Friday, 2026-01-23

-@gerrit:opendev.org- Zuul merged on behalf of Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org: [opendev/system-config] 974381: Grant fungi access to statusbot in Matrix https://review.opendev.org/c/opendev/system-config/+/97438100:32
-@gerrit:opendev.org- László Jánosi proposed: [opendev/glean] 974346: Fix vlan and bond config generation https://review.opendev.org/c/opendev/glean/+/97434608:00
@harbott.osism.tech:regio.chatseems we were having issues on the docs server yesterday, too, see e.g. https://zuul.opendev.org/t/openstack/build/58c14c070bd749a5943e303cb9c004da will see whether the recheck will work now, manually checking some of the affected urls did work fine for me right now09:25
@harbott.osism.tech:regio.chatregarding statusbot auth, I don't think I'd want my current matrix account visible in a repo, do you think we could consider handling that via a secret?09:28
@harbott.osism.tech:regio.chatI'm also noticing that I've had to completely mute this channel in order to avoid being overloaded by notifications, sadly this also seems to result in the `infra-root` trigger to be muted. for now please ping me in IRC if something urgent comes up09:29
@fungicide:matrix.orgwe could probably set the statusbot nicklist in a private hostvar rather than the public inventory14:32
@clarkb:matrix.orgfungi: Jens Harbott yes, that same variable can be set in private vars and I believe the private var version should win due to variable precedence. But that is also why I didn't blindly assume I should add everyone to that list in the repo15:45
@clarkb:matrix.orgalso for notifications element lets you set the notification to on which is just a visual indicator or noisy which is the audible ping as well as the visual indicator15:47
@clarkb:matrix.orgthat may be helpful if the problem is the audio portion of the notification15:47
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [opendev/irc-meetings] 974444: Update irc meetings index page with Matrix https://review.opendev.org/c/opendev/irc-meetings/+/97444415:54
@clarkb:matrix.orgfungi: I've double checked screenshots for https://review.opendev.org/c/opendev/system-config/+/970451 and did a quick skim of the bazel build logs. I don't see anything concerning with the newer version of nodejs. Is today still a good day for a Gerrit restart on your end? And if so should I approve this change now?15:56
@fungicide:matrix.orgyeah, let's go for it15:57
@clarkb:matrix.orgchange is approved15:59
@fungicide:matrix.orgstanding by to schedule a restart once it promotes/deploys16:00
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [openstack/project-config] 974462: Simplify Zuul tenant config for providers https://review.opendev.org/c/openstack/project-config/+/97446216:07
-@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [opendev/irc-meetings] 974444: Update irc meetings index page with Matrix https://review.opendev.org/c/opendev/irc-meetings/+/97444416:22
@fungicide:matrix.orglooks like the gerrit change is about to merge17:19
-@gerrit:opendev.org- Zuul merged on behalf of Monty Taylor https://matrix.to/#/@mordred:inaugust.com: [opendev/system-config] 970451: Update node to 24 for Gerrit builds https://review.opendev.org/c/opendev/system-config/+/97045117:19
@fungicide:matrix.orgdeploy jobs are already running17:21
@fungicide:matrix.organd done17:22
@fungicide:matrix.orgi guess we can prepare to restart any time now17:22
@clarkb:matrix.orgelodilles mentioned trying to get things prepped for some Openstack release stuff and mentioned being done by 1800 UTC. I said that we could wait until then17:23
@fungicide:matrix.orgyep, should be fine17:23
@clarkb:matrix.orgBut yes I think we should start prepping and then restart around then17:23
@fungicide:matrix.orgi have a root screen session open on review0317:24
@fungicide:matrix.orgquay.io/opendevorg/gerrit       3.11      f43f988a6736   6 days ago     718MB17:25
@clarkb:matrix.orgThe thing we didn't do last time that we should be better about is pausing the Gerrit results queue 17:25
@fungicide:matrix.orgthat's what we're running now, i think17:25
@clarkb:matrix.orgcorvus: do you know if the queue pausing script has been fixed in zuul or do we still ahve to use the special copy on zuul01?17:25
@clarkb:matrix.orgfungi: yes that image look correct17:26
@clarkb:matrix.orgas the current image I mean17:26
@fungicide:matrix.orgi'll go ahead and pull unless there are objections17:26
@clarkb:matrix.orgfungi: I think you can pull17:27
@jim:acmegating.comClark: refresh my memory please, why was it broken?  or alternatively, where's the special copy?17:27
@clarkb:matrix.org`/root/zuul-client manage-events --all-tenants --reason "Gerrit restart in progress" pause-result` and `/root/zuul-client manage-events --all-tenants normal` are the two pause commands17:27
@clarkb:matrix.orgcorvus: ^ that /root/zuul-client special copy is on zuul01. I don't recall what the specific fix was17:28
@jim:acmegating.comthat looks like it's just a docker run command to simplify using zuul-client17:28
@clarkb:matrix.orgah ok so maybe it did get fixed17:28
@fungicide:matrix.orgquay.io/opendevorg/gerrit       3.11      5bebd6c38d59   51 minutes ago   716MB17:28
@fungicide:matrix.org`"RepoDigests": [ "quay.io/opendevorg/gerrit@sha256:9081c24e5ca0b9f87dcacc038cc9f7a33fdee739ee11f4acdcbce0e62424d9aa", "quay.io/opendevorg/gerrit@sha256:fce878a59a63f1decf78c5d330f6bbaae453354e04ffeae3694221605feb0277"`17:29
@clarkb:matrix.orgfungi: the inspect command on 5beb... seems to match the hashes at https://quay.io/repository/opendevorg/gerrit/manifest/sha256:9081c24e5ca0b9f87dcacc038cc9f7a33fdee739ee11f4acdcbce0e62424d9aa17:30
@jim:acmegating.comit doesn't look like there was an update to zuul-client manage-events after the initial merge...17:30
@clarkb:matrix.orghuh those are the commands in root's history for `manage-events` on zuul01. Let me see if I can find more info in IRC historiy17:31
@jim:acmegating.comis there a possibility it wasn't broken?  :)17:31
@fungicide:matrix.orgi've got this queued up in the screen, ready to execute at 18:00 utc: `docker compose -f /etc/gerrit-compose/docker-compose.yaml down && mv ~gerrit2/review_site/data/replication/ref-updates/waiting ~gerrit2/tmp/waiting_queue_2026-01-23 && rm ~gerrit2/review_site/cache/{gerrit_file_diff,git_file_diff,git_modified_files,modified_files,comment_context}.* && sudo docker compose -f /etc/gerrit-compose/docker-compose.yaml up -d`17:32
@clarkb:matrix.orgcorvus: https://review.opendev.org/c/opendev/system-config/+/967111 this was the fix according to IRC history. The bug was in the helper script not the zuul-client tool17:32
@jim:acmegating.comi do notice that zuul-client is building from dockerhub.  and https://review.opendev.org/953231 never merged but i think events have passed that one by and we need a new change now?17:32
@fungicide:matrix.orghow's this look for the announcement? `status notice The Gerrit service on review.opendev.org will be offline momentarily while we restart it for an upgrade to NodeJS 24`17:34
@clarkb:matrix.orgfungi: that command lgtm. We probably don't need to clear out all of those caches but ti didn't seem to hurt last time so I think we can proceed with that17:34
@clarkb:matrix.orgfungi: In the announcement I would say something like "while we restart to pick up new container images after updating the build tooling" or similar. Nodejs isn't a runtime thing onl a build side item17:35
@clarkb:matrix.orgmostly I don't want to confuse people and have them think gerrit is now a nodejs application17:35
@fungicide:matrix.orgso more like: `status notice The Gerrit service on review.opendev.org will be offline momentarily while we restart it onto new images after an update to the build toolchain`17:37
@clarkb:matrix.org++17:37
@clarkb:matrix.orgfungi: I can run the zuul-client commands to pause then unpause things17:51
@fungicide:matrix.orgthanks17:51
@fungicide:matrix.orgready for me to send the notice?17:59
@clarkb:matrix.orgyup17:59
@fungicide:matrix.org#status notice The Gerrit service on review.opendev.org will be offline momentarily while we restart it onto new images after an update to the build toolchain17:59
@status:opendev.org@fungicide:matrix.org: sending notice17:59
@fungicide:matrix.orgyay, it's working this time18:00
@clarkb:matrix.orglet me know when I should pause zuul queues18:00
-@status:opendev.org- NOTICE: The Gerrit service on review.opendev.org will be offline momentarily while we restart it onto new images after an update to the build toolchain18:02
@status:opendev.org@fungicide:matrix.org: finished sending notice18:02
@clarkb:matrix.orgmaybe now that the notice is done I should stop queues then we restart gerrit?18:03
@fungicide:matrix.orgClark: i guess go ahead, and then i'll restart gerrit once you're done18:03
@clarkb:matrix.orgI'm getting a token has expired error from zuul-client18:03
@clarkb:matrix.orgcorvus:  is this something I can do logged into the web ui?18:04
@fungicide:matrix.orgi bet that's fallout from the rotation18:04
@clarkb:matrix.orgyes, I'm guessing the old token has no expiration or it was like 10 years in the future and this one doesn't18:04
@fungicide:matrix.orgi updated the token for zc18:04
@fungicide:matrix.orgoh, maybe that's my fault18:04
@fungicide:matrix.orgi think i just used the command in the zuul docs18:05
@clarkb:matrix.orgI tried on zuul02 and got the same thing18:05
@clarkb:matrix.orglooks like I can do it through the web ui so give me a minute18:06
@clarkb:matrix.orgfungi: that is done18:08
@clarkb:matrix.orgcorvus: looks like with the cli you can pause only result queues but the web ui only lets you do trigger + result queues (so I did that)18:08
@clarkb:matrix.orgoh and it is per tenant?18:08
@fungicide:matrix.orghttps://zuul-ci.org/docs/zuul/latest/client.html#create-auth-token is what i did previously18:08
@jim:acmegating.comClark: the web ui is per-tenant, the script does all tenants18:09
@clarkb:matrix.orgcorvus: ack I did openstack and zuul tenants as they are the active ones right now18:09
@clarkb:matrix.orgI think that should be sufficient for this restart18:09
@fungicide:matrix.orgi wonder if i actually copied `--expires-in 1800`18:10
@clarkb:matrix.orgfungi: looks like an extra character snuck into your command in screen18:10
@jim:acmegating.comClark: pause-result includes pause-trigger18:11
@clarkb:matrix.orgI don't knwo if I did that with my xmonad window switching18:11
@clarkb:matrix.orgfungi: its -n d instead of -d18:11
@fungicide:matrix.orgah, i did `sudo docker run --rm --volume=/etc/zuul:/etc/zuul -it quay.io/zuul-ci/zuul-scheduler:latest zuul-admin create-auth-token --auth-config local --user localauth` but not specifying `--expires-in` defaulted to 600 (10 minutes i guess)18:11
@fungicide:matrix.orgmea culpa!18:11
@clarkb:matrix.orgfungi: I think we can proceed if we fix the docker compose up -d command18:12
@fungicide:matrix.organd yeah, fixed the command in screen already18:12
@clarkb:matrix.orgoh yup it is fixed now18:12
@fungicide:matrix.orgrunning18:12
@fungicide:matrix.orgfunny that stopping gerrit is taking longer this time even though it's only been a week18:14
@fungicide:matrix.org106.3s18:14
@clarkb:matrix.orggit_file_diff.h2.db is the last h2 db file with a lock file in place so I think it is processing that one18:14
@clarkb:matrix.org`[2026-01-23T18:14:47.668Z] [main] INFO  com.google.gerrit.pgm.Daemon : Gerrit Code Review 3.11.7-22-g66e6095dcb-dirty ready`18:15
@fungicide:matrix.orgwebui is up18:15
@fungicide:matrix.orgshall i trigger reindexing or wait?18:16
@clarkb:matrix.organd diffs are loading for me. Should I go ahead and unpause event queues in openstack and zuul tenants?18:16
@jim:acmegating.comi have a patch to push up, lmk when you want it.18:16
@fungicide:matrix.orgyeah should be fine to unpause18:16
@clarkb:matrix.orgcorvus: let me unpause, then you can push then if that all looks good (replication etc) then we can trigger reindexing18:16
@fungicide:matrix.orgwfm18:16
@clarkb:matrix.orgcorvus: openstack and zuul tenants should be unpaused now18:17
@fungicide:matrix.orgi guess we still need a new longer-term token generated and added to hostvars, i can do that when we're done18:18
-@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [opendev/system-config] 974540: Remove statusbot from #opendev irc https://review.opendev.org/c/opendev/system-config/+/97454018:18
@clarkb:matrix.org974540 replicated to gitea and I can see its matching diffs in gitea and gerrit. Zuul is still working through the event queues18:19
@clarkb:matrix.orgcorvus: I don't see 974540 in openstack's zuul tenant check queue yet18:20
@fungicide:matrix.orgare event queues still burning down?18:21
@clarkb:matrix.orgoh now it says there are 2 events though (it said 0 before but maybe I caught it while it was still shuffling stuff around)18:21
@fungicide:matrix.orgah18:21
@clarkb:matrix.orgit want from 2 -> 6 -> 11 -> 12 so maybe i caught it before it actually started seeing those events18:21
@clarkb:matrix.org* it went from 2 -> 6 -> 11 -> 12 so maybe i caught it before it actually started seeing those events18:22
@jim:acmegating.comit's there now.  looks like there was some stuff going on in the openstack tenant.18:22
@jim:acmegating.com(as in, like, a lot of changes to process and query)18:22
@fungicide:matrix.orgcool, proceeding with reindexing in that case, unless i should wait longer18:22
@clarkb:matrix.orgI think it is probably fine. The gerrit side of things seem to be working (replicating pushes voting)18:23
@fungicide:matrix.org"Reindexer started"18:23
@clarkb:matrix.orgif zuul is struggling then it is likely a side effect of pausing then unpausing which could be normal (things didn't process as they arrived so there is a backlog) or there is a bug and we'd have to fix that in zuul anyway18:23
@clarkb:matrix.orgin this case I think it falls under normal behavior when we stop letting it chew on things as they arrive as everything looks happy now18:24
@jim:acmegating.comi think we would expect recovery to be slow due to lack of caches in gerrit18:24
@clarkb:matrix.orgcorvus: oh yup every api request is going to be a little bit slower too18:25
@fungicide:matrix.orgdoes this look right for 10 years? `sudo docker run --rm --volume=/etc/zuul:/etc/zuul -it quay.io/zuul-ci/zuul-scheduler:latest zuul-admin create-auth-token --auth-config local --user localauth --expires-in=315569260`18:25
@clarkb:matrix.organd jgit cachesi n memory that would be wiped out too18:25
@clarkb:matrix.orgfungi: your seconds number seems slightly larger than my `60 * 60 * 24 * 356 * 10` value which is probably correct ot handle leap years so ya18:26
@jim:acmegating.comhaha i'm over here trying to reverse engineer that with like ... "did he use 365.2425? nope"18:27
@fungicide:matrix.org`units 10years seconds`18:27
@fungicide:matrix.orgi rely on the units utility a lot18:27
@fungicide:matrix.orgbecause: lazy18:28
@jim:acmegating.comi'm going to double check something, 1 sec18:28
@fungicide:matrix.orgnp18:28
@jim:acmegating.comlgtm.  wanted to double check the field really is in seconds (because javascript frequently isn't, but this really is)18:30
@fungicide:matrix.orgcool, thanks18:30
@fungicide:matrix.orgupdated the zuul-scheduler group vars with a token that should expire in a decade now18:34
@clarkb:matrix.orgthanks the next hourly run should apply it18:34
@clarkb:matrix.orgfor some reason autohold-list works on zuul02 maybe that doesn't require auth so its fine18:35
@fungicide:matrix.orgautoholds are also listed in the webui without authentication so that seems expected18:36
@clarkb:matrix.orgyup that must be it then. I was just hoping for an easy test method. Maybe we pause then unpause again or just trust that it will work18:36
@clarkb:matrix.orgoh actually we can create an autohold and then delete it that is non impactful18:36
@fungicide:matrix.orgyeah18:37
@clarkb:matrix.orgreindexing is ~22% done. I also detached from the screen session18:37
@fungicide:matrix.orgi'll leave it up reporting the reindex progress until that finishes18:38
@clarkb:matrix.orgthat was the last chagne in the Gerrit 3.11 upgrade stack. Now we can start looking more closely at the 3.12 upgrade18:39
@clarkb:matrix.orgwhich should get us post quantum algorithm support from MINA for the folks asking about that18:40
@jim:acmegating.com```18:40
year tropicalyear
tropicalyear 365.242198781 day
The time needed for the mean Sun as defined above to increase its
longitude by 360 degrees. Most references defined the tropical year
as the interval between vernal equinoxes, but this is misleading. The
length of the season changes over time because of the eccentricity of
Earth's orbit. The time between vernal equinoxes is approximately
365.24237 days around the year 2000. See "Mathematical Astronomy
Morsels" for more details.
```
@jim:acmegating.com(units source code)18:41
@clarkb:matrix.org`increase its longitude by 360 degrees` this is relative to the sun?18:42
@clarkb:matrix.orgit must be18:42
@clarkb:matrix.orgoh yes that is what `mean Sun` refers to18:43
@jim:acmegating.comyeah there was a whole other paragraph about the mean sun18:43
@jim:acmegating.com```18:43
Astronomical time measurement is a complicated matter. The length of the
true day at a given place can be 21 seconds less than 24 hours or 30 seconds
over 24 hours. The two main reasons for this are the varying speed of
Earth in its elliptical orbit and the fact that the Sun moves on the ecliptic
instead of along the celestial equator. To devise a workable system for time
measurement, Simon Newcomb (1835-1909) used a fictitious "mean Sun".
Consider a first fictitious Sun traveling along the ecliptic at a constant
speed and coinciding with the true Sun at perigee and apogee. Then
considering a second fictitious Sun traveling along the celestial equator at
a constant speed and coinciding with the first fictitious Sun at the
equinoxes. The second fictitious Sun is the "mean Sun". From this equations
can be written out to determine the length of the mean day, and the tropical
year. The length of the second was determined based on the tropical year
from such a calculation and was officially used from 1960-1967 until atomic
clocks replaced astronomical measurements for a standard of time. All of the
values below give the mean time for the specified interval.
See "Mathematical Astronomy Morsels" by Jean Meeus for more details
and a description of how to compute the correction to mean time.
```
@clarkb:matrix.orgorbital mechanics are super complicated, something which playing kerbal space program illustrates really well18:43
@fungicide:matrix.orgyeah, my hat's off to the units community for being total measurement geeks18:54
@fungicide:matrix.orgi rely on it constantly when cooking/baking too, or doing carpentry/cabinetry, not just for computing tasks18:55
-@gerrit:opendev.org- Clark Boylan proposed: [opendev/system-config] 974555: Send statusbot notices to #starlingx-general:opendev.org https://review.opendev.org/c/opendev/system-config/+/97455519:04
@clarkb:matrix.orgfungi: you should double check I got the room name correct in that change19:05
@fungicide:matrix.orgyep, looks right19:08
@fungicide:matrix.org"Reindex changes to version 86 complete"19:18
@fungicide:matrix.orgclosing out the root screen session on review0319:19
@clarkb:matrix.orgI was just going to report that. Three errors as expected19:19
@fungicide:matrix.org"Failed 3/967957 changes" yes19:19
@clarkb:matrix.orgI was able to create a zuul autohold using `zuul-client` on zuul02 then delete it. I think the new token is working19:31
-@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/system-config] 974555: Send statusbot notices to #starlingx-general:opendev.org https://review.opendev.org/c/opendev/system-config/+/97455520:03
@fungicide:matrix.orgstatus has joined that room now20:39
-@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul-jobs] 973689: Validate crc32 checksum for S3 image upload https://review.opendev.org/c/zuul/zuul-jobs/+/97368922:25

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!