Wednesday, 2020-07-22

« Tuesday, 2020-07-21 Index Thursday, 2020-07-23 »
*** bpradipt has joined #kata-dev04:09
*** OutBackDingo has joined #kata-dev04:44
*** bpradipt_ has joined #kata-dev05:21
*** bpradipt has quit IRC05:24
*** sgarzare_ has joined #kata-dev06:47
*** fgiudici has joined #kata-dev07:09
kata-irc-bot1<thierry> Quick status update on SoB/WIP checks -- the Zuul jobs now run again but systematically fail, as some tricks we played are not compatible with the new container-based job runner. Hopefully will have a fix in the next 8 hours when the US wakes up... but feel free to accelerate the transition to using GH actions for those tests07:55
kata-irc-bot1<aadam> Going back to this thread, I think Eric's approach of doing one thing at a time makes sense so we can start with cleaning the backlog and maybe in the future change processes based on the needs :slightly_smiling_face:.08:03
*** davidgiluk has joined #kata-dev08:07
*** sgarzare_ has quit IRC09:17
*** sgarzare_ has joined #kata-dev09:17
*** sgarzare_ has quit IRC09:27
*** sgarzare_ has joined #kata-dev09:28
*** pcaruana has quit IRC09:50
*** sgarzare_ has quit IRC09:57
*** sgarzare_ has joined #kata-dev09:58
*** pcaruana has joined #kata-dev10:02
*** sgarzare_ has quit IRC10:27
*** sgarzare has joined #kata-dev10:28
kata-irc-bot1<thierry> amshinde: would be great to merge https://github.com/kata-containers/kata-containers/pull/438 asap -- I can help disabling the required checks if you need. Just ping me when you are ready10:35
fidencio@thierry, Archana is on the West Coast, IIRC, so may be too early for her. And we still need another approval in that PR.10:38
*** pcaruana has quit IRC11:03
kata-irc-bot1<thierry> ok, will keep an eye on it11:35
*** devimc has joined #kata-dev12:27
*** sgarzare has quit IRC12:42
*** pcaruana has joined #kata-dev12:54
*** vgoyal has joined #kata-dev13:08
*** devimc_ has joined #kata-dev13:18
kata-irc-bot1<thierry> PR approved, I just switched the required approvals to the new checks. Blocked by travis-CI, but should merge soon13:19
*** devimc has quit IRC13:20
*** devimc_ is now known as devimc13:23
kata-irc-bot1<thierry> (except travisCI seems to currently fail all tests)13:47
*** sgarzare has joined #kata-dev14:12
*** ailan has joined #kata-dev14:30
kata-irc-bot1<thierry> gabyct: the WIP/SoB checks are no longer required... maybe you can push a new version of https://github.com/kata-containers/kata-containers/pull/433 to refresh the test card15:30
kata-irc-bot1<thierry> (that will allow to merge https://github.com/kata-containers/kata-containers/pull/438 afterwards)15:30
*** sgarzare has quit IRC15:53
*** sgarzare has joined #kata-dev15:54
*** sgarzare has quit IRC16:21
*** sgarzare_ has joined #kata-dev16:21
*** fgiudici has quit IRC17:17
kata-irc-bot1<jose.carlos.venegas.m> @archana.m.shinde @chen.bo could you ack on  https://github.com/kata-containers/packaging/pull/1098. ?17:38
kata-irc-bot1<jose.carlos.venegas.m> https://github.com/kata-containers/runtime/pull/2843 is passing so we are safe to merge17:39
kata-irc-bot1<jose.carlos.venegas.m> ```INFO: Applying patch /tmp/jenkins/workspace/kata-containers-runtime-ubuntu-18-04-virtiofs-PR/go/src/github.com/kata-containers/packaging/kernel/patches//kata-v5.6-april-09-2020.x/0001-net-virtio_vsock-Fix-race-condition-between-bind-and.patch```17:41
*** sgarzare_ has quit IRC17:52
*** sgarzare_ has joined #kata-dev17:52
kata-irc-bot1<chen.bo> @jose.carlos.venegas.m Thanks for updating the PRs. For PR #1098 to packaging, it looks like we are only carrying one custom kernel path, a.k.a fix race condition for virtio_vsock. Does it mean the other custom kernel patches we had are already in the updated kernel (e.g. the patch to enable evged)?18:02
kata-irc-bot1<jose.carlos.venegas.m> @chen.bo I am not sure, this is the only patch that today we have for virtiofs kernel18:41
kata-irc-bot1<jose.carlos.venegas.m> we have not run into any issue today without those18:43
kata-irc-bot1<jose.carlos.venegas.m> I see you added this for v0.318:43
kata-irc-bot1<jose.carlos.venegas.m> in master we dont have use it18:44
kata-irc-bot1<jose.carlos.venegas.m> https://github.com/kata-containers/packaging/tree/master/kernel/patches/virtio-fs-dev.virtio-fs-dev.x18:44
kata-irc-bot1<jose.carlos.venegas.m> @chen.bo https://github.com/cloud-hypervisor/cloud-hypervisor/blob/master/docs/hotplug.md#kernel-support18:47
kata-irc-bot1<jose.carlos.venegas.m> `available in 5.5-rc1 and later`  yep we the kernel is on v5.618:47
kata-irc-bot1<chen.bo> @jose.carlos.venegas.m Great to confirm that. I saw an earlier commits of your PR had those patches, so wanted to double check.18:48
kata-irc-bot1<jose.carlos.venegas.m> yes, thanks for asking I was not aware of it18:49
kata-irc-bot1<jose.carlos.venegas.m> so we are good to merge those patches18:49
kata-irc-bot1<jose.carlos.venegas.m> @archana.m.shinde should remove the WIP and SoB checks  and merge?18:49
kata-irc-bot1<archana.m.shinde> @jose.carlos.venegas.m yes, remove those checks and merge18:51
kata-irc-bot1<archana.m.shinde> I'll add the actions checks today for all our repos18:51
kata-irc-bot1<jose.carlos.venegas.m> sounds good18:52
kata-irc-bot1<jose.carlos.venegas.m> @archana.m.shinde this PR requires one more ack18:53
kata-irc-bot1<jose.carlos.venegas.m> https://github.com/kata-containers/packaging/pull/109818:53
kata-irc-bot1<archana.m.shinde> taking a look18:55
kata-irc-bot1<archana.m.shinde> why do I see `File renamed without changes.`18:55
kata-irc-bot1<jose.carlos.venegas.m> I just moved the dir to a new path18:55
kata-irc-bot1<jose.carlos.venegas.m> so the build script can find patches for the new kernel version/tag18:55
kata-irc-bot1<jose.carlos.venegas.m> it should be the same kernel18:56
kata-irc-bot1<jose.carlos.venegas.m> so not changes needed just rename to keep all as before18:56
kata-irc-bot1<jose.carlos.venegas.m> @archana.m.shinde ^18:56
kata-irc-bot1<archana.m.shinde> ok18:57
kata-irc-bot1<archana.m.shinde> strange git does not show the directory path of the new location18:57
kata-irc-bot1<archana.m.shinde> confusing18:57
kata-irc-bot1<jose.carlos.venegas.m> yeah , I feel the same18:58
kata-irc-bot1<jose.carlos.venegas.m> ```kernel/patches/virtio-fs-dev.virtio-fs-dev.x/0001-net-virtio_vsock-Fix-race-condition-between-bind-and.patch → kernel/patches/kata-v5.6-april-09-2020.x/0001-net-virtio_vsock-Fix-race-condition-between-bind-and.patch```18:58
kata-irc-bot1<jose.carlos.venegas.m> this should be the rename18:58
kata-irc-bot1<archana.m.shinde> yeah18:58
kata-irc-bot1<archana.m.shinde> @jose.carlos.venegas.m Would be good if we could get this merged soon as well : https://github.com/kata-containers/packaging/pull/109718:59
kata-irc-bot1<jose.carlos.venegas.m> yep, this CI problem was a blocker to test18:59
kata-irc-bot1<archana.m.shinde> I have left my comments on that PR18:59
kata-irc-bot1<jose.carlos.venegas.m> let me merge both kernel patches now an trigger18:59
kata-irc-bot1<archana.m.shinde> you will need to disable `vmx-rdseed-exit` as well for the new qemu tag19:00
kata-irc-bot1<archana.m.shinde> since it is based on 5.x19:00
kata-irc-bot1<jose.carlos.venegas.m> oh thanks for the heads-up19:01
kata-irc-bot1<chen.bo> @jose.carlos.venegas.m @archana.m.shinde Yay, the two PRs are landed. We are ready to land another PR from packaging that build clh statically. It looks like we are not running clh CI for the packaging repo. So should I use a dmmy PR from runtime/test to verify this PR to packaging?19:15
*** bpradipt_ has quit IRC19:21
*** davidgiluk has quit IRC19:23
*** sgarzare_ has quit IRC19:33
*** sgarzare_ has joined #kata-dev19:33
kata-irc-bot1<archana.m.shinde> chenb : sure go ahead19:35
kata-irc-bot1<archana.m.shinde> and add a `Depends-on` on the pacakaging PR19:36
kata-irc-bot1<archana.m.shinde> @jose.carlos.venegas.m Can you ack this: https://github.com/kata-containers/tests/pull/272919:36
kata-irc-bot1<chen.bo> @archana.m.shinde ok. Got it. Thanks.19:37
*** sgarzare_ has quit IRC19:46
*** devimc has quit IRC21:13
kata-irc-bot1<fidencio> @jose.carlos.venegas.m, @archana.m.shinde, we don't have cri-o ci enabled for 2.0, right?21:17
kata-irc-bot1<fidencio> We also don't have 1.x with shimv2, do we?21:18
kata-irc-bot1<jose.carlos.venegas.m> no, as far as I know, let me double check21:18
kata-irc-bot1<jose.carlos.venegas.m> for shimv2 there is only a few integration test for containerd only21:19
kata-irc-bot1<jose.carlos.venegas.m> not all k8s stack running with shimv221:19
kata-irc-bot1<fidencio> Can you point me to that, in case you have that handy?21:19
kata-irc-bot1<fidencio> I'd like to ensure that everything working with containerd will be working the same with CRI-O21:20
kata-irc-bot1<fidencio> so, I guess the logical (and easier) setps are: • Enable the same tests we have for shimv2 & containerd, to run with shimv2 & cri-o; • Get shimv2 & crio CI working against 2.021:21
kata-irc-bot1<jose.carlos.venegas.m> For 2.0 this is what it does today21:25
kata-irc-bot1<jose.carlos.venegas.m> https://gist.github.com/jcvenegas/22205111fd41f8d995435bd9779970fd21:25
kata-irc-bot1<jose.carlos.venegas.m> I copied from the jenkins job21:25
kata-irc-bot1<jose.carlos.venegas.m> probably what you want to take a look is21:25
kata-irc-bot1<jose.carlos.venegas.m> ```export CI_JOB="CRI_CONTAINERD_K8S" export SHIMV2_TEST="true"```21:25
kata-irc-bot1<jose.carlos.venegas.m> that is going to hit this case21:26
kata-irc-bot1<jose.carlos.venegas.m> https://github.com/kata-containers/tests/blob/2.0-dev/.ci/jenkins_job_build.sh#L15621:26
kata-irc-bot1<jose.carlos.venegas.m> I have talked with gaby to try to move all that kind of variables to the test jobs to know that the CI does from repositories21:26
kata-irc-bot1<jose.carlos.venegas.m> for cloud-hypervisor in 1.x I already have all as part of the test repository so we can know what is tested21:27
kata-irc-bot1<fidencio> Yeah, that would help the life of the newcomers (as myself ;-))21:27
kata-irc-bot1<jose.carlos.venegas.m> another part of the hidden behaviour from the CI are those env variables21:28
kata-irc-bot1<jose.carlos.venegas.m> https://github.com/kata-containers/tests/blob/100f3855f524d17a342cbf246cc63f2f92f31786/integration/containerd/cri/integration-tests.sh#L2721:29
kata-irc-bot1<fidencio> And we also have those issues blocking a possible CRI-O CI integration • https://github.com/kata-containers/kata-containers/issues/336 • https://github.com/kata-containers/kata-containers/issues/29921:29
kata-irc-bot1<fidencio> The first one I've spent some time, but got stuck into *not* being able to get agent logs at all, which is something I have to revist21:30
kata-irc-bot1<jose.carlos.venegas.m> If that variable is exported this test should run  for containerd21:30
kata-irc-bot1<fidencio> Cool, I'll give this a try Tomorrow. We (as in Red Hat) really need to have at least the same tests passing with crio that we have with containerd21:31
kata-irc-bot1<fidencio> Or, at least, have a well documented list of failures (and the reasons)21:31
kata-irc-bot1<jose.carlos.venegas.m> I see, I have not played with rust agent yet  :(  - not sure about the easy way to get logs21:31
kata-irc-bot1<fidencio> At least that one is turned out to not be cri-o specific, as it was also reproduced with containerd21:32
kata-irc-bot1<jose.carlos.venegas.m> sounds good yeah, at least initially identify why is a good starting point  :slightly_smiling_face:21:32
kata-irc-bot1<fidencio> cool, I will try to focus on those for the next few days Let's see how it goes.21:33
kata-irc-bot1<fidencio> Thanks for the info, Carlos!21:38
*** vgoyal has quit IRC22:10
kata-irc-bot1<archana.m.shinde> Can I get acks on these: https://github.com/pulls?q=is%3Aopen+is%3Apr+author%3Aamshinde+archived%3Afalse+actions23:05
kata-irc-bot1<archana.m.shinde> @eric.ernst @chen.bo @fidencio ^23:06
kata-irc-bot1<eric.ernst> tim-actions will get our token, github token, right?23:07
kata-irc-bot1<eric.ernst> do we want to pin / fork that action?23:07
kata-irc-bot1<eric.ernst> Sai was sending us info on that just yesterday IIRC23:07
kata-irc-bot1<archana.m.shinde> @eric.ernst yes, tim-actions gets a token scoped to the action workflow23:39
kata-irc-bot1<archana.m.shinde> that gives read/write permission to the repo itself and not org23:39
kata-irc-bot1<archana.m.shinde> all our branches are protected by means of needing approval23:40
kata-irc-bot1<archana.m.shinde> I suppose we could fork the action for additional security23:40
kata-irc-bot1<archana.m.shinde> btw, the tim in tim-actions is from ant-financial23:41
« Tuesday, 2020-07-21 Index Thursday, 2020-07-23 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!