Monday, 2024-01-22

tkajinam--2024-01-22 08:38:39--  (try:19)
tkajinamConnecting to (||:443... failed: Connection timed out.08:58
tkajinamseems this host is not reachable.08:58
fricklertkajinam: ack, it does not even respond to icmp, checking the provider now10:22
tkajinamfrickler, thanks !10:26
fricklerinfra-root: ^^ seems the last successful used node was on 2024-01-08 if I read correctly10:28
fricklerI can't get the console log for the mirror server, trying to reboot it now. which also check nodepool logs for that cloud10:30
fricklernodepool ops on inmotion get server timeouts, same for my reboot request, but the instance still shows up as "rebooting". anyway that cloud seems to be rather borked and need further debugging10:43
fricklerI was wondering how a job even managed to run on that cloud, but it looks like it ran on ovh-bhs1. tkajinam: is that mirror hardcoded into the job configuration somehow? that should not happen10:45
tkajinamfrickler, hmm. let me look into it10:47
fricklerso yes10:48
* frickler kicks TheJulia a bit ;) but also nice that it worked for two years without issues. shows how stable our infra is :-)10:52
tkajinamso should unblock us. I'll dig into the way to replace the upstream mirror by an appropriate local CI mirror, though11:07
* tkajinam is looking into shell hacks done in puppet11:07
tkajinamok so should replace the "bad" hard-coded mirror host13:12
priteauHello. Anyone know if I can end an IRC meeting if I am not the one who started it?14:45
fungipriteau: yes, you just have to wait at least 60 minutes from the #startmeeting14:46
priteauEasy, thanks fungi14:47
fungithat's the main failsafe for cases where a meeting chair forgets to explicitly end a meeting or loses their internet mid-meeting14:48
clarkbI've noticed that in my note taking file and todo list I've been marking down the last week as december not january. Clearly an indication of how that week went16:17
fungifunny, i put it down as a week of mondays16:43
clarkbAs probably my last followup on the haproxy thing 2.9 will not be an lts. 2.10 will be (even releases are LTS I guess).17:43
clarkbI guess that also explains why 2.9 got so many changes that seem to have disrupted things17:43
clarkband reading more about their release strategy I think the LTS tag is correct for us17:44
clarkbthat said they are only supposed to break users when they go to the 3.x series. should only add features17:44
*** dhill is now known as Guest1494417:46
clarkbdid anyone look to see what errors the inmotion servers are getting? Is it the no valid host found error? if so we may need to clean up the placement allocatiosn again18:02
fricklerfrom what I saw in nodepool logs and also my trying to do openstack commands against the mirror host, these were really API timeouts, apache waiting too long for a response from nova18:07
fricklerserver list/show did work fine though, if I were to guess some issue with rabbitmq blocking RPC requests18:07
clarkback so probably something new to us18:09
clarkbpoking at gitea12 with find I believe the repo archive cleanup ran properly over the weekend18:31
clarkbthe only things in the repo-archive/ dir older than 3 days are directories (which don't get claered out by the clenaup only the endpoint files get removed)18:31
clarkbany thoughts on February 7+8 or February 14+15 or February 21+22 for pre ptg meetup? Then scheduling wise I think we can do 2 blocks each day something like 14:00-16:00 UTC and 22:00-00:00 UTC ?18:42
clarkbI think any of those days and times will work for me. I'm happy to defer to others if any are particularly good or bad18:44
clarkbthen I plan to sign us up for the PTG proper on a more "as you're able" basis and maybe treat that as more user facing rather than admin dev time18:44
fungiall of that sounds fine to me. i don't have any significant conflicts with those dates/times18:51
fungilooks like openstack feature freeze is upon us:
fricklerrebase of a stack in nova with ... 46 patches?
clarkbthats quite the stack22:13
clarkbI've taken a pass at updating the meeting agenda22:23
clarkbanything else to add?22:23
opendevreviewJeremy Stanley proposed opendev/system-config master: Switch from legacy to new style keycloak container

Generated by 2.17.3 by Marius Gedminas - find it at!