| *** dklyle has joined #kata-dev | 00:24 | |
| *** sameo has joined #kata-dev | 06:20 | |
| *** lpetrut has joined #kata-dev | 07:09 | |
| *** lpetrut has quit IRC | 07:10 | |
| *** lpetrut has joined #kata-dev | 07:11 | |
| *** sgarzare has joined #kata-dev | 07:49 | |
| *** pcaruana has joined #kata-dev | 08:20 | |
| *** jodh has joined #kata-dev | 08:31 | |
| *** sgarzare has quit IRC | 08:34 | |
| *** sgarzare has joined #kata-dev | 08:36 | |
| *** gwhaley has joined #kata-dev | 09:15 | |
| *** gwhaley has quit IRC | 12:02 | |
| *** pcaruana has quit IRC | 12:28 | |
| *** pcaruana has joined #kata-dev | 12:42 | |
| *** fuentess has joined #kata-dev | 13:18 | |
| *** gwhaley has joined #kata-dev | 13:25 | |
| *** devimc has joined #kata-dev | 13:40 | |
| *** canyounot has joined #kata-dev | 13:57 | |
| *** crobinso has joined #kata-dev | 14:43 | |
| *** lpetrut has quit IRC | 15:20 | |
| kata-irc-bot | <eric.ernst> Were you able to launch a spot based VM, @salvador.fuentes ? | 16:17 |
|---|---|---|
| kata-irc-bot | <salvador.fuentes> @eric.ernst let me check, I think I did, but now I am in doubt | 16:24 |
| *** devimc has quit IRC | 16:29 | |
| *** devimc has joined #kata-dev | 16:30 | |
| kata-irc-bot | <salvador.fuentes> @eric.ernst you're right... I couldn't using the kata-containers account | 16:31 |
| kata-irc-bot | <eric.ernst> Okay. I think we need to request a quota for our account, then. | 16:35 |
| kata-irc-bot | <eric.ernst> I think this would reduce our CI cost pressure considerably. Usually ~80% discount. | 16:36 |
| kata-irc-bot | <salvador.fuentes> yeah, would that be Michael Kelly? | 16:36 |
| kata-irc-bot | <eric.ernst> We can crypto mine w all the savings. | 16:36 |
| kata-irc-bot | <salvador.fuentes> hahaha | 16:37 |
| kata-irc-bot | <eric.ernst> I.... think so. | 16:37 |
| kata-irc-bot | <eric.ernst> I don’t have his email anymore. :) | 16:37 |
| kata-irc-bot | <salvador.fuentes> I have it, I'll send him an email... anyway we need that they add spot support to the jenkins plugin... hopefully soon | 16:37 |
| kata-irc-bot | <eric.ernst> Ok. Sure. | 16:40 |
| *** devimc has quit IRC | 17:01 | |
| *** pcaruana has quit IRC | 17:11 | |
| *** devimc has joined #kata-dev | 17:13 | |
| kata-irc-bot | <graham.whaley> hi @bergwolf @julio.montes et. al. - I have some simple questions I wonder if you can help me with - I'm not sure we document the.... | 17:24 |
| kata-irc-bot | <graham.whaley> q1. how can I tell if I am running shimv2 and not shim(v1) ? | 17:24 |
| kata-irc-bot | <graham.whaley> q2. does shimv2 only work with containerd, and not cri-o? | 17:25 |
| kata-irc-bot | <graham.whaley> thx! | 17:25 |
| *** sgarzare has quit IRC | 17:25 | |
| kata-irc-bot | <archana.m.shinde> @graham.whaley shimv2 does work with crio2, but we have a few gaps with restarting containers | 17:25 |
| *** pcaruana has joined #kata-dev | 17:26 | |
| kata-irc-bot | <archana.m.shinde> we dont have CI for it yet | 17:26 |
| kata-irc-bot | <archana.m.shinde> for shimv2, you should be able to see it in the `ps |grep shim` output | 17:26 |
| kata-irc-bot | <graham.whaley> I see 4 'kata-shim's when I am running 2 single container pods - they are called `kata-shim` though. will they be called `shimv2` if in shimv2 mode? | 17:27 |
| kata-irc-bot | <graham.whaley> this is with cri-o as well I think... | 17:28 |
| kata-irc-bot | <graham.whaley> ah, I have two `kata-proxy` running, which means I guess not shimv2? | 17:29 |
| kata-irc-bot | <archana.m.shinde> I believe it is `containerd-shim-kata-v2` | 17:29 |
| kata-irc-bot | <archana.m.shinde> thats the binary | 17:29 |
| kata-irc-bot | <archana.m.shinde> if you are running shimv2, you should not see any `kata-shim` and `kata-proxy` processes | 17:30 |
| kata-irc-bot | <graham.whaley> ack. thanks @archana.m.shinde I'll flip my install to containerd tomorrow hopefully, and see how that looks. | 17:30 |
| kata-irc-bot | <archana.m.shinde> ok | 17:30 |
| kata-irc-bot | <archana.m.shinde> had a brief look at your fluentd doc today | 17:31 |
| kata-irc-bot | <archana.m.shinde> it looks good @graham.whaley | 17:31 |
| kata-irc-bot | <archana.m.shinde> I am glad we had this figured out | 17:31 |
| kata-irc-bot | <graham.whaley> thx. we'll have to see how robust it is. I really should run up a local soak. it is also the driver for me looking at shimv2, to see how/where it's logs go (if anywhere). iirc, we may have just had some doc updates to say you have to enable containerd debug to get them - I want to see if we get anything as is, and they are flippable to json. I'll see what I find. we have gaps - but the runtime/JSON/fluentd is probably the most | 17:32 |
| kata-irc-bot | important big first step. | 17:32 |
| *** jodh has quit IRC | 18:02 | |
| *** gwhaley has quit IRC | 18:02 | |
| *** igordc has joined #kata-dev | 18:36 | |
| *** pcaruana has quit IRC | 19:29 | |
| crobinso | fidencio: you said there's a PR pending to finish kata-runtime cgroupv2 work. got a link? or is it merged? | 19:33 |
| kata-irc-bot | <archana.m.shinde> crobinso: I think this was the one https://github.com/kata-containers/runtime/pull/2441 | 19:34 |
| kata-irc-bot | <archana.m.shinde> Thats been merged | 19:35 |
| kata-irc-bot | <eric.ernst> @julio.montes ^^ | 19:35 |
| fidencio | crobinso: exactly! | 19:36 |
| crobinso | thanks | 19:36 |
| fidencio | crobinso: you'll need that one for runtime (so, cgroups working on the host), and use a new enough agent (1.11.0-alpha0) with an up-to-date libcontainer in order to have it working on the guest side | 19:37 |
| *** igordc has quit IRC | 19:58 | |
| devimc | crobinso: you need to enable sandbox_cgroup_only | 20:07 |
| crobinso | devimc: okay thanks. I'm not attempting to try it right now, but I will sometime this week | 20:08 |
| devimc | crobinso: ok, thx | 20:09 |
| devimc | crobinso: I'm going to close https://github.com/kata-containers/runtime/issues/2107 , feel free to reopen it | 20:10 |
| crobinso | devimc: please do! | 20:10 |
| *** devimc has quit IRC | 20:14 | |
| *** devimc has joined #kata-dev | 20:19 | |
| fidencio | devimc: thanks a ton for working on that! | 20:19 |
| devimc | fidencio: yw :) | 20:20 |
| fidencio | devimc: before replying to your issue on gitlab ... | 21:08 |
| devimc | fidencio: gitlab? | 21:08 |
| fidencio | devimc: github, sorry, I'm used to gitlab more :-) | 21:08 |
| devimc | haha | 21:08 |
| devimc | ok | 21:08 |
| fidencio | devimc: what are the distros officially support to run as the guest? | 21:09 |
| devimc | fidencio: we ship an two images,one with clearlinux and other with alpine | 21:09 |
| fidencio | both support cgroupsv2? | 21:09 |
| devimc | nop | 21:10 |
| fidencio | devimc: I can image forcing that change causing issues | 21:10 |
| fidencio | devimc: on centos7, for instace | 21:10 |
| devimc | even with a recent kernel? | 21:10 |
| fidencio | devimc: I do think so, but I'll double check | 21:11 |
| fidencio | devimc: so, I would *not* enforce the change to cgroupsv2 and leave that up to the distros (with some documentation on how to enable that, in case the distros want to) | 21:11 |
| fidencio | I'll also add this info to the github issue, but does this make sense? | 21:11 |
| devimc | fidencio: yeah, please | 21:12 |
| fidencio | devimc: thanks for the feedback | 21:12 |
| devimc | fidencio: I just raised an issue, because I'm documenting the cgroup v2 implementation | 21:12 |
| devimc | fidencio: this issues will be part of the `limitations` section | 21:16 |
| fidencio | devimc: aha, cool. | 21:17 |
| devimc | *issue | 21:25 |
| fidencio | I dropped my thoughts there, thanks for cc'ing me in the issue | 21:26 |
| *** igordc has joined #kata-dev | 21:30 | |
| devimc | fidencio: thanks for the feedback, | 21:31 |
| fidencio | anytime! | 21:33 |
| fidencio | devimc: JFYI, when using f31+ kernel, we're already taking advantage of cgroupsv2 in the guest | 21:36 |
| fidencio | devimc: I've tested that with 1.11.0-alpha0 agent | 21:36 |
| fidencio | devimc: but on Fedora we're using systemd and system.unified_cgroup_hierarch=1 is the default | 21:37 |
| devimc | fidencio: yeah, I'll include that in the document, thx | 21:38 |
| *** canyounot has quit IRC | 21:48 | |
| *** fuentess has quit IRC | 22:06 | |
| *** devimc has quit IRC | 22:14 | |
| *** crobinso has quit IRC | 23:46 | |
| *** igordc has quit IRC | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!