| opendevreview | Pierre Riteau proposed opendev/irc-meetings master: Update Git URLs to opendev.org https://review.opendev.org/c/opendev/irc-meetings/+/883108 | 05:21 |
|---|---|---|
| opendevreview | Pierre Riteau proposed opendev/irc-meetings master: Update Git URLs to opendev.org https://review.opendev.org/c/opendev/irc-meetings/+/883108 | 07:18 |
| *** gthiemon1e is now known as gthiemonge | 08:59 | |
| opendevreview | Merged opendev/irc-meetings master: Update Git URLs to opendev.org https://review.opendev.org/c/opendev/irc-meetings/+/883108 | 12:35 |
| clarkb | infra-root after sleeping on it for the weekend I do think that using the skopeo preload hack is a reasonable workaround. Ideally for the short term but any move to podman or something else is going to need more time. We can also build all of our images with buildx instead of regular docker build as buildx seems to understand multiple registry mirrors | 15:28 |
| clarkb | However, if others feel more comfortable without workarounds like that I think we should roll back the quay.io image moves and stick to docker hub until after we convert to something like podman | 15:28 |
| clarkb | let me know what yall think I can write appropriate changes from there | 15:29 |
| fungi | did we arrive at a conclusion for impact to multi-arch images? | 15:36 |
| clarkb | I don't think there is one. It should be mostly orthogonal | 15:36 |
| clarkb | on the image build size we need to use buildx everywhere (multi arch is already using buildx) and on the consumption side we preload with skopeo. Neither will affect multiarch | 15:37 |
| clarkb | converting to !docker may have an impact on multiarch since I don't think we currently have tooling to build images multiarch with buildah. But it is theoretically possible as buildah does support it. Mostly just an effort thing I suspect | 15:38 |
| clarkb | my monday software updates are interrupted by pipewire, pulse, and alsa packages fighting over conflicting versions :/ | 15:40 |
| clarkb | "Make pipewire-alsa conflict with alsa-plugins-pulse as both provide /etc/alsa/conf.d/99-*-default.conf" it would be nice if that came with a suggested path forward | 15:41 |
| clarkb | I guess pipewire is the new thing and I should have it talk to alsa for compat instead of pulse? | 15:42 |
| fungi | yeah, or it can also talk to pa i think | 15:48 |
| fungi | though if you don't have stuff that only works with pa, then sandwiching it in between pw and alsa is probably silly | 15:48 |
| clarkb | I'm going to reboot and see if sound still works :) | 16:07 |
| clarkb | (that should restart all daemons with new configs and make sure it isn't working via some zombie that is currently running) | 16:08 |
| clarkb | seems to be working. I can probably do all my meetings tomorrow :) | 16:14 |
| clarkb | fungi: any preference for rolling forward with quay.io using the buildx and skopeo workarounds or rolling back to docker hub to swap out underlying container tech more completely? | 16:15 |
| fungi | i'm good with not rolling back | 16:16 |
| clarkb | ok cool if corvus doesn't object in the near future I'll work on retrofitting what has already been done to accoodate that plan then udpate the not yet merged changes to match | 16:19 |
| clarkb | one thing that makes this easier is we don't do a lot of integration type testing outside of system-config so we can keep the workarounds mostly centralized | 16:20 |
| clarkb | (buildx enablement will go everywhere though) | 16:20 |
| slittle1_ | Please add me as first core for starlingx-app-harbor-core | 16:29 |
| fungi | slittle1_: done | 16:42 |
| *** Guest1107 is now known as atmark | 17:04 | |
| *** d34dh0r5- is now known as d34dh0r53 | 17:07 | |
| *** dmellado92 is now known as dmellado9 | 20:07 | |
| *** dmellado92 is now known as dmellado9 | 20:16 | |
| clarkb | I've just updated hte meeting agenda. If there is anything missing please get it updated soonish so I can send that out before my day ends | 22:20 |
| fungi | thanks! | 22:21 |
| clarkb | for the backup failures from today they appear to be connection resets. dmesg and uptime and memory use etc all look ok on the backup host so I suspect this was something outside of the server having a sad | 22:27 |
| clarkb | we should just keep an eye on it to see if it persists | 22:27 |
| clarkb | the agenda is sent | 23:25 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!