| elibrokeit | Gentoo has a pretty good balance for x86-64-v3, we just build all packages twice, once for a baseline repo and once for -v3 | 03:07 |
|---|---|---|
| elibrokeit | curious that the distro maintained by community hobbyists has more resources than a big company... | 03:08 |
| opendevreview | Dale Smith proposed opendev/irc-meetings master: Update magnum meeting time and frequency https://review.opendev.org/c/opendev/irc-meetings/+/950374 | 04:11 |
| tristanC[m] | fungi: clarkb thanks, that's good to know! What I couldn't tell from the matrix next-gen auth was if refresh token were going to be mandatory and/or if m.login.password would become unavailable. | 05:13 |
| opendevreview | Michal Nasiadka proposed opendev/glean master: Add support for CentOS 10 keyfiles https://review.opendev.org/c/opendev/glean/+/941672 | 05:49 |
| opendevreview | Merged opendev/irc-meetings master: Update magnum meeting time and frequency https://review.opendev.org/c/opendev/irc-meetings/+/950374 | 12:36 |
| mnasiadka | tonyb: From my perspective it would be fantastic if we could get some x86-64-v3 labels in Zuul for Kolla to start working on CS10 builds | 12:55 |
| *** ralonsoh is now known as ralonsoh_out | 15:08 | |
| mnasiadka | Anything happened to OSUOSL (the sole aarch64 provider)? | 15:17 |
| clarkb | mnasiadka: https://grafana.opendev.org/d/42e9cd74b4/nodepool3a-osuosl?orgId=1&from=now-6h&to=now&timezone=utc&var-region=$__all it seems to be in use at 100% capacity | 15:18 |
| mnasiadka | Ah, nice - thanks for the grafana link, did not think about checking it ;-) | 15:56 |
| mnasiadka | clarkb: I'm planning to do some more work on CS10 in the coming days - should I also include adding stream 10 mirroring to the list of items? | 16:09 |
| clarkb | mnasiadka: that is likely to be something to plan for but I don't think we should do that until we know how we are dealing with the hardware requirements | 16:11 |
| clarkb | mostly I don't want us to mirror 250GB of pacakges then find it isn't possible to boot the nodes reliably forcing us to clean all that data up | 16:11 |
| tonyb | So on that ..... [please hold] | 16:13 |
| tonyb | https://paste.opendev.org/show/827859 | 16:13 |
| tonyb | I was going to bring that up in the meeting in open discussion | 16:13 |
| tonyb | 40% of our donated resources are able to supply x86_64-v3 nodes, If you/we request a "nested-virt-ubuntu-noble" node we should be fine | 16:15 |
| tonyb | and that's supplied by several clouds/regions | 16:15 |
| mnasiadka | Ok, can try that out | 16:16 |
| tonyb | mnasiadka: For now don't worry about mirroring and the existing DIB patch for CS-10 already uses the correct nodeset | 16:16 |
| mnasiadka | Just trying to estimate how many days that's going to be :) | 16:17 |
| mnasiadka | And I have no clue if mirroring makes sense - e.g. how many centos stream based jobs anyone is running | 16:17 |
| tonyb | mnasiadka: For now the DIB change needs some hacks in nodepool but they're already listed in Depends-On and I'm working on cleaning then up | 16:18 |
| mnasiadka | And Glean needs some additional unit testing based on clarkb's comments - but I can do that tomorrow. | 16:18 |
| tonyb | mnasiadka: IMO, for now I don't think it does make sense, *if* we (opendev) do provide centos-10 nodes for general CI ... at that point it would make sense. | 16:19 |
| tonyb | mnasiadka: Also there is at least one review out there that we could merge to add the mirroring so that part should be quick | 16:19 |
| mnasiadka | good, I'll focus on glean patch for now | 16:20 |
| tonyb | mnasiadka: Thank you! | 16:21 |
| clarkb | tonyb: looking at your paste it looks like only rax classic is not new enough hardware? | 16:24 |
| clarkb | thinking out loud here: maybe we ensure that the nested virt labels are on all the other clouds at this point and then ask anyone running centos 10 continue to adhere to the high level rules for nested virt (use it when necessary be prepared for breakage) | 16:26 |
| clarkb | then if we add centos 10 as a test image it would only be a nested virt label | 16:26 |
| clarkb | (if we adhere to that proposal anyway) | 16:26 |
| tonyb | clarkb: That all seems reasonable. Just because the label is 'nested-virt' doesn't mean you *need* to use nested-virt. | 16:27 |
| clarkb | ya and this way we don't need an additional matrix axis | 16:28 |
| tonyb | clarkb: and yes it does look like it's only rax classic | 16:28 |
| tonyb | My paste/data is based on booting a node in each cloud/region so it's possible I got lucky and not all hypervisors in a given cloud/region are the same but I think that's unlikely and also beyond our sphere of influence | 16:33 |
| clarkb | tonyb: we know in rax classic you sometimes don't get nodes that can v2 | 16:33 |
| clarkb | but that isn't a problem in this case since v2 is too old even | 16:34 |
| tonyb | yeah | 16:35 |
| opendevreview | Clark Boylan proposed opendev/engagement master: Properly query for opened changes during date range https://review.opendev.org/c/opendev/engagement/+/950442 | 16:40 |
| opendevreview | Goutham Pacha Ravi proposed opendev/irc-meetings master: Fix meeting id on openstack-eventlet-removal https://review.opendev.org/c/opendev/irc-meetings/+/950443 | 16:40 |
| opendevreview | Merged opendev/irc-meetings master: Fix meeting id on openstack-eventlet-removal https://review.opendev.org/c/opendev/irc-meetings/+/950443 | 17:14 |
| corvus | do we have zuul jobs that do that now? | 17:45 |
| clarkb | corvus: note only your second message went across the bridge :/ | 17:45 |
| clarkb | corvus: but no I think its a cronjob on one of the servers | 17:45 |
| clarkb | corvus: the afs-release role seems to install it on mirror-update | 17:46 |
| clarkb | job: '/opt/afs-release/release-volumes.py -d >> /var/log/afs-release/afs-release.log 2>&1' | 17:46 |
| corvus | yeah, i think it restarted. first msg was: someone noticed the zuul docs are stale; we may have a stuck afs volume release | 17:47 |
| corvus | 2025-05-20 00:05:02,662 release INFO ... no release required | 17:48 |
| corvus | hrm, the root marker for both the RO and RW volume are the same | 17:50 |
| corvus | Project: zuul/zuul Branch: master Build: 6e3bc6f82adf497aad911a2b61cd2622 Revision: refs/changes/43/949943/1 | 17:50 |
| corvus | that looks like no fault with the vos release, and more like just not having the right content in the RW volume | 17:50 |
| corvus | however, that is the right build | 17:51 |
| clarkb | ya that chagne is newer than the missing content that was called out. So I guess something to do with promoting the docs content | 17:51 |
| fungi | could two changes have raced one another during publication? | 17:51 |
| corvus | https://zuul.opendev.org/t/zuul/build/6e3bc6f82adf497aad911a2b61cd2622 | 17:51 |
| corvus | ya | 17:51 |
| corvus | i think the missing content is like weeks old? | 17:52 |
| corvus | so even if there was a race (there shouldn't be) i think all the race contenders should have had it | 17:52 |
| fungi | we see this ~all the time in openstack due to the volume of changes getting approved in rapid succession | 17:52 |
| clarkb | fungi: that maycause temporary issues but the missing content is from ~ April 14 | 17:52 |
| clarkb | corvus: yup exactly | 17:52 |
| fungi | yeah, so if something merged in the past month that's ruled out | 17:53 |
| clarkb | maybe the builds themselves are excluding the content for some reason | 17:53 |
| clarkb | https://80209368867f82b34e3d-c28848b5e68478b097129cefc05acf76.ssl.cf2.rackcdn.com/zuul/525c746b1a754a4baf4375752e112d07/docs/drivers/gerrit.html ya the content doesn't seem to be in there from the gate build for 949943 | 17:54 |
| corvus | oh yep i see it | 17:54 |
| clarkb | its a bug in the change itself I think | 17:54 |
| clarkb | missing the attr:: I wonder why that doesn't render as just content belonging to the previous entry or error as an invalid list item | 17:55 |
| corvus | questionable rst implementation choice :/ | 17:56 |
| corvus | it's a comment | 17:56 |
| clarkb | oh ha | 17:56 |
| mnasiadka | so I see I'm not the only one having problems with Matrix OFTC bridge missing people/messages | 18:05 |
| clarkb | yes, it mostly seems to affect new irc users | 18:07 |
| clarkb | they don't cross the bridge | 18:07 |
| clarkb | there is an open issue upstream, but I suspect that the lack of input/followup is an indication they have deprioritized the bridges due to the lack of funding (they haven't made any public announcement since the original warning though as far as I can tell) | 18:07 |
| opendevreview | Merged opendev/system-config master: Rotate mailman-core logs https://review.opendev.org/c/opendev/system-config/+/948478 | 19:59 |
| fungi | corvus: clarkb: i'll stack another change to reorder the sections a bit and adjust the prose for better flow | 20:02 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Move donors list to the very bottom of the page https://review.opendev.org/c/opendev/system-config/+/950462 | 20:14 |
| clarkb | fungi: /etc/logrotate.d/45e68e.conf is in place on lists.o.o now | 20:28 |
| clarkb | fungi: and /etc/cron.daily/logrotate.conf runs `/usr/sbin/logrotate /etc/logrotate.conf` I wonder if we can manually run that ahead of schedule or if that will upset the timestamping (would we end up with skipping rotation during the first daily run then not actually automatically rotating until tomorrow's run?) | 20:29 |
| fungi | i'm fine checking in on it in the morning. we already did it for some of the logfiles (for mailman web) just not these yet (mailman-core) | 20:33 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Reword the free tools section of the main page https://review.opendev.org/c/opendev/system-config/+/950464 | 20:55 |
| clarkb | fungi: re that stack of gitea updates it might be a good idea to squash together the set we want to proceed with to avoid unnecessary rolling update of gitea. The main risk there is probably with docker rate limits causing us to have half the servers with one frontpage and half with another. Not the end of the world | 20:57 |
| clarkb | both changes lgtm on the surface but I'll wait for when I can review the preview image before voting | 20:59 |
| opendevreview | Clark Boylan proposed opendev/engagement master: Fix data accuracy bugs https://review.opendev.org/c/opendev/engagement/+/950471 | 23:26 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!