Wednesday, 2026-01-07

opendevreviewMerged zuul/zuul-jobs master: Rewrite upload-pypi test-playbook  https://review.opendev.org/c/zuul/zuul-jobs/+/97225200:40
*** darmach3 is now known as darmach12:55
clarkblooks like the honeypot change has been approved15:45
clarkbI'm going to check on intermediate registry pruning shortly as well15:45
clarkbthe smaller backup server needs pruning as well. I can probably get to that if someone doesn't beat me to it15:46
clarkbthe prune log doesn't end in a traceback and the last blob path it looks at has a fff prefix so I think it very likely it managed to iterate through the entire list this time around15:52
clarkbI'm going to recheck the change from mordred which kicked off the debugging of this issue now as the images it depends on should still exist after pruning15:53
clarkbhttps://review.opendev.org/c/opendev/system-config/+/970451 this change to be specific15:53
opendevreviewMichal Nasiadka proposed openstack/project-config master: kolla: Introduce a version of Ironic core two tier structure  https://review.opendev.org/c/openstack/project-config/+/97250815:58
opendevreviewMerged opendev/system-config master: Add modsecurity waf rules to docs.opendev.org  https://review.opendev.org/c/opendev/system-config/+/97067416:09
opendevreviewMichal Nasiadka proposed openstack/project-config master: kolla: Introduce a version of Ironic core two tier structure  https://review.opendev.org/c/openstack/project-config/+/97250816:10
opendevreviewJeremy Stanley proposed opendev/system-config master: Update documentation for Matrix move  https://review.opendev.org/c/opendev/system-config/+/97251016:15
clarkbthe honeypot change records successful deployment16:15
fungiinfra-root: ^ we can merge that in advance, or wait until closer to the cut-over and i'll wip it in the meantime16:16
clarkbI can still reach https://docs.opendev.org/opendev/system-config/latest/ so we haven't broken things terribly16:16
fungiyeah, i'm going to try blocking one of my machines in a moment16:21
fungigetting 'ERROR 404: Not Found.' from the honeypot url and the client is still able to fetch normal page content16:23
fungithe client is 162.242.225.14316:24
clarkbfungi: we don't log requests to the initial honeypot url that add to the table (but maybe we should) but we should log subsequent requests that get denied16:25
fungioh, it's only for the docs vhost, let me try again16:26
clarkbfungi: yes only for docs.opendev.org16:26
clarkb(I wanted to limit the scope for the initial testing)16:26
fungimakes sense, sure16:26
clarkbalso normal apache request logging is in place. The logs I talk about are specific to mod security actions16:27
fungiERROR 403: Forbidden.16:27
opendevreviewMichal Nasiadka proposed openstack/project-config master: kolla: Introduce a version of Ironic core two tier structure  https://review.opendev.org/c/openstack/project-config/+/97250816:27
clarkbfungi: excellent. So I think it is operating as expected. The main question now is whether or not any of the crawlers will abuse robots.txt to fetch things16:28
fungiand now that client gets 403 for any subsequent docs.opendev.org url16:28
clarkband we may have to wait a few days to get data on that16:28
fungiright16:28
fungiso far so good16:28
clarkbfungi: can you make a note to test fetches to docs.opendev.org from that IP address 24 hours from now?16:29
clarkbfungi: in theory the table entries are only good for 24 hours16:29
fungii also (inadvertently) tested triggering it with docs.openstack.org and security.openstack.org and it didn't happen there, so all working as intended16:29
fungiand the blocked client can still get to those16:29
fungiyeah i'll try non-honeypot urls from docs.opendev.org again after 16:30 utc tomorrow16:30
clarkbthanks16:30
clarkbhrm looks like the existing images for that gerrit build recheck may have already been pruned based on logs. So I need to recheck the whole stack to build new data then recheck again in a couple of days? I'll start that process now16:49
clarkb(the latest run of pruning got "Unknown manifest" logged which is what happens when there isn't any manifest data in the location due to a 404)16:50
clarkbbut the pruning pass also indicates it wants to keep the manifest so I think the problem is due to the old bug16:50
clarkbfungi: do we want to proceed with https://review.opendev.org/c/opendev/system-config/+/970919 and https://review.opendev.org/c/opendev/system-config/+/970920 today as well to finish up the project renaming followups? I think the logging change is basically safe. The homedir perms change should be safe but has more potential for unexpected impacts16:55
fungiyeah, would be good to get those out of the way before we forget about them16:56
clarkbmaybe we do the permissions change when we switch gerrit to java 21 just in case there is some podman interaction with bind mount perms changing while the container is running? Though I think it should just act like normal file perms updates and in this case we're giving the running process more access so should be safe16:56
clarkbI approved the log collection update change as I believe it to be completely safe (it makes the one job act like all the others)16:59
clarkbit also had two +2s16:59
opendevreviewMerged opendev/system-config master: Stop collecting manage-projects logs as regular zuul job logs  https://review.opendev.org/c/opendev/system-config/+/97092017:08
fungideploy succeeded17:14
clarkbfungi: I think it was a noop though as the manage-projects job didn't run17:21
clarkbfungi: if we land mnasiadka's kolla-reviewers group addition we'll exercise it though17:22
clarkbmnasiadka: is kolla ready for that?17:22
fungiah yeah17:22
clarkbI'm going to prune the backup server now17:24
mnasiadkaclarkb: sure, that’s only an additonal group17:25
mnasiadkaBut the job that is checking acl normalization is failing17:27
mnasiadkaSo if you need me to update it - then it’s going to be tomorrow17:27
clarkbmnasiadka: ack. I'm not sure its urgent but I can take a look and push an update if the problem has an easy solution17:27
clarkbbackup pruning is in progress now17:27
mnasiadkaclarkb: not urgent, but would be nice to get it in this week17:28
clarkbmnasiadka: did you see JayF's comment? I won't address that and will only fix the acl test job17:30
clarkbI'll defer to you on whether you want to take that suggestion17:30
mnasiadkaClark[m]: we’re fine with the main group staying as is, I can respond in the patch17:31
opendevreviewClark Boylan proposed openstack/project-config master: kolla: Introduce a version of Ironic core two tier structure  https://review.opendev.org/c/openstack/project-config/+/97250817:31
clarkbmnasiadka: I think the diff issue is that the formatting wants the rules in alphabetical order so I swapped the -reviewers rules to come after the -core rules17:31
mnasiadkaAh right, my alphabet flipped on my first day of work this year, I still don’t know what I’m doing - thanks :)17:33
slittle1_I'm trying to update starlingx feature branch f/portable-dc from it's parent branch via merge across all our gits.   A fair number are failing in zuul, even though zuul was happy with the same code when merged to master prior to the Xmas break.   https://review.opendev.org/q/topic:%22merge-master-as-of-20260106T051000Z%22+status:open17:37
slittle1_has anything changed on the back end?17:37
clarkbslittle1_: these look like job specific errors: `sysinv.common.exception.SysinvException: Invalid metadata.yaml: version should be a valid version string.` and `ModuleNotFoundError: No module named 'cgcs_patch'` for example17:38
clarkbpossibly related to the merges themselves if not done correctly (eg using cgcs_patch module which is called somethen else on the target side of the merge)17:39
clarkbslittle1_: I think you need to look at the end result of the merges and see if they are valid. These don't look like zuul problems to me17:40
slittle1_cgcs_patch errors I might understand, others like https://zuul.opendev.org/t/openstack/build/307dc5b5b47c4efe9c967405815ba7d4 baffle me ... zero code delta vs master17:41
clarkb`/home/zuul/src/opendev.org/starlingx/update/patch-scripts/start-scripts/24.09.400/pre-start.sh:42:1: E004 File did not end with a newline` is another17:41
clarkbslittle1_: https://review.opendev.org/c/starlingx/update/+/972405/1//MERGE_LIST this is the merge list for that one which is quite large. Is there really zero code delta?17:42
clarkbbut again those are errors in your choice of tools as applied to the code base. I don't think zuul has done anything wrong there17:42
clarkbit is also a non voting job so it may be the case that those errors have been there all along and just ignored17:43
slittle1_i'll try to sort cgcs_patch ... which did hit master .. just thought they had it resolved17:46
slittle1_I'll see whats left after that.17:46
clarkbbackup pruning is done. Got us back down to 68% usage. Logs are in the typical location and exit code waas 018:30
clarkbfungi: stephenfin looks like setuptools never made that end of october release (last release is from may)19:13
clarkbthat new pep fungi pointed out made me curious19:14
clarkbinfra-root as a reminder 23.253.20.107 is the held etherpad 2.6.0 node and there is a clarkb-test pad on it if we want to manually check https://review.opendev.org/c/opendev/system-config/+/972001 to upgrade to etherpad 2.6.019:15
clarkbas mentioned I couldn't figure out the session transfer feature and wonder if we have to use the other skin for that. But am curious if anyone else figures it out19:15
fungiit might be covered in the api docs if those are being kept up to date at all these days19:20
clarkbI only see api methods for create, delete, get, and list sessions22:07
clarkbdoes make me wonder if this is only half baked at this point. The changelog does hint at bigger changes to come22:07
clarkbfor some reason I thought gitea was upgraded already. Probably beacuse we did a coupel upgrades in quick succession in december https://review.opendev.org/c/opendev/system-config/+/971469 is another one I should clear off my backlog (a minor gitea bugfix update)22:15
clarkbfwiw fungi's test is the only thing that has hit the honeypot so far today22:42
clarkbinfra-root ^ it is the error_log not the access log that you need to check for mod security logs just fyi (the access log should show any 403s too)22:42
fungimaybe the ai scourge has finally come to a screeching halt22:43
clarkbit does seem to be cyclical22:43
clarkbthings will be fine for a time then the next horde arrives22:44

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