| *** egernst has quit IRC | 00:14 | |
| *** egernst has joined #kata-dev | 00:15 | |
| *** egernst has quit IRC | 00:31 | |
| *** egernst has joined #kata-dev | 00:32 | |
| *** egernst has quit IRC | 00:36 | |
| *** egernst has joined #kata-dev | 01:08 | |
| *** snir_ has quit IRC | 01:23 | |
| *** egernst has quit IRC | 01:27 | |
| *** snir_ has joined #kata-dev | 01:28 | |
| *** bpradipt has joined #kata-dev | 04:12 | |
| *** sgarzare has joined #kata-dev | 06:47 | |
| *** jodh has joined #kata-dev | 07:30 | |
| *** davidgiluk has joined #kata-dev | 07:39 | |
| *** sameo has joined #kata-dev | 07:43 | |
| kata-irc-bot1 | <bergwolf> Any special reason that CRI-O need to keep kata shim after a pod dies? It seems to be a waste of node resources. And nice spot, we do need to disconnect the agent when stopping a pod. Thanks! | 07:44 |
|---|---|---|
| *** fgiudici has joined #kata-dev | 08:07 | |
| *** pcaruana has joined #kata-dev | 08:27 | |
| *** Yarboa has quit IRC | 09:08 | |
| *** Yarboa has joined #kata-dev | 09:08 | |
| *** Dweezahr has joined #kata-dev | 10:26 | |
| Dweezahr | Hi, kata works with qemu but not with firecracker, I get error messages in containerd like: Aug 24 12:47:14 devserver containerd[18609]: time="2020-08-24T12:47:14.213952720+02:00" level=error msg="getting vm status failed" ID=fffc22bda7f7468e47eb01cfe614ae28c3f150177aaa432fd57bc5c729dfb559 error="Get \"http://localhost/\": dial unix | 10:48 |
| Dweezahr | /run/vc/firecracker/fffc22bda7f7468e47eb01cfe614ae28/root/run/firecracker.socket: connect: no such file or directory" source=virtcontainers subsystem=firecracker | 10:48 |
| Dweezahr | ID=fffc22bda7f7468e47eb01cfe614ae28c3f150177aaa432fd57bc5c729dfb559 error="Get \"http://localhost/\": dial unix /run/vc/firecracker/fffc22bda7f7468e47eb01cfe614ae28/root/run/firecracker.socket: connect: no such file or directory" source=virtcontainers subsystem=firecracker | 10:48 |
| Dweezahr | this command runs just fine: docker run --runtime=kata-fc -itd --name=oh-sweet-fc alpine sh | 10:51 |
| *** devimc has joined #kata-dev | 12:11 | |
| *** fuentess has joined #kata-dev | 12:28 | |
| *** crobinso has joined #kata-dev | 13:25 | |
| *** crobinso has joined #kata-dev | 13:26 | |
| *** devimc has quit IRC | 13:38 | |
| *** devimc has joined #kata-dev | 13:38 | |
| *** egernst has joined #kata-dev | 14:25 | |
| *** egernst has quit IRC | 14:32 | |
| *** egernst has joined #kata-dev | 14:32 | |
| kata-irc-bot1 | <eric.ernst> Dweezahr - are you using a block based snapshotter / graphdriver for docker? | 14:32 |
| kata-irc-bot1 | <bartsmink> @eric.ernst I use a docker config file daemon.json with devicemapper | 14:49 |
| kata-irc-bot1 | <bartsmink> All the rest I have not touched any config, this is Archlinux, so I think not? | 14:50 |
| *** Dweezahr has quit IRC | 14:53 | |
| egernst | davidgiluk: around? I was looking @ old conversation on boot time: https://github.com/kata-containers/runtime/issues/1917#issuecomment-623910451 | 14:54 |
| egernst | ...*seems* that we'd want to carry this patch -- right? devimc have you tested/looked at this? I haven't checked boot-time recently... | 14:55 |
| kata-irc-bot1 | <eric.ernst> I haven't tested with fc in a bit, Bart. | 14:55 |
| kata-irc-bot1 | <eric.ernst> @bartsmink what version of docker? | 14:55 |
| davidgiluk | egernst: Just give me a sec and I'll have a look | 14:56 |
| egernst | Thank! | 14:57 |
| devimc | egernst, nop, I haven't | 14:57 |
| davidgiluk | ok, so you're asking for a backport of the kick send_pkt into the kata kernel branch? | 15:00 |
| *** fgiudici_ has joined #kata-dev | 15:00 | |
| kata-irc-bot1 | <bartsmink> 19.03.12-ce is the docker version | 15:01 |
| egernst | I can try to backport it -- just seemed like it was a miss that we didn't carry it already in our 5.4 kernel. | 15:02 |
| davidgiluk | yeh I was about to ask which kernel you needed it in | 15:03 |
| *** fgiudici has quit IRC | 15:03 | |
| *** crobinso has quit IRC | 15:03 | |
| egernst | jodh: you around? when's the last time you saw agent tracing working with 1.x? :D | 15:09 |
| egernst | I'm looking @ boot-time, and hard to know where time goes when we're waiting on agent. | 15:10 |
| egernst | boot time is a bit longer than I expected, with ~2.4 seconds spent waiting for the initial gRPC check after starting the VM. | 15:11 |
| jodh | egernst: I am around but snowed under I'm afraid. We have an issue to add agent tracing support to the tracing test, but nobody to work on it atm: https://github.com/kata-containers/tests/issues/1283 ;((( | 15:12 |
| *** dklyle has joined #kata-dev | 15:14 | |
| *** dklyle has quit IRC | 15:17 | |
| jodh | egernst: you might be hitting this possibly? https://github.com/kata-containers/kata-containers/issues/413 | 15:19 |
| *** iamweswilson has quit IRC | 15:26 | |
| *** iamweswilson has joined #kata-dev | 15:26 | |
| kata-irc-bot1 | <simon.kaegi> *gulp* ... there are a number of updates that we need for how we run kata in kernel's later than 5.4. The 5.9 LTS branch is going to get created in the next few days. How crazy would it be to start using that instead of 5.4. | 15:36 |
| *** dklyle has joined #kata-dev | 15:42 | |
| kata-irc-bot1 | <efoster> Ideally the shim would be killed when the pod stopped, but I'm having a hard time figuring out how to do that in CRI-O. I don't see a convenient way to kill the shim without deleting containers, which isn't the correct behavior for pod stop. | 15:49 |
| kata-irc-bot1 | <efoster> I could always just send a signal to the shim process, but I'm concerned that we might miss out on some cleanup. | 15:52 |
| *** sgarzare has quit IRC | 16:06 | |
| kata-irc-bot1 | <eric.ernst> @simon.kaegi are there other config changes you're aware of, or things you're looking forward to wrt 5.4 versus master/soon-to-be-but-not-too-soon LTS? | 16:32 |
| kata-irc-bot1 | <simon.kaegi> As of right now I'm trying to use the same config but that's actually what's motivating me to see if we can move onto a more recent kernel. I'm concerned my config is not optimal. | 16:43 |
| *** crobinso has joined #kata-dev | 17:04 | |
| *** jodh has quit IRC | 17:04 | |
| *** dklyle has quit IRC | 17:06 | |
| *** david-lyle has joined #kata-dev | 17:07 | |
| *** dklyle has joined #kata-dev | 17:14 | |
| *** david-lyle has quit IRC | 17:15 | |
| *** fgiudici_ has quit IRC | 17:18 | |
| *** egernst has quit IRC | 17:33 | |
| *** egernst has joined #kata-dev | 17:34 | |
| *** egernst has quit IRC | 17:38 | |
| *** egernst has joined #kata-dev | 17:49 | |
| kata-irc-bot1 | <eric.ernst> Did you start with the frags that are defined in github.com/kata-containers/packaging? | 17:58 |
| kata-irc-bot1 | <eric.ernst> or with distro's, or? | 17:58 |
| *** pcaruana has quit IRC | 17:58 | |
| kata-irc-bot1 | <eric.ernst> @simon.kaegi | 17:58 |
| kata-irc-bot1 | <eric.ernst> ^ | 17:58 |
| kata-irc-bot1 | <simon.kaegi> Everything we use is based on github.com/kata-containers/packaging | 17:59 |
| kata-irc-bot1 | <simon.kaegi> We're currently using the 1.11.2 branch with a couple of very small tweaks to let us build with recent kernels and qemu versions -- coincidentally today I'm building with kernel 5.8.3 and qemu 5.1.0 to start the process to validate for an internal release | 18:04 |
| *** pcaruana has joined #kata-dev | 18:11 | |
| *** davidgiluk has quit IRC | 18:43 | |
| *** pcaruana has quit IRC | 19:02 | |
| *** egernst has quit IRC | 19:14 | |
| *** pcaruana has joined #kata-dev | 19:15 | |
| *** egernst has joined #kata-dev | 19:19 | |
| *** egernst_ has joined #kata-dev | 19:21 | |
| *** egernst has quit IRC | 19:24 | |
| *** bpradipt has quit IRC | 19:31 | |
| *** egernst_ has quit IRC | 19:48 | |
| *** egernst has joined #kata-dev | 19:48 | |
| *** egernst has quit IRC | 20:02 | |
| *** egernst_ has joined #kata-dev | 20:02 | |
| *** egernst_ has quit IRC | 20:06 | |
| *** egernst has joined #kata-dev | 20:07 | |
| *** egernst has quit IRC | 20:15 | |
| *** egernst_ has joined #kata-dev | 20:15 | |
| *** Yarboa has quit IRC | 20:18 | |
| *** Yarboa has joined #kata-dev | 20:19 | |
| *** devimc has quit IRC | 20:33 | |
| *** devimc has joined #kata-dev | 20:37 | |
| *** devimc has quit IRC | 20:43 | |
| *** devimc has joined #kata-dev | 20:53 | |
| *** devimc has quit IRC | 21:12 | |
| *** fuentess has quit IRC | 21:39 | |
| *** egernst has joined #kata-dev | 22:30 | |
| *** egernst_ has quit IRC | 22:34 | |
| *** crobinso has quit IRC | 23:13 | |
| kata-irc-bot1 | <archana.m.shinde> @simon.kaegi We are planning on a release this week for both stable branches | 23:16 |
| kata-irc-bot1 | <archana.m.shinde> @eric.ernst I heard you are up for it? | 23:17 |
| kata-irc-bot1 | <eric.ernst> Yep. | 23:17 |
| kata-irc-bot1 | <eric.ernst> still looking at what is in flight atm | 23:17 |
| *** egernst has quit IRC | 23:30 | |
| *** sameo has quit IRC | 23:44 | |
| *** egernst has joined #kata-dev | 23:49 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!