| kata-irc-bot | <feng.wang> Questions on the `tests` repository and general development practices. I have been trying to follow the README and run the tests on an Azure instance but the result seems flaky. And the experience is a bit painful. I can’t even get `kubernetes` target pass. Is that expected? Can people share their development environment setup and workflow for Kata project? | 05:52 |
|---|---|---|
| kata-irc-bot | <fidencio> All, can I get one review on https://github.com/kata-containers/kata-containers/pull/2453? I'd like to have it merged & backported for the releases Tomorrow. | 09:29 |
| kata-irc-bot | <samuel.ortiz> @fidencio Merged! | 11:54 |
| kata-irc-bot | <fidencio> thanks a whole lot! | 11:55 |
| kata-irc-bot | <simon.kaegi> yes no problem will do -- we do our own builds so it's easy and can test in our real staging env | 12:59 |
| kata-irc-bot | <gkurz> What's the policy to introduce new patches for QEMU ? Asking because I see that https://github.com/kata-containers/kata-containers/pull/2706 is adding one for QEMU 6.1 in main that doesn't come from upstream QEMU. An equivalent patch for QEMU 5.1 was already added in the past by https://github.com/kata-containers/kata-containers/issues/1841. | 14:02 |
| kata-irc-bot | <jakob.naucke> In one of two cases, yes. I don't think there's a formal policy -- you add whichever patches you need, don't miscredit signoff lines. | 14:55 |
| kata-irc-bot | <jakob.naucke> I suppose we could have co-authored Git commits, but it doesn't look like we have been doing that up to now. | 14:56 |
| kata-irc-bot | <gkurz> What about the correctness of the patch ? It seems it would be a lot safer to have the QEMU community to review it at least | 15:04 |
| kata-irc-bot | <gkurz> and if it fixes an issue, QEMU community might be interested to merge the fix in the first place | 15:04 |
| kata-irc-bot | <jakob.naucke> Hm. About #2706, one is here https://lore.kernel.org/all/20191004155302.4632-1-shameerali.kolothum.thodi@huawei.com/ but I think that kind of got lost? Not super familiar with mailing list workflows though. The other one I can't find anywhere else, but I'm not aware of a formal policy we would have. You can approach Yuanzhe Liu directly if you're concerned. | 15:41 |
| kata-irc-bot | <gkurz> sorry got scheduled... well my concern is that if patches don't reach upstream QEMU, this should be considered as a risk and we should probably not take them either in the first place... unless the QEMU patch is carefully reviewed by knowledgeable people | 16:02 |
| kata-irc-bot | <gkurz> which I might be for some QEMU subsystems but not ARM for example | 16:03 |
| kata-irc-bot | <gkurz> I'll look deeper in these patches tomorrow | 16:04 |
| kata-irc-bot | <gkurz> ttyl | 16:05 |
| kata-irc-bot | <fidencio> devimc, @jose.carlos.venegas.m, @gabriela.cervantes.te, may I bother one of you for a review on (all from @chelsea.e.mafrica): • https://github.com/kata-containers/kata-containers/pull/2662 • https://github.com/kata-containers/kata-containers/pull/2664 • https://github.com/kata-containers/kata-containers/pull/2668 | 16:28 |
| *** devimc_ is now known as devimc | 16:53 | |
| kata-irc-bot | <chelsea.e.mafrica> thanks @fidencio | 17:00 |
| kata-irc-bot | <fidencio> I wish I could take credit by someone's else work, but all kudo go to devimc for quickly reviewing and merging those! :slightly_smiling_face: | 17:02 |
| devimc | lol | 17:03 |
| kata-irc-bot | <chelsea.e.mafrica> and thanks @julio.montes! | 17:04 |
| kata-irc-bot | <jakob.naucke> @wmoschet are the cached builds (tests#3999) necessarily static? | 17:32 |
| kata-irc-bot | <wmoschet> yes @jakob.naucke, the cached builds are static. If you need caching the dynamic ones, can be done too | 18:09 |
| kata-irc-bot | <simon.kaegi> worked perfectly! -- note: I cherry picked your change on top of 2.2.0 as that's what we're using currently | 18:27 |
| kata-irc-bot | <fidencio> Perfect! | 18:27 |
| kata-irc-bot | <fidencio> Then we will get that released as part of 2.2.1 Tomorrow! | 18:27 |
| kata-irc-bot | <simon.kaegi> I tested with both no configmap and an empty configmap, adding a new entry, then deleting the entry or just deleting the configmap -- same tests for secrets too | 18:28 |
| kata-irc-bot | <fidencio> @eric.ernst, can you open the backport for that to the stable-2.2 branch and I can ensure the CIs will be green and we can have it merged later Today / Tomorrow morning. | 18:29 |
| kata-irc-bot | <simon.kaegi> Note: I cherry-picked just @eric.ernst’s latest commit (e.g. 272771d) as there were other incompatible to 2.2.x changes in watcher.rs -- thanks so much here guys | 18:31 |
| kata-irc-bot | <fidencio> All kudos to Eric! :slightly_smiling_face: | 18:31 |
| kata-irc-bot | <eric.ernst> Can you clarify @simon.kaegi | 18:31 |
| kata-irc-bot | <eric.ernst> changes in the stable-2.2 versus main? | 18:32 |
| kata-irc-bot | <eric.ernst> (we need to add tests for this — any chance you want to add your yaml to a unit test skaegi?) | 18:33 |
| kata-irc-bot | <simon.kaegi> Sure this was my setup for my build... ```ARG kata_version=2.2.0 RUN \ mkdir -p /go/src/github.com/kata-containers && \ cd /go/src/github.com/kata-containers && \ git clone --branch $kata_version --depth=1 https://github.com/kata-containers/kata-containers.git && \ cd /go/src/github.com/kata-containers/kata-containers && \ git fetch origin pull/2703/head:watcher_fix && \ git cherry-pick -n 272771d``` | 18:33 |
| kata-irc-bot | <eric.ernst> there were specific changes/issues in watcher.rs ? | 18:34 |
| kata-irc-bot | <simon.kaegi> Just meant the 3 previous changes... https://github.com/kata-containers/kata-containers/commits/main/src/agent/src/watcher.rs ... changed crates/names that wouldn't let me build without fully switching from 2.2.0 to main | 18:36 |
| kata-irc-bot | <simon.kaegi> so... I just cherry picked your fix and re-built and tested | 18:37 |
| kata-irc-bot | <simon.kaegi> Happy to contribute the yaml -- not so sure how to write/contribute the unit test. | 18:42 |
| kata-irc-bot | <eric.ernst> I would just look at any of the others in integration/kubernetes | 18:44 |
| kata-irc-bot | <eric.ernst> I *think* / hope those are stragith forward. | 18:44 |
| kata-irc-bot | <eric.ernst> but, i haven’t dug deep enough so I could be off base; let me know! | 18:45 |
| kata-irc-bot | <simon.kaegi> yep -- I'll take a look -- might take me a bit of time but will do. | 18:46 |
| kata-irc-bot | <fidencio> Those should be straight forward and if you need some help adding that, let me know. Or if you prefer that we do that, please, just open an issue. | 19:26 |
| kata-irc-bot | <simon.kaegi> yep -- i can imitate -- jst am not setup to test locally before i create a pr but... should be able to do it correctly. | 20:10 |
| kata-irc-bot | <gabriela.cervantes.te> @bergwolf any idea if --privileged flag should work with containerd ctr https://github.com/kata-containers/kata-containers/issues/2467? | 21:34 |
| kata-irc-bot | <gabriela.cervantes.te> I saw an issue that you raised long time ago on containerd repository to support this feature but it seems that is not running with kata | 21:35 |
| kata-irc-bot | <gabriela.cervantes.te> https://github.com/containerd/cri/pull/1225 | 21:36 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!