| *** sameo has quit IRC | 00:18 | |
| kata-irc-bot | <archana.m.shinde> Thanks @simon.kaegi for the pointer. | 02:07 |
|---|---|---|
| *** vgoyal has quit IRC | 02:14 | |
| kata-irc-bot | <liubin0329> And I also created an issue https://github.com/kata-containers/kata-containers/issues/1030 , if use the CRI interface, it will not rely on upper-runtime like containerd or CRI-O | 02:55 |
| *** dklyle has quit IRC | 07:02 | |
| *** david-lyle has joined #kata-dev | 07:03 | |
| *** ailan has joined #kata-dev | 07:25 | |
| *** sameo has joined #kata-dev | 07:46 | |
| *** david-lyle has quit IRC | 07:54 | |
| *** pcaruana has joined #kata-dev | 08:15 | |
| *** jodh has joined #kata-dev | 08:37 | |
| *** davidgiluk has joined #kata-dev | 09:01 | |
| *** pcaruana has quit IRC | 09:11 | |
| *** ailan has quit IRC | 09:28 | |
| *** ailan has joined #kata-dev | 10:20 | |
| *** devimc has joined #kata-dev | 13:08 | |
| *** fuentess has joined #kata-dev | 14:02 | |
| *** crobinso has joined #kata-dev | 14:40 | |
| kata-irc-bot | <archana.m.shinde> Hi all, I dont see any agenda items for Kata AC | 14:41 |
| kata-irc-bot | <archana.m.shinde> do we want to still meet this week? | 14:41 |
| kata-irc-bot | <fidencio> I added a note saying it was cancelled a few minutes ago. | 14:53 |
| kata-irc-bot | <fidencio> But if folks want to meet, that's fine as well | 14:53 |
| kata-irc-bot | <fidencio> Let's cancel this one and meet next week with some topics to discuss. | 14:54 |
| kata-irc-bot | <archana.m.shinde> ohh, didnt see the note @fidencio | 14:55 |
| kata-irc-bot | <archana.m.shinde> ya that sgtm | 14:55 |
| kata-irc-bot | <archana.m.shinde> @fidencio Can you send out a mail to the mailing list as well? | 14:56 |
| kata-irc-bot | <fidencio> Sure! | 14:56 |
| kata-irc-bot | <archana.m.shinde> thanks | 14:58 |
| kata-irc-bot | <fidencio> Done! | 14:59 |
| *** devimc has quit IRC | 15:23 | |
| *** devimc has joined #kata-dev | 15:23 | |
| *** ailan has quit IRC | 15:31 | |
| *** dklyle has joined #kata-dev | 15:34 | |
| *** jodh has quit IRC | 18:01 | |
| kata-irc-bot | <eric.ernst> :wave: | 18:43 |
| kata-irc-bot | <eric.ernst> devimc , @archana.m.shinde around? | 18:43 |
| kata-irc-bot | <eric.ernst> I am seeing odd behavior on cleanup path, and want a pair of eyes to check for me. | 18:44 |
| kata-irc-bot | <eric.ernst> Any idea why we would try to store sandbox when we are deleting the sandbox? ie: https://github.com/kata-containers/kata-containers/blob/2.0-dev/src/runtime/virtcontainers/sandbox.go#L1554 | 18:46 |
| kata-irc-bot | <eric.ernst> The call to store is problematic, since it is (wasting time) trying to read component status/state for things like the VM, which is no longer running, and already cleaned up on the filesystem. | 18:47 |
| kata-irc-bot | <eric.ernst> In particular, I see error messages for failure to read the VMM Pid file, which no longer exists because we already stopped the VM and cleanup vm path | 18:49 |
| kata-irc-bot | <fidencio> Not saying much, but this is the commit which added that code: https://github.com/kata-containers/kata-containers/commit/b42fde69c04d098ca3ff91ccfba1dcf6dcb203db | 18:51 |
| kata-irc-bot | <fidencio> Maybe that's ring a bell why it's needed | 18:51 |
| kata-irc-bot | <eric.ernst> yea, i git blamed to wei too. | 18:56 |
| devimc | @eric.ernst afaik just in case something "catastrophic" happen and the runtime aborts before actually deleting the sandbox/container | 18:56 |
| devimc | for example `s.agent.disconnect()` can fail or abort | 18:56 |
| kata-irc-bot | <eric.ernst> Not sure how graceful we need to be, but I think we need to look closer. | 18:58 |
| kata-irc-bot | <eric.ernst> best case we're just wasting our cycles on teh delete path. | 18:58 |
| kata-irc-bot | <eric.ernst> But, the actual case is we're illogical and throwing errors when we're on the 'correct' expected path. | 18:58 |
| kata-irc-bot | <eric.ernst> I'll open an issue so we can review closer | 19:00 |
| kata-irc-bot | <fidencio> The reasoning (persisting) seem reasonable | 19:06 |
| kata-irc-bot | <eric.ernst> devimc: fyi: https://github.com/kata-containers/kata-containers/issues/1179 | 19:12 |
| devimc | @eric.ernst thx | 19:15 |
| devimc | btw @eric.ernst @fidencio I have enabled github's discussions section | 19:22 |
| devimc | ptal https://github.com/kata-containers/kata-containers/discussions | 19:22 |
| kata-irc-bot | <eric.ernst> oh interesting. | 19:22 |
| kata-irc-bot | <eric.ernst> Haven't seen that used before. One more place to communicate :) :) | 19:22 |
| fidencio | devimc: I like it! | 19:23 |
| kata-irc-bot | <eric.ernst> that cleanup flow isn't perfect as is. I'm not really sure the value we get in calling that persist per container as well, at least in the context of deleting an already stopped sandbox. | 19:25 |
| kata-irc-bot | <eric.ernst> esp. in shim-v2; where AFAIU we don't really utilize the persist as well (unless something like migrating?) | 19:25 |
| kata-irc-bot | <eric.ernst> Yeah.. the persist.json should be deleted shortly after as well... | 19:29 |
| fidencio | @eric.ernst, we don't support migration AFAIU, so we could get rid of that before people assume we do support migration | 19:30 |
| kata-irc-bot | <eric.ernst> Sorry, I think it was more for live upgrade, actually. | 19:31 |
| *** davidgiluk has quit IRC | 20:11 | |
| kata-irc-bot | <eric.ernst> devimc fyi: https://github.com/kata-containers/kata-containers/pull/1180 | 20:27 |
| *** devimc has quit IRC | 22:00 | |
| *** crobinso has quit IRC | 23:27 | |
| *** sameo has quit IRC | 23:32 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!