in some backports to stable/ussuri | 10:57 |
| fungi | got it. thanks for confirming! | 10:59 |
| *** pojadhav is now known as pojadhav|sick | 11:05 | |
| *** ysandeep is now known as ysandeep|brb | 13:19 | |
| *** dasm|off is now known as dasm | 13:20 | |
| *** ysandeep|brb is now known as ysandeep | 13:30 | |
| *** ysandeep is now known as ysandeep|out | 14:37 | |
| *** dviroel is now known as dviroel|lunch | 14:52 | |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Redirect openstack-infra specs to opendev docs https://review.opendev.org/c/opendev/system-config/+/856828 | 15:21 |
| opendevreview | Merged openstack/project-config master: Add STS-Silicom app to StarlingX https://review.opendev.org/c/openstack/project-config/+/856533 | 15:35 |
| clarkb | fungi: looking at mailman list configuration options at https://docs.mailman3.org/projects/mailman/en/latest/src/mailman/rest/docs/listconf.html one thing we could possibly do is set all lists to not be advertised and not accept posts | 15:35 |
| clarkb | then assuming the import will override those settings we'll be good and new lists can avoid setting the "disable" flag in the first place | 15:35 |
| clarkb | There doesn't seem to be a "this list exists but does not operate" flag though | 15:36 |
| clarkb | I think the set advertised to false and allow_list_posts to false option ins't perfect because people could still potentially subcsribe before migration and maybe the migration would undo their subscription if not subscribed on the mm2 side? also the list admin could potential start it up | 15:37 |
| clarkb | Another option is that we script up the list creation and add it to the import process. I kinda like this option because it helps keep control on the migration side of things | 15:38 |
| clarkb | I know you're AFK for much of today so no rush. But if you can give that some thought and let me know what you think I'm happy to try and update changes or import scripts etc to match | 15:38 |
| fungi | clarkb: alternative... we could just make separate child changes which add the lists for those sites and then if we want to test with them we hold a node from the one at the end of the series, but keep them wip and merge when we're ready for each migration? | 15:43 |
| fungi | less worried about the test nodes because we block outbound e-mail anyway | 15:44 |
| clarkb | ya that should work too. Then merge the changes near when we plan to migrate each site | 15:45 |
| clarkb | separately my java keystore is apparently not a valid keystore in the jvb change :? | 15:45 |
| clarkb | https://zuul.opendev.org/t/openstack/build/e58057733ac94114aae84c31c2e8f26c/log/meetpad01.opendev.org/docker/jitsi-meet-docker_jvb_1.txt#41 | 15:46 |
| *** dviroel|lunch is now known as dviroel | 15:53 | |
| opendevreview | Clark Boylan proposed opendev/system-config master: Update colibri for all the JVBs https://review.opendev.org/c/opendev/system-config/+/856553 | 15:55 |
| clarkb | I'm hoping it was just a case of container vs host paths being different for the same bind mounted file | 15:55 |
| *** marios is now known as marios|out | 15:59 | |
| opendevreview | Brian Rosmaita proposed opendev/system-config master: Redirect openstack-specs to git repo on opendev https://review.opendev.org/c/opendev/system-config/+/856834 | 16:19 |
| *** jpena is now known as jpena|off | 16:35 | |
| clarkb | ok the colibri websocket jvb change looks happy now | 16:50 |
| clarkb | I think that is ready for review and cautious testing/approval | 16:50 |
| fungi | nice! | 16:55 |
| fungi | i can try to look at it once i'm home (definitely not from this rest area parking lot) | 16:56 |
| clarkb | its still an open quetion if the keystore needs coordinated content for authentication/trust, but otherwise I don't see errors in startup | 16:57 |
| clarkb | fungi: I've just double checked on the lists stuff and we do actually define them all on the test node. It was only for the production server where I removed them | 17:29 |
| clarkb | fungi: so they should all be there. When you get a chance can you point at the error you saw? | 17:29 |
| fungi | clarkb: this was the import log... https://paste.opendev.org/show/816668 | 18:48 |
| fungi | unfortunately i seem to have missed redirecting stderr | 18:48 |
| fungi | so i guess i didn't capture it | 18:48 |
| fungi | but i can do another run for another site real quick and do so | 18:49 |
| fungi | i think it's not happening for every list | 18:50 |
| fungi | for example, i saw it when importing legal-discuss@lists.openstack.org | 18:51 |
| clarkb | oh its happening for the lists we don't create? I took the list from the existing mm2 ansible | 18:55 |
| fungi | yeah, that was never complete | 18:56 |
| clarkb | fungi: is our list in ansible incomplete then? That may be the issue we need to reconcile it against relaity. Probably best to start iwth mm2 and then port that over to the mm3 change | 18:56 |
| fungi | we used to just treat the puppet manifest as a convenient way to ask the automation to create a new list, but we never added older lists to it | 18:57 |
| fungi | i can update the ansible to add all the missing ones if that's desirable (seems like it probably is for creating the new server, at least) | 18:58 |
| clarkb | ya I think at the very least we seem to want it for the mm3 change | 18:59 |
| fungi | though that's also an even better reason to split the list creation up into different changes per-site. i'm hoping to do one more pass of retiring unused mailing lists for openstack before we end up migrating lists nobody is using, so that would give us time to do final cleanup between when opendev/zuul have earlier migrations and when the other sites move | 19:02 |
| fungi | i'm happy to do the splitting too while i'm cross-checking all the lists | 19:03 |
| clarkb | fungi: I think we should continue to have all the lists in hte test node. Then have split up changes for the prod node host vars adding each site | 19:03 |
| fungi | oh, right that makes sense | 19:04 |
| fungi | yes i agree with that | 19:04 |
| fungi | i didn't dawn on me those were separated | 19:04 |
| clarkb | I think the mm3 change as is only has opendev lists in the prod hots vars but attempted to have all lists on the test node (and failed at that because the mm2 listing is incomplete) | 19:04 |
| fungi | sure, i can add them | 19:06 |
| *** dviroel is now known as dviroel|out | 20:02 | |
| fungi | clarkb: do we want to migrate https://lists.openstack.org/cgi-bin/mailman/listinfo/test-list do you think? or create a new one for lists.opendev.org? | 21:40 |
| fungi | it has no posts, apparently | 21:41 |
| fungi | i'm leaning toward just deleting it | 21:41 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Add a mailman3 list server https://review.opendev.org/c/opendev/system-config/+/851248 | 21:50 |
| clarkb | fungi: ++ to cleaning it up | 21:50 |
| fungi | clarkb: ^ that's adding all the missing lists and alpha-ordering everything (sites and lists) except for the opendev site is still first since that becomes the default domain | 21:51 |
| fungi | a bit of cleanup we could consider is moving the listdomain parameter into the site name since the name is no longer used | 21:53 |
| fungi | so just name the sites by their fqdns in the future | 21:53 |
| fungi | not urgent we do it before migration though | 21:53 |
| fungi | #status log Deleted unused OpenStack "test-list" mailing list | 21:55 |
| opendevstatus | fungi: finished logging | 21:55 |
| fungi | i can't find any reference to it in our configs, so it may have been created manually | 21:55 |
| clarkb | fungi: we do use it for the apache vhost aliases | 22:00 |
| clarkb | it being listdomain. We might be able to refactor that though | 22:01 |
| clarkb | fungi: for those new lists you can set 'private: true' in the list def for private lists. I thought I had done that before for osme lists but apparently not | 22:14 |
| clarkb | fungi: I know it isn't critical for these lists due to the import process but still probably a good idea to serve as an exampl eto new lists and to help us keep that metadta straight | 22:14 |
| fungi | clarkb: yeah, that's what i mean, refactor to use name instead, and move the listdomain values into name, since the current name is useless, but yes it's not super important | 22:27 |
| clarkb | fungi: right its useless for creating lists but those values inform the apache vhost config as to what aliases to accept | 22:27 |
| clarkb | otherwise the vhosting wouldn't work | 22:27 |
| fungi | and yeah, we can add the private flag, i don't really have an opinion, stuff like description and owner already gets out of sync quickly | 22:27 |
| fungi | clarkb: the listdomain values are important yes, i mean the name values aren't | 22:28 |
| fungi | they used to be the names of the separate directory trees we used for the old multi-domain setup we had | 22:28 |
| clarkb | oh I see the name on the site | 22:29 |
| clarkb | ya I'm not sure if we use that anywhere anymore. Probably not | 22:29 |
| fungi | though maybe we could just drop the "name" parameters entirely, i don't suppose we need to play musical variable names | 22:29 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Add a mailman3 list server https://review.opendev.org/c/opendev/system-config/+/851248 | 23:04 |
| fungi | clarkb: ^ flagged the private lists and also tried removing the name params to see if it still works | 23:04 |
| clarkb | thanks | 23:05 |
| *** dasm is now known as dasm|off | 23:13 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!