| *** sameo has joined #kata-dev | 00:01 | |
| *** sameo has quit IRC | 01:05 | |
| *** eernst has joined #kata-dev | 03:11 | |
| *** eernst has quit IRC | 03:12 | |
| *** dklyle has joined #kata-dev | 03:23 | |
| *** david-lyle has quit IRC | 03:25 | |
| *** stackedsax has quit IRC | 04:11 | |
| *** stackedsax has joined #kata-dev | 05:48 | |
| *** stackedsax has quit IRC | 08:07 | |
| *** jodh has joined #kata-dev | 08:35 | |
| *** gwhaley has joined #kata-dev | 09:03 | |
| *** sameo has joined #kata-dev | 09:26 | |
| *** sameo has quit IRC | 10:04 | |
| *** sameo has joined #kata-dev | 10:18 | |
| *** zerocoolback has joined #kata-dev | 11:15 | |
| *** zerocoolback has quit IRC | 11:20 | |
| *** sameo has quit IRC | 11:34 | |
| *** gwhaley has quit IRC | 12:03 | |
| *** shrasool has joined #kata-dev | 12:29 | |
| *** shrasool has quit IRC | 12:47 | |
| *** shrasool has joined #kata-dev | 12:52 | |
| *** sameo has joined #kata-dev | 13:19 | |
| *** sameo has quit IRC | 13:19 | |
| *** sameo has joined #kata-dev | 13:20 | |
| *** gwhaley has joined #kata-dev | 13:23 | |
| *** shrasool has quit IRC | 13:31 | |
| *** irclogbot_0 has quit IRC | 14:15 | |
| *** irclogbot_0 has joined #kata-dev | 14:53 | |
| *** shrasool has joined #kata-dev | 15:50 | |
| *** sameo has quit IRC | 15:56 | |
| *** shrasool has quit IRC | 16:02 | |
| *** jodh has quit IRC | 16:07 | |
| *** shrasool has joined #kata-dev | 16:19 | |
| *** shrasool has quit IRC | 17:02 | |
| *** shrasool has joined #kata-dev | 17:03 | |
| *** shrasool has quit IRC | 17:17 | |
| *** shrasool has joined #kata-dev | 17:32 | |
| *** sameo has joined #kata-dev | 17:54 | |
| *** gwhaley has quit IRC | 17:58 | |
| *** shrasool has quit IRC | 18:26 | |
| *** shrasool has joined #kata-dev | 18:27 | |
| *** shrasool has quit IRC | 19:55 | |
| *** stackedsax has joined #kata-dev | 20:17 | |
| *** shrasool has joined #kata-dev | 20:55 | |
| *** sameo has quit IRC | 21:23 | |
| kata-irc-bot | <joshua.reese> is there anyone that could point me in the direction of where sandbox teardown occurs? specifically, i'm running into an issue where CNI teardown is executed after all containers are destroyed, but before the sandbox is destroyed. this is an issue in the particular setup we have where we are leveraging sriov-cni which attempts to decommission the VF, but is unable to locate it as it's still bound to the vfio driver by kata. i'm | 21:30 |
|---|---|---|
| kata-irc-bot | willing to dig in to this, but can also share some logs if someone wants to jump in. any help is appreciated. | 21:30 |
| *** shrasool has quit IRC | 21:38 | |
| kata-irc-bot | <jose.carlos.venegas.m> @jainprasuk1996: you may be interested in https://github.com/kata-containers/runtime/blob/master/virtcontainers/sandbox.go#L1534 | 21:38 |
| kata-irc-bot | <jose.carlos.venegas.m> does it expect do it before the sandbox is destroyed?? | 21:42 |
| kata-irc-bot | <joshua.reese> that's what it's doing right now, i would expect it to wait until the sandbox is destroyed before executing the CNI deletions | 21:43 |
| kata-irc-bot | <jose.carlos.venegas.m> yeah meke sense | 21:46 |
| kata-irc-bot | <jose.carlos.venegas.m> probably a race condition, where CNI deletion is run but VM is still running / turning off | 21:47 |
| kata-irc-bot | <jose.carlos.venegas.m> are you using cri-o + kuberentes ? or only cri-o ? or continerd? | 21:48 |
| kata-irc-bot | <jose.carlos.venegas.m> @archana.m.shinde: could know more about it | 21:49 |
| kata-irc-bot | <joshua.reese> cri-o + kubernetes | 21:50 |
| *** shrasool has joined #kata-dev | 22:02 | |
| kata-irc-bot | <archana.m.shinde> @joshua.reese The network should be teared down while the sandbox is being deleted | 22:14 |
| kata-irc-bot | <archana.m.shinde> iirc the function is removeNetwork | 22:14 |
| kata-irc-bot | <archana.m.shinde> unless something has changed that recently | 22:15 |
| kata-irc-bot | <archana.m.shinde> taking a look | 22:15 |
| kata-irc-bot | <archana.m.shinde> https://github.com/kata-containers/runtime/blob/0911331974fbb82b75d0ecbe21d4eac292cfd5cd/virtcontainers/sandbox.go#L1562 | 22:17 |
| kata-irc-bot | <joshua.reese> @archana.m.shinde from the logs, the cni work completes before the `Stopping VM` message there. also, the kata-runtime and crio logs don't interleave, what i see is a few seconds of logs from kata-runtime when tearing down the containers, then logs from crio (the cni specifically) trying to tear down the network, then logs from kata-runtime tearing down the sandbox | 22:31 |
| *** shrasool has quit IRC | 22:46 | |
| kata-irc-bot | <archana.m.shinde> @joshua.reese we take down the network while the sandbox is being deleted, crio should call CNI cleanup after the sandbox has been deleted | 22:58 |
| kata-irc-bot | <archana.m.shinde> so not sure why you are seeing that order | 22:58 |
| kata-irc-bot | <joshua.reese> so this may be a bug in crio? | 22:58 |
| kata-irc-bot | <archana.m.shinde> not sure, will have to dig a little into this | 22:59 |
| kata-irc-bot | <archana.m.shinde> but have you tried this with containerd? | 22:59 |
| kata-irc-bot | <joshua.reese> i have not | 22:59 |
| kata-irc-bot | <archana.m.shinde> might be worth trying that out to confirm if this is a kata issue | 23:02 |
| kata-irc-bot | <archana.m.shinde> you tried this with sriov-cni plugin? | 23:03 |
| kata-irc-bot | <joshua.reese> correct - via multus | 23:05 |
| kata-irc-bot | <joshua.reese> i'll see if i can get a test with containerd | 23:05 |
| kata-irc-bot | <archana.m.shinde> ok let me know how it goes | 23:05 |
| kata-irc-bot | <joshua.reese> will do, thanks for assisting | 23:06 |
| *** shrasool has joined #kata-dev | 23:29 | |
| *** shrasool has quit IRC | 23:30 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!