| *** rdale has quit IRC | 01:09 | |
| *** noisecell has joined #baserock | 05:01 | |
| *** gtristan has joined #baserock | 05:41 | |
| *** ctbruce has joined #baserock | 07:38 | |
| *** toscalix has joined #baserock | 07:44 | |
| *** jude_ has joined #baserock | 07:44 | |
| *** jonathanmaw has joined #baserock | 08:03 | |
| *** rdale has joined #baserock | 08:19 | |
| *** juergbi has quit IRC | 08:20 | |
| *** juergbi has joined #baserock | 08:23 | |
| *** adds68 has quit IRC | 08:59 | |
| *** ssam2 has joined #baserock | 08:59 | |
| *** ChanServ sets mode: +v ssam2 | 08:59 | |
| *** adds68 has joined #baserock | 08:59 | |
| *** locallycompact has joined #baserock | 09:29 | |
| ssam2 | the 8 hour CI timeout is causing various troubles | 09:34 |
|---|---|---|
| ssam2 | e.g. https://gitlab.com/baserock/definitions/-/jobs/34738586 | 09:34 |
| ssam2 | ideally we wouldn't have a ci pipeline that could ever take more than 8 hours .... i think pulling the freedesktop SDK all the time slows things down a lot | 09:34 |
| ssam2 | so hopefully https://gitlab.com/baserock/definitions/merge_requests/58 should improve matters | 09:35 |
| gtristan | ssam2, in buildstream we use a cache for that, not the same here ? | 09:35 |
| gtristan | like, gitlab has a cache directory thingy | 09:35 |
| ssam2 | yeah, that's set up | 09:35 |
| ssam2 | but it may not be working | 09:35 |
| gtristan | so it should not download, or at least; that doesnt really cause issue for us | 09:35 |
| gtristan | maybe | 09:36 |
| ssam2 | it definitely seems to fetch the SDK a lot | 09:36 |
| ssam2 | so perhaps it's not working | 09:36 |
| ssam2 | i need to work out how to debug that, though ... | 09:36 |
| ssam2 | ohh | 09:37 |
| ssam2 | seems we are only using the cache for YBD | 09:37 |
| *** jude_ has quit IRC | 09:40 | |
| ssam2 | ok, i've pushed a change for .gitlab-ci.yml to master so it should share the buildstream cache | 09:44 |
| ssam2 | i can't believe i was looking at all these slow builds and didn't check that! it such an obvious thing I assumed we must have done it already :-) | 09:44 |
| *** jude_ has joined #baserock | 09:52 | |
| benbrown_ | kswapd0 is hogging cpu on g.b.o | 10:06 |
| ssam2 | yeah, seems to be out of memory | 10:06 |
| ssam2 | `top` just shows loads of git processes | 10:06 |
| * benbrown_ nods | 10:06 | |
| ssam2 | seems like a couple of `git pack-objects` processes have been running for over an hour | 10:07 |
| ssam2 | those can get pretty resource intensive | 10:07 |
| ssam2 | it maybe an issue from the dozens of buildstream builds that have been running in CI and not caching their sources | 10:08 |
| benbrown_ | Quite possibly :) | 10:10 |
| *** noisecell has quit IRC | 11:10 | |
| gitlab-br-bot | ybd: merge request (benbrown/kbas-silence->master: kbas: fixup and silence tar/touch output) #391 changed state ("opened"): https://gitlab.com/baserock/ybd/merge_requests/391 | 11:49 |
| gitlab-br-bot | ybd: merge request (benbrown/kbas-silence->master: kbas: fixup and silence tar/touch output) #391 changed state ("opened"): https://gitlab.com/baserock/ybd/merge_requests/391 | 12:51 |
| gitlab-br-bot | ybd: merge request (benbrown/kbas-silence->master: kbas: fixup and silence tar/touch output) #391 changed state ("merged"): https://gitlab.com/baserock/ybd/merge_requests/391 | 12:55 |
| *** rdale has quit IRC | 13:20 | |
| *** rdale has joined #baserock | 13:21 | |
| *** ctbruce has quit IRC | 13:49 | |
| *** ctbruce has joined #baserock | 13:49 | |
| *** gtristan has quit IRC | 15:01 | |
| * benbrown_ eagerly awaits the completion of https://gitlab.com/baserock/definitions/-/jobs/34881013 | 15:34 | |
| ssam2 | heh yeah | 15:37 |
| ssam2 | oh, nearly there | 15:37 |
| *** locallycompact has quit IRC | 15:58 | |
| ssam2 | hah, except that it won't cache stuff outside the build directory | 15:58 |
| ssam2 | seems the ybd source caching must have never worked either, since that is in the same place | 15:59 |
| ssam2 | hmm, but i remember something about /cache being some kind of other cache | 16:00 |
| benbrown_ | /cache is the runners "persistent" cache | 16:01 |
| ssam2 | right | 16:01 |
| benbrown_ | our runners being ephemeral by nature | 16:01 |
| ssam2 | right | 16:01 |
| ssam2 | but buildstream was already storing all its sources in there | 16:02 |
| ssam2 | and clearly not getting them back ever | 16:02 |
| ssam2 | this makes no sense to me | 16:03 |
| *** gtristan has joined #baserock | 16:03 | |
| benbrown_ | Have things changed enough in either definitions/buildstream that nothing is cached? Hence why it's seemingly fetching and building *everything* | 16:03 |
| ssam2 | not really | 16:09 |
| ssam2 | it's been pulling the same version of freedesktop SDK since basically forever | 16:09 |
| ssam2 | yet it pulls it every single time | 16:09 |
| *** ssam2 has quit IRC | 16:25 | |
| *** ssam2 has joined #baserock | 16:27 | |
| *** ChanServ sets mode: +v ssam2 | 16:27 | |
| *** jonathanmaw has quit IRC | 17:02 | |
| *** ssam2 has quit IRC | 17:02 | |
| *** toscalix has quit IRC | 17:11 | |
| *** jude_ has quit IRC | 18:08 | |
| *** ctbruce has quit IRC | 18:19 | |
| *** jude_ has joined #baserock | 18:57 | |
| *** jude_ has quit IRC | 21:38 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!