| *** dviroel|rover|afk is now known as dviroel|out | 00:00 | |
| opendevreview | Ian Wienand proposed opendev/system-config master: ptgbot: have apache cache backend https://review.opendev.org/c/opendev/system-config/+/813910 | 00:23 |
|---|---|---|
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: epel: match replacement better https://review.opendev.org/c/openstack/diskimage-builder/+/813922 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu-systemd-container: deprecate and remove jobs https://review.opendev.org/c/openstack/diskimage-builder/+/814068 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu: add Focal test https://review.opendev.org/c/openstack/diskimage-builder/+/814072 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop apt-sources https://review.opendev.org/c/openstack/diskimage-builder/+/814074 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: centos7 : drop functional testing https://review.opendev.org/c/openstack/diskimage-builder/+/814075 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop minimal tests in the gate https://review.opendev.org/c/openstack/diskimage-builder/+/814078 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Remove extras job, put gentoo job in gate https://review.opendev.org/c/openstack/diskimage-builder/+/814079 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Simplify functests job https://review.opendev.org/c/openstack/diskimage-builder/+/814080 | 00:28 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 00:28 |
| opendevreview | Ian Wienand proposed zuul/zuul-jobs master: ensure-podman: support Debian bullseye https://review.opendev.org/c/zuul/zuul-jobs/+/814088 | 00:39 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 00:40 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 01:07 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 02:05 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu-systemd-container: deprecate and remove jobs https://review.opendev.org/c/openstack/diskimage-builder/+/814068 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu: add Focal test https://review.opendev.org/c/openstack/diskimage-builder/+/814072 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop apt-sources https://review.opendev.org/c/openstack/diskimage-builder/+/814074 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: centos7 : drop functional testing https://review.opendev.org/c/openstack/diskimage-builder/+/814075 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop minimal tests in the gate https://review.opendev.org/c/openstack/diskimage-builder/+/814078 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Remove extras job, put gentoo job in gate https://review.opendev.org/c/openstack/diskimage-builder/+/814079 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Simplify functests job https://review.opendev.org/c/openstack/diskimage-builder/+/814080 | 02:25 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 02:26 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu: Fix unbound DIB_LOCAL_IMAGE https://review.opendev.org/c/openstack/diskimage-builder/+/814094 | 02:26 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 02:49 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 03:11 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 03:38 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 04:20 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 04:45 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Revert "Allowing ubuntu element use local image" https://review.opendev.org/c/openstack/diskimage-builder/+/814094 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu-systemd-container: deprecate and remove jobs https://review.opendev.org/c/openstack/diskimage-builder/+/814068 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: ubuntu: add Focal test https://review.opendev.org/c/openstack/diskimage-builder/+/814072 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop apt-sources https://review.opendev.org/c/openstack/diskimage-builder/+/814074 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: centos7 : drop functional testing https://review.opendev.org/c/openstack/diskimage-builder/+/814075 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: functests: drop minimal tests in the gate https://review.opendev.org/c/openstack/diskimage-builder/+/814078 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Remove extras job, put gentoo job in gate https://review.opendev.org/c/openstack/diskimage-builder/+/814079 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Simplify functests job https://review.opendev.org/c/openstack/diskimage-builder/+/814080 | 05:43 |
| opendevreview | Ian Wienand proposed openstack/diskimage-builder master: Run functional tests on Debian Bullseye https://review.opendev.org/c/openstack/diskimage-builder/+/814081 | 05:43 |
| *** odyssey4me is now known as Guest2971 | 06:56 | |
| *** jpena|off is now known as jpena | 07:30 | |
| *** ysandeep|out is now known as ysandeep | 08:01 | |
| *** ysandeep is now known as ysandeep|brb | 10:39 | |
| *** ysandeep|brb is now known as ysandeep | 11:06 | |
| *** jpena is now known as jpena|lunch | 11:31 | |
| *** dviroel|out is now known as dviroel|rover | 11:41 | |
| *** jpena|lunch is now known as jpena | 12:22 | |
| *** dviroel|rover is now known as dviroel|afk | 12:54 | |
| ecsantos[m] | hello folks | 13:32 |
| ecsantos[m] | I asked this in the zuul channel, but I wanted to ask here as well to get more details if possible | 13:32 |
| ecsantos[m] | I'm facing this issue [1] when running the devstack job [2] (I have a local Zuul instance with the same job, roles, diskimages, everything set up just as opendev) | 13:33 |
| ecsantos[m] | [1] https://bugs.launchpad.net/devstack/+bug/1946340 | 13:33 |
| ecsantos[m] | [2] https://zuul.opendev.org/t/openstack/job/devstack | 13:33 |
| ecsantos[m] | on the zuul channel, corvus said to me that the reason for this is probably that opendev has old wheel builds in the caches that it's using | 13:33 |
| ecsantos[m] | could someone clarify what that means? I'm not sure. I see opendev uses a custom mirror [3], does this have something to do with how opendev isn't affected by the issue? | 13:35 |
| ecsantos[m] | [3] http://mirror.dfw.rax.opendev.org/ubuntu | 13:35 |
| ecsantos[m] | because apparently everyone trying to deploy devstack are currently hitting this issue, except the opendev ci | 13:36 |
| ecsantos[m] | s/are/is/ | 13:36 |
| ecsantos[m] | s/in/on/ | 13:36 |
| fungi | ecsantos[m]: yes, in order to speed up our jobs, we build platform-specific python wheels for any direct or transitive dependencies of the projects listed in the openstack/requirements repository and we cache those indefinitely | 13:38 |
| fungi | a recent change to setuptools (starting in 58.0.0) caused it to no longer be possible to build some old wheels for python projects which rely on deprecated (and now removed) features of setuptools | 13:39 |
| fungi | however, we built those wheels before setuptools 58 was released and so still have them sitting around in our cache | 13:39 |
| fungi | an example is the suds-jurko python library. it can't be installed from the sdist (source tarball) published for it on pypi when used with setuptools >= 58 | 13:40 |
| fungi | but we precompiled and cached some platform-specific wheels of it long ago, well before setuptools broke it | 13:41 |
| fungi | however, we can't run jobs which rely on it on newer platforms we're adding, for the same reason: we don't have wheels of it pre-built for those new platforms | 13:42 |
| fungi | there are workarounds, depending on what's orchestrating the installation of python packages. for example tox can be sold to use an older version of itself which installs an older virtualenv that includes setuptools 57.x | 13:43 |
| fungi | or devstack can be told to downgrade the environment to setuptools<58 before installing anything else | 13:44 |
| fungi | the "solution" on our side is working out a way to purge now-unbuildable wheels from our own cache, so that we don't silently mask problems like this. we have some designs in progress to achieve that | 13:46 |
| Clark[m] | And the fix on the openstack side is https://opendev.org/openstack/oslo.vmware/commit/7bc343173751330d2e989c28bc6fefed56d358d6 but I think that needs a release next? I wonder if anyone has requested that yet | 13:46 |
| fungi | the proper solution of course is to get projects to stop using those old, abandoned dependencies like suds-jurko, and switch to maintained alternatives, or to revive development on them and fix them to work with modern setuptools | 13:47 |
| fungi | ecsantos[m]: does that answer your questions? | 13:48 |
| opendevreview | Cody Lee Cochran proposed opendev/bindep master: feat: Add support for Manjaro Linux https://review.opendev.org/c/opendev/bindep/+/814171 | 13:55 |
| Clark[m] | Speaking of releases did we ever do that gear release? I think I forgot, but should put a post it up for it I guess | 13:58 |
| ecsantos[m] | fungi: your explanation was very clear, thank you! it helped a lot. I'd prefer waiting for the proper fix, but my team is on a tight schedule to review some downstream patches in our local CI. do you think it's feasible for us to "mirror" the setup opendev has in their mirrors? with the old cached wheels | 13:58 |
| ecsantos[m] | Clark: I saw that this change [4] was merged, but it didn't fix the issue for me (the only conflict I'm having is with oslo.vmware). does anything else needs to be merged for the problem to be solved? | 14:01 |
| ecsantos[m] | [4] https://review.opendev.org/c/openstack/oslo.vmware/+/813377 | 14:01 |
| ecsantos[m] | s/needs/need/ | 14:02 |
| ecsantos[m] | * it's feasible/doable for | 14:02 |
| ecsantos[m] | s/fungi: your explanation was very clear, thank you! it helped a lot. I'd prefer waiting for the proper fix, but my team is on a tight schedule to review some downstream patches in our local CI. do you think it's feasible for us to "mirror" the setup opendev has in their mirrors? with the old cached wheels/fungi: your explanation was very clear, thank you! it helped a lot. I'd prefer waiting for the proper fix, but my team is on a tight | 14:03 |
| ecsantos[m] | schedule to review some downstream patches in our local CI. do you think it'd feasible/doable for us to "mirror" the setup opendev has in their mirrors? with the old cached wheels. not sure how much work was that for you guys/ | 14:03 |
| ecsantos[m] | s/fungi: your explanation was very clear, thank you! it helped a lot. I'd prefer waiting for the proper fix, but my team is on a tight schedule to review some downstream patches in our local CI. do you think it's feasible for us to "mirror" the setup opendev has in their mirrors? with the old cached wheels/fungi: yes, your explanation was very clear, thank you! it helped a lot. I'd prefer waiting for the proper fix, but my team is on a | 14:03 |
| ecsantos[m] | tight schedule to review some downstream patches in our local CI. do you think it'd feasible/doable for us to "mirror" the setup opendev has in their mirrors? with the old cached wheels. not sure how much work was that for you guys/ | 14:03 |
| Clark[m] | ecsantos: they need to make a release of oslo.vmware with that fix in it. Then they need to update constraints to install that new release | 14:04 |
| fungi | Clark[m]: yeah, we keep remembering on fridays that we wanted to make a release of gear, and then saying we should wait for monday... *sigh* | 14:06 |
| ecsantos[m] | Clark: oh I see... oslo.vmware is not cloned from master, my mistake. thanks! | 14:09 |
| fungi | ecsantos[m]: you can either ask in #openstack-oslo for them to propose a release, or propose one to the openstack/releases repository yourself if you're familiar with the process | 14:10 |
| Clark[m] | https://review.opendev.org/c/openstack/releases/+/814012 now you need the constraints update then everything is probably fine | 14:10 |
| fungi | ahh, so it's already in progress | 14:11 |
| Clark[m] | https://review.opendev.org/c/openstack/requirements/+/814148 | 14:11 |
| fungi | yep, merged an hour ago | 14:11 |
| fungi | the release i mean | 14:11 |
| fungi | i've pinged the requirements project lead asking to expedite approval of that | 14:14 |
| ecsantos[m] | thanks fungi and Clark for the info and help! if the release is on the way I believe my team can wait :) | 14:17 |
| fungi | the release is actually done for the past hour, it's the constraints update in requirements which is the remaining piece | 14:18 |
| fungi | once 814148 merges you should hopefully be working again (or at least find out what the next problem is) | 14:18 |
| ecsantos[m] | lol I hope it's only one! :D | 14:20 |
| *** ysandeep is now known as ysandeep|out | 14:22 | |
| Clark[m] | fungi you should probably go ahead and approve https://review.opendev.org/c/opendev/system-config/+/813886 so that it is ready for us later today. Also we should update the etherpad to make sure there is a check ssh connectivity step and add instructions and a list for running the gitea sync playbook | 14:28 |
| Clark[m] | I can help with all that after the school run, but figured getting that approved early was a good idea | 14:29 |
| fungi | yep, approving now | 14:31 |
| fungi | psf is seeking user feedback on pypi: https://discuss.python.org/t/pypi-user-feedback | 14:38 |
| rosmaita | is ethercalc.openstack.org down? | 14:50 |
| fungi | i can ssh to it and there's a node process running for the past week | 14:51 |
| fungi | but it's possible the service is hung | 14:51 |
| fungi | #status log Restarted the Ethercalc service on ethercalc.openstack.org as it was not responding to proxied connections from Apache any longer | 14:54 |
| opendevstatus | fungi: finished logging | 14:54 |
| fungi | rosmaita: should be working now, i'm still trying to dig into logs for it | 14:54 |
| rosmaita | fungi: ty! | 14:55 |
| clarkb | fungi: I've updated the etherpad with notes on running the sync-gitea-projects.yaml playbook. Can you double check my use of -e and a vars file in that? I think I got it right | 15:27 |
| clarkb | re Gerrit MINA server-sig-algs someone has posted kotlin to the bug showing how to make it work in existing MINA sshd installations. I guess in theory we could translate that to Java and incorporate it into Gerrit. I doubt I'll have time to start looking at that today, but good to know we have that option if we can sort out the java for it | 15:30 |
| fungi | clarkb: have the pad url handy? otherwise i can grab it out of the channel history just need to dig | 15:30 |
| fungi | also now i need to look up kotlin | 15:31 |
| opendevreview | Merged opendev/system-config master: Allow gitea_create_repos always_update to be list https://review.opendev.org/c/opendev/system-config/+/813886 | 15:32 |
| clarkb | fungi: https://etherpad.opendev.org/p/project-renames-2021-10-15 | 15:32 |
| fungi | thanks! | 15:32 |
| clarkb | kotlin is a newer jvm language that is gaining popularity due to android | 15:32 |
| clarkb | you can actually link between the two, but as far as I know gerrit is java only so far and I wouldn't want to also propose adding kotlin to their code base as part of this | 15:32 |
| clarkb | it also sounds like MINA itnends on fixing this by default without users needing to manually add support so if we wait long enough it should get fixed on its own as well | 15:33 |
| clarkb | java/com/google/gerrit/sshd/SshDaemon.java's SshDaemon is the class to update I think | 15:34 |
| fungi | clarkb: the amended plan for the rename addressing the metadata updates looks right to me | 15:35 |
| clarkb | cool, I think we're about as ready as we can be then. Just have to work through the steps in the etherpad at 18:00UTC | 15:36 |
| fungi | i guess we can use https://meetpad.opendev.org/project-renames-2021-10-15 to coordinate | 15:44 |
| fungi | though i'll also take some notes in here | 15:44 |
| clarkb | ya thats fine | 15:46 |
| fungi | and 813886 deployed, so the system-config checkout on bridge should properly include it now | 15:47 |
| *** dviroel|afk is now known as dviroel|rover | 15:52 | |
| *** marios is now known as marios|out | 16:03 | |
| fungi | i'm going to be testing meetpad with firefox... assuming that goes well, we could maybe consider adding that to the list of browsers it doesn't warn about | 16:17 |
| fungi | we've had quite a few folks in recent months, including at the last ptg, mention they were working great on ff | 16:18 |
| fungi | i tried it myself a few minutes ago and it seemed to be functioning even with my paranoid settings and blockers, as long as i allowed it to request mic and cam access | 16:19 |
| *** jpena is now known as jpena|off | 16:41 | |
| clarkb | neat | 16:51 |
| clarkb | I got completely sucked into rewriting this kotlin in java | 16:51 |
| clarkb | I shouldn't have but it nerd sniped me | 16:51 |
| fungi | hah | 16:58 |
| fungi | it totally happens | 16:58 |
| tristanC | clarkb: couldn't gerrit use the kotlin implementation? | 17:03 |
| fungi | sounds like it would need a different jvm for that? | 17:07 |
| fungi | or at least different compiler | 17:07 |
| clarkb | tristanC: it can, but I don't want to figure out how to integrate kotlin into their java nor do I want to argue that they should write kotlin | 17:10 |
| clarkb | tristanC: the existing code is java 8 compat too, I don't know if that matters | 17:10 |
| clarkb | tristanC: basically I'd spend days doing the kotlin integration or just a few hours figuring out the rewrite | 17:10 |
| clarkb | I think I'm close. The kotlin version has a bug as session doesn't have a signatureFactories member which I'm trying to sort out now. | 17:15 |
| clarkb | aha I think kotlin might automagic getters and setters. So I have to explicitly call get in this case | 17:23 |
| yuriys | is there a tempest/refstack irc channel? | 17:24 |
| tristanC | clarkb: oh, i thought kotlin could easily be integrated in a java codebase | 17:25 |
| fungi | yuriys: there are both | 17:26 |
| yuriys | ah, im assuming they're hidden from public listing just like this one, ty | 17:26 |
| fungi | yuriys: looks like there's a #refstack | 17:27 |
| fungi | probably better to start there if the questions are refstack-specific | 17:27 |
| fungi | yuriys: if you have more general questions about the tempest framework, then #openstack-qa is the primary place its developers hang out | 17:28 |
| yuriys | yeah that probably fits more, ty fungi | 17:28 |
| clarkb | tristanC: I'm sure it can be if you understand the codebase and bazel and java and kotlin :) | 17:29 |
| clarkb | tristanC: I understand only one of those things and not very well (java) | 17:29 |
| fungi | yuriys: also a convenient place to look for channels we're publicly logging is https://meetings.opendev.org/irclogs/ | 17:29 |
| yuriys | oh nice! | 17:29 |
| fungi | we do set all the channels to "hidden" in the ircd because apparently spammers use the public channels list as an invitation to... spam | 17:29 |
| fungi | so setting them hidden from the list cuts down on the work for our chanops | 17:30 |
| clarkb | fungi: I guess we're up in about 25 minutes. I'll put the java down for now (it compiles. next step is linting, then making a patch against our gerrit builds so we can get a test gerrit to test against) | 17:34 |
| fungi | yeah, i need to gear up | 17:37 |
| clarkb | fungi: you should review my rename yaml file and copy it to bridge if it looks good to you | 17:43 |
| clarkb | I'm going to grab a quick snack and then can hop on the meetpad for that etherpad | 17:43 |
| fungi | review it in the etherpad? | 17:43 |
| fungi | oh, the opendev/project-config change | 17:43 |
| fungi | on it | 17:44 |
| clarkb | looks like we're late on adding things to emergency file. | 17:46 |
| fungi | oh, i can do that now too | 17:46 |
| clarkb | thats ok nothing is running right now so if we add things to the file now we're good | 17:46 |
| * clarkb loads ssh keys | 17:46 | |
| clarkb | oh shoot I just remembered a thing. Zuul secrets | 17:48 |
| fungi | i've added the three indicated entries to /etc/ansible/hosts/emergency.yaml but double-checking when you have a second can't hurt so we don't get an unpleasant surprise | 17:48 |
| clarkb | oh I guess we updated the rename playbook to do zuul secret renames already. yay old us | 17:49 |
| clarkb | fungi: I think I've always written out the giteas as I'm not sure how globbing/regexes work in that file | 17:50 |
| clarkb | historically we've had problems with treating regexes as globs and vice versa iirc | 17:50 |
| clarkb | that always makes me paranoid enough to just list them out | 17:50 |
| fungi | i'll do that then | 17:50 |
| fungi | better? | 17:51 |
| clarkb | yup thanks | 17:52 |
| clarkb | I've added the contents of the opendev/project-config rename record file to bridge in the spot where we will -e repolist=$path | 17:52 |
| clarkb | if you can check the contents of the change and that file that would be great | 17:52 |
| fungi | status notice The Gerrit service on review.opendev.org will be offline starting in 5 minutes, at 18:00 UTC, for scheduled project rename maintenance, which should last no more than an hour (but will likely be much shorter): http://lists.opendev.org/pipermail/service-announce/2021-October/000024.html | 17:53 |
| fungi | for sending in a minute or two | 17:53 |
| fungi | and yeah i was trying to compare the renames file against the proposed projects.yaml changes | 17:54 |
| clarkb | fungi: you should double check the inventory name for storyboard too. i think it may be more complex than storyboard.opendev.org | 17:54 |
| clarkb | storyboard01.opendev.org ? | 17:54 |
| clarkb | I can ssh to all of the hosts from here | 17:54 |
| fungi | yeah, fixing | 17:55 |
| fungi | that's what i get for copying from the pad without checking | 17:55 |
| fungi | good with the proposed status notice? | 17:56 |
| fungi | the renames list has 20 repos, but i only count 19 between the two projects.yaml patches | 17:57 |
| clarkb | ya the notice lgtm | 17:57 |
| clarkb | fungi: there are three patches | 17:57 |
| fungi | #status notice The Gerrit service on review.opendev.org will be offline starting in 5 minutes, at 18:00 UTC, for scheduled project rename maintenance, which should last no more than an hour (but will likely be much shorter): http://lists.opendev.org/pipermail/service-announce/2021-October/000024.html | 17:57 |
| opendevstatus | fungi: sending notice | 17:57 |
| -opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline starting in 5 minutes, at 18:00 UTC, for scheduled project rename maintenance, which should last no more than an hour (but will likely be much shorter): http://lists.opendev.org/pipermail/service-announce/2021-October/000024.html | 17:58 | |
| fungi | oh, the third one doesn't have our usual topic set | 17:58 |
| clarkb | that may have been my mistake when I rebased it | 17:58 |
| clarkb | it confliced with the other change so I did the rebase | 17:58 |
| fungi | added the topic now | 17:58 |
| fungi | renames list lgtm, will copy to bridge | 17:59 |
| clarkb | fungi: I copied it, you just need to double check the contents | 17:59 |
| fungi | oh, will do | 17:59 |
| fungi | i've also started a root screen session on bridge | 18:00 |
| clarkb | cool I'll join that momentarily. I'm on the meetpad too fi you want to join | 18:00 |
| *** ysandeep|out is now known as ysandeep | 18:00 | |
| clarkb | I'm on the screen | 18:00 |
| *** dviroel|rover is now known as dviroel|rover|afk | 18:01 | |
| fungi | service should be going down and up any minute now | 18:05 |
| *** dviroel|rover|afk is now known as dviroel|rover | 18:09 | |
| opendevreview | Merged openstack/project-config master: Rename inspur/*venus* project to openstack/*venus* https://review.opendev.org/c/openstack/project-config/+/808149 | 18:20 |
| opendevreview | Merged openstack/project-config master: Move osf/ namespace repos to openinfra/ namespace https://review.opendev.org/c/openstack/project-config/+/808479 | 18:21 |
| opendevreview | Merged openstack/project-config master: Move ansible-role-refstack-client from x/ to openinfra/ https://review.opendev.org/c/openstack/project-config/+/765787 | 18:22 |
| opendevreview | Merged opendev/project-config master: Record renames being performed on October 15, 2021 https://review.opendev.org/c/opendev/project-config/+/813902 | 19:53 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Fix gitea repo syncing always_update flag type https://review.opendev.org/c/opendev/system-config/+/814227 | 19:58 |
| clarkb | fungi: ^ there is the slight chance that our list input is also being cast to a string. If that is happening then we will fail the existing testing for ^. I'm going to go look at old job logs to try and figure out if that is happening and then we can head scratch more. Probably a revert is the first step if that is a problem. Then we fix it properly with a new chnage | 19:59 |
| clarkb | fungi: "always_update": "['opendev/disk-image-builder']", | 20:01 |
| clarkb | I think that is indeed an issue? | 20:01 |
| clarkb | fungi: I'll push a revert whcih we can land today then regroup next week and decide how we want to address this | 20:01 |
| clarkb | (its also quick enough that maybe we consider always update to just be set to true instead) | 20:02 |
| clarkb | though not very quick it is quick enough maybe | 20:02 |
| clarkb | https://8145bb2d9385627a8710-4ce80c083e58c5170f82b282f505c84d.ssl.cf2.rackcdn.com/813886/10/gate/system-config-run-gitea/828e2ab/bridge.openstack.org/ara-report/results/492.html shows it updating settings for all the repos like we saw in prod | 20:03 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Revert "Allow gitea_create_repos always_update to be list" https://review.opendev.org/c/opendev/system-config/+/814228 | 20:04 |
| clarkb | fungi: ^ I think we should land the revert then rethink this. | 20:04 |
| clarkb | and now I must lunch | 20:05 |
| fungi | yeah, i agree, ansible may make this hard. i was expecting to be able to rely on yaml's type inference, did not expect that not telling ansible the variable's type would cause it to always recast to a string unconditionally regardless of what type yaml interprets it to be | 20:12 |
| fungi | which means to do a ternary there we have to invent our own serialization language now, because ansible won't let us use the serialization language it's based on | 20:13 |
| fungi | when i repropose i'll also add the additional tests we considered | 20:15 |
| fungi | which would have showed the problem | 20:15 |
| fungi | alternatively, we don't bother to unrevert, we just run the full update after any rename | 20:16 |
| fungi | though being able to have it dynamically update the issues url when a projects.yaml change lands to change the tracker for a particular project would be nice, or for description changes | 20:17 |
| fungi | might be we just want to drop the false default and let it always update all projects any time the manage-projects playbook is called, now that we have evidence it only takes a couple minutes to complete across the cluster | 20:18 |
| clarkb | fungi: ya I think we should consider that | 20:20 |
| clarkb | If we do that we should rever your change then flip the default to true | 20:20 |
| clarkb | why don't we start with the revert then we can discuss in a change that flips it to true? and then the testing from your old change will be helpful to verify things do what we want | 20:21 |
| fungi | yeah, i already approved the revert | 20:22 |
| clarkb | cool sounds like a plan | 20:22 |
| fungi | just thinking that having it always run for all projects might be the simplest way forward next | 20:22 |
| fungi | because really, we don't land changes which trigger it all that often | 20:23 |
| fungi | once or twice a day on average, if that | 20:23 |
| fungi | it was a problem back when we first migrated from cgit, but it seems like gitea has gotten many orders of magnitude more efficient at handling those calls in the years since | 20:24 |
| fungi | or more than an order of magnitude faster at the least | 20:25 |
| clarkb | I think the switch to using gitea's true api rather than manipulating the web ui helped alot | 20:25 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Push a patch to test a MINA server-sig-algs fix in Gerrit https://review.opendev.org/c/opendev/system-config/+/814230 | 20:34 |
| clarkb | I'm going to put a hold in place for the system-config-run-gerrit-3.3 job for that change. | 20:34 |
| fungi | cool | 20:40 |
| clarkb | neat my patch doesn't apply/work with 3.4 or master | 20:59 |
| clarkb | one thing at a time | 20:59 |
| *** dviroel|rover is now known as dviroel|rover|afk | 21:08 | |
| fungi | hah | 21:50 |
| opendevreview | Merged opendev/system-config master: Revert "Allow gitea_create_repos always_update to be list" https://review.opendev.org/c/opendev/system-config/+/814228 | 21:59 |
| clarkb | Looks like upstream mina has a fix up so my time might be better spent figuring out how to update mina in gerrit. | 22:45 |
| clarkb | tools/nongoogle.bzl <- is the file to update to get a newer mina | 22:48 |
| clarkb | nongoogle beacuse google gerrit doesn't use those libs so they get hidden away in a corner rather than google adding/removing what they need from the open source project :/ | 22:48 |
| clarkb | hrm and my test has an NPE | 22:54 |
| fungi | the best kind of pointer exception | 23:17 |
| clarkb | unfortunately mina does a good job of supressing those too :/ | 23:21 |
| clarkb | https://github.com/apache/mina-sshd/blob/bf4000ce5b56ce8c5b893ad2620858e1309c1338/sshd-core/src/main/java/org/apache/sshd/common/session/helpers/SessionHelper.java#L1236-L1237 is catching it and not giving me a traceback | 23:22 |
| clarkb | if I was a proper java dev I'd have my IDE open and set a breakpoint on that line and inspect it from thee | 23:22 |
| clarkb | At this point probably better off waiting for MINA to release their update and then work with gerrit to update MINA versioning there | 23:26 |
| clarkb | https://github.com/apache/mina-sshd/pull/204/ is the MINA fix | 23:27 |
| fungi | there is rather a bit of docstring boilerplate in that | 23:36 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!