18 Oct 2024 |
raitobezarius | (I try to keep all the deployments proper wrt to git state but I do sometimes push hotfixes without pushing to git, please yell at me if I do that too often) | 23:19:49 |
jade_ | yeah, i was scared that there was a loadbearing one of those | 23:20:01 |
jade_ | so i had assumed you had a lock on the machine | 23:20:05 |
jade_ | if whatever is in main is clean, i can go from there | 23:20:16 |
raitobezarius | I lied ; give me a se | 23:21:11 |
raitobezarius | +c | 23:21:13 |
raitobezarius | it's not a lie anymore: https://git.lix.systems/the-distro/infra/pulls/143 | 23:23:00 |
raitobezarius | .oO(really need to fix that "lock machine" concept once and for all) | 23:23:28 |
raitobezarius | apologies | 23:23:45 |
Dusk | is https://buildbot.lix.systems/#/buildrequests/197977 a transient failure? | 23:29:07 |
jade_ | In reply to @bb010g:matrix.org is https://buildbot.lix.systems/#/buildrequests/197977 a transient failure? that's what usually happens if you push your CL | 23:29:37 |
jade_ | did it happen in another condition? | 23:29:44 |
Dusk | In reply to @jade_:matrix.org did it happen in another condition? i'm not totally sure what you mean by "condition", but i don't believe so. tests reliably pass for me and nothing related changed AFAICT. | 23:30:19 |
jade_ | In reply to @bb010g:matrix.org i'm not totally sure what you mean by "condition", but i don't believe so. tests reliably pass for me and nothing related changed AFAICT. like, this happens if you push a new revision of your CL and the old one gets cancelled | 23:30:40 |
Dusk | ah, i see | 23:30:46 |
jade_ | there should be a new CL run for the new revision | 23:30:55 |
Dusk | i haven't pushed since about an hour ago, so would that be an old queued run? | 23:31:26 |
jade_ | possible | 23:31:37 |
Dusk | ah, https://buildbot.lix.systems/#/builders/1/builds/6144 actually reports it as obsoleted. my bad. | 23:32:28 |
jade_ | yeah. to be clear, we are also aware that the buildbot is incredibly slow :/ | 23:32:52 |
raitobezarius | We should probably consider some regular synchronous meeting on Lix infra someday, @jade_:matrix.org: so we can sync on things and I can take care of things synchronously as well | 23:35:53 |
jade_ | yeah | 23:36:00 |
raitobezarius | (e.g. pyroscoping the Buildbot, etc.) | 23:36:12 |
jade_ | yeah | 23:36:18 |
raitobezarius | Will send a meet.dgnum.eu later so we can decide on a date and maybe set this as recurring based on the needs | 23:37:49 |
jade_ | looking at the profiler code (which i didn't notice was kinda abandoned looking till now), we should just pyroscope it | 23:47:20 |
jade_ | there's not a whole lot of reason to need buildbot specific integration for the profiling besides knowing why a particular endpoint is bad | 23:47:42 |
jade_ | but our problems are so egregious that we should find plenty enough from overall profiling | 23:47:58 |
19 Oct 2024 |
jade_ | raitobezarius: reminder on https://gerrit.lix.systems/c/lix/+/1828 | 00:06:25 |
jade_ | going to roll a 2.91.1 finally, i think | 00:34:15 |