| *** LinuxMe has quit IRC | 00:09 | |
| *** sameo has quit IRC | 00:29 | |
| kata-irc-bot | <xu> They work on this for years, and had some cool C++ projects. | 00:57 |
|---|---|---|
| *** fuentess has quit IRC | 03:11 | |
| *** dims has quit IRC | 04:47 | |
| *** dims has joined #kata-dev | 04:48 | |
| *** dims has quit IRC | 04:56 | |
| *** jodh has joined #kata-dev | 07:50 | |
| *** sameo has joined #kata-dev | 08:22 | |
| *** davidgiluk has joined #kata-dev | 09:03 | |
| *** gwhaley has joined #kata-dev | 09:29 | |
| *** sameo has quit IRC | 10:10 | |
| *** sameo has joined #kata-dev | 10:23 | |
| *** lpetrut has joined #kata-dev | 10:26 | |
| *** lpetrut has quit IRC | 11:26 | |
| *** dims has joined #kata-dev | 13:23 | |
| kata-irc-bot | <graham.whaley> hi @thierry @sebastien.boeuf - wrt pullapprove, shall we start collecting requirements (what do we currently get from pullapprove), and alternative solutions (what are the options, what do they get us from the requirements). Either I guess in a github Issue or maybe a launchpad? | 13:31 |
| kata-irc-bot | <graham.whaley> We already have a couple of options we can add - our own checkcommits tool, 'probot' (on github), mergify.io, and githubs own ack tools for instance (yeah, ttx and I have discussed/looked into this a touch already, but not extensively searched or compared) | 13:32 |
| kata-irc-bot | <sebastien.boeuf> @graham.whaley yes we can track this through a github issue if we feel this is gonna be easier. I don't mind the mailing list either | 13:34 |
| kata-irc-bot | <sebastien.boeuf> but you summarized it well, the first step is to ensure we will list what we need | 13:35 |
| kata-irc-bot | <sebastien.boeuf> from what I can think of, we use it to parse LGTM and +1, in combination with the list of approvers/reviewers | 13:36 |
| kata-irc-bot | <sebastien.boeuf> are we using it for `/test` ? | 13:36 |
| kata-irc-bot | <graham.whaley> '/test' is done on jenkins, so nope, not pullapprove | 13:37 |
| kata-irc-bot | <sebastien.boeuf> cool | 13:37 |
| kata-irc-bot | <sebastien.boeuf> thus what we need is very limited | 13:37 |
| kata-irc-bot | <graham.whaley> I'll open an Issue for it and gather what we know already.. | 13:37 |
| kata-irc-bot | <sebastien.boeuf> I'd be surprised if Github didn't have another alternative | 13:37 |
| kata-irc-bot | <sebastien.boeuf> thx! | 13:38 |
| kata-irc-bot | <graham.whaley> github itself will do the ack bit, but only a count of group member acks, not separate group counts iirc - that'll be in the Issue with links ;) so, I think that gets us 80% of that | 13:38 |
| kata-irc-bot | <sebastien.boeuf> oh yes sure, but by alternative, I meant I would be surprised if no other opensource project compatible with Github was not already out there to replace pullapprove, especially when the level of feature we need is so reduced | 13:48 |
| kata-irc-bot | <sebastien.boeuf> worst case, we create a Github plugin I guess | 13:48 |
| kata-irc-bot | <thierry> ++ | 13:58 |
| kata-irc-bot | <claire> If you plan to attend the weekly AC meetings, please take this poll *today* to help influence the scheduled call time moving forward: https://framadate.org/katacontainers2019ac Thanks! | 14:11 |
| *** fuentess has joined #kata-dev | 14:35 | |
| *** eernst has joined #kata-dev | 14:43 | |
| *** eernst has quit IRC | 15:05 | |
| *** eernst has joined #kata-dev | 15:14 | |
| *** eernst has quit IRC | 15:50 | |
| *** eernst has joined #kata-dev | 15:51 | |
| *** sameo has quit IRC | 15:54 | |
| *** LinuxMe has joined #kata-dev | 16:02 | |
| *** LinuxMe has quit IRC | 16:47 | |
| *** LinuxMe has joined #kata-dev | 16:47 | |
| *** gwhaley has quit IRC | 18:00 | |
| *** jodh has quit IRC | 18:00 | |
| *** sameo has joined #kata-dev | 18:03 | |
| *** lsm5 has quit IRC | 19:03 | |
| kata-irc-bot | <eric.ernst> @salvador.fuentes i'm seeing a good amount of CI failures for kernel right now | 19:19 |
| kata-irc-bot | <eric.ernst> example: https://github.com/kata-containers/runtime/pull/1010 | 19:20 |
| kata-irc-bot | <salvador.fuentes> checking... | 19:24 |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m any idea of this error? ```Kernel: arch/x86/boot/bzImage is ready (#1) ERROR: path to kernel does not exist, use build-kernel.sh setup``` | 19:25 |
| *** sameo has quit IRC | 19:31 | |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes looking | 19:52 |
| kata-irc-bot | <salvador.fuentes> @eric.ernst, @jose.carlos.venegas.m, in addition, the new 4.19.10 kernel needs flex to be built. please take a look at https://github.com/kata-containers/tests/pull/1032 | 19:55 |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m seems like the `build-kernel.sh` script does not take the latest values of the kernel version in a PR. | 19:57 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes but that is not the reason it is failing ? | 19:57 |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m that is a new failure :O | 19:57 |
| kata-irc-bot | <eric.ernst> keren lversion bump in runtime's verson file merged this am | 19:57 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes what values is getting? | 19:58 |
| kata-irc-bot | <salvador.fuentes> the ones from master, e.g. checking the logs from: https://github.com/kata-containers/runtime/pull/1030 | 19:59 |
| kata-irc-bot | <salvador.fuentes> ``` | 20:00 |
| kata-irc-bot | <salvador.fuentes> ``` INFO: Current Kernel version 4.19.10.21 INFO: Get packaged kernel version INFO: Packaged Kernel version 4.14.67.18 INFO: Install kernel from sources INFO: kernel path does not exist, will download kernel % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- | 20:00 |
| kata-irc-bot | --:--:-- 0 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0 100 272k 100 272k 0 0 1630k 0 --:--:-- --:--:-- --:--:-- 1622k INFO: Download kernel version 4.14.67 INFO: Download kernel``` | 20:00 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes ok, so the kernel was not tested at all in the @sebastien.boeuf PR | 20:03 |
| kata-irc-bot | <salvador.fuentes> yeap :( | 20:03 |
| kata-irc-bot | <jose.carlos.venegas.m> need to update: https://github.com/kata-containers/packaging/bloab/master/scripts/lib.sh#L41 | 20:03 |
| *** davidgiluk has quit IRC | 20:04 | |
| kata-irc-bot | <jose.carlos.venegas.m> but now that is merged is being tested, right? | 20:04 |
| kata-irc-bot | <salvador.fuentes> yes, but now it says that it needs flex package | 20:04 |
| kata-irc-bot | <jose.carlos.venegas.m> but is failing due to the error you are fixing about flex? | 20:04 |
| kata-irc-bot | <salvador.fuentes> right | 20:04 |
| kata-irc-bot | <jose.carlos.venegas.m> ok, so after merge your PR we fix master CI | 20:04 |
| kata-irc-bot | <jose.carlos.venegas.m> And we have pending a validation to not take all the time master | 20:05 |
| kata-irc-bot | <jose.carlos.venegas.m> correct ? | 20:05 |
| kata-irc-bot | <salvador.fuentes> it should be fixed, right, about the other error that @eric.ernst reported, not sure if that is the same root cause | 20:05 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes ok, so not reproducible in all PRs | 20:06 |
| kata-irc-bot | <jose.carlos.venegas.m> the rest of PRs are failing with the issue you are fixing | 20:07 |
| kata-irc-bot | <eric.ernst> ew | 20:11 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst I can not reproduce the error you see | 21:16 |
| kata-irc-bot | <jose.carlos.venegas.m> but seems that fuentess PR get all working again | 21:17 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes @eric.ernst https://github.com/kata-containers/packaging/pull/300 | 21:32 |
| kata-irc-bot | <salvador.fuentes> thanks @jose.carlos.venegas.m, @eric.ernst we have 2 issues right now: there is no arm config for this new kernel, and it seems that there is a performance regression using this kernel | 21:39 |
| kata-irc-bot | <jose.carlos.venegas.m> Just found that we need to re-revert the PR https://github.com/kata-containers/packaging/pull/295/files | 21:42 |
| *** jugs has quit IRC | 21:42 | |
| *** jugs has joined #kata-dev | 21:43 | |
| kata-irc-bot | <jose.carlos.venegas.m> For the kernel regression, well need to open an issue and find a way to reduce, probably ask graham tweak to not fail all the time now | 21:44 |
| *** LinuxMe has quit IRC | 21:44 | |
| kata-irc-bot | <eric.ernst> what's the regression on? | 21:47 |
| kata-irc-bot | <eric.ernst> @jose.carlos.venegas.m what's the regression? | 21:49 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst *F* | memory-footprint | 85.0% | 128.1% | 115.0% | 30.0% | 128.1% | 128.1% | 0.0% | 0.0% | 1 | 21:51 |
| kata-irc-bot | <eric.ernst> i see it | 21:51 |
| kata-irc-bot | <jose.carlos.venegas.m> and seems that boott ime did not worked | 21:51 |
| kata-irc-bot | <jose.carlos.venegas.m> *F* | boot-times | Failed to load JSON | exit status 2 | | | | | | | | | 21:51 |
| kata-irc-bot | <eric.ernst> yeah. | 21:52 |
| kata-irc-bot | <jose.carlos.venegas.m> So unless we want to first fix the kernel, I suggest fix the CI | 21:52 |
| kata-irc-bot | <jose.carlos.venegas.m> then will take longer to jump to a new kernel | 21:52 |
| kata-irc-bot | <jose.carlos.venegas.m> if jump to a new kernel is not priority then revert | 21:53 |
| kata-irc-bot | <eric.ernst> since ARM doesn't have config yet either, I think fixing the CI and waiting to update makes most sense. | 21:53 |
| kata-irc-bot | <eric.ernst> I think we should have a better process in place for kernel update going forward, this one has been a CF | 21:54 |
| kata-irc-bot | <jose.carlos.venegas.m> https://github.com/kata-containers/packaging/pull/302 | 21:54 |
| kata-irc-bot | <eric.ernst> We should take a look at the resulting config. and maybe when updating to oldnoconfig instead of olddefconfig | 21:54 |
| kata-irc-bot | <jose.carlos.venegas.m> now we have config | 21:55 |
| kata-irc-bot | <eric.ernst> maybe just adding instead of revert revert? | 21:55 |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m I think that the one that we need to revert is https://github.com/kata-containers/runtime/pull/1030 | 21:55 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes if you want to add a depends-on | 21:56 |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m @eric.ernst afaik, there were issues with that config last week, right? that is why it was reverted | 21:56 |
| kata-irc-bot | <jose.carlos.venegas.m> yeah that may work to | 21:56 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes 1. The CI was not testing the PRs on packaging | 21:58 |
| kata-irc-bot | <jose.carlos.venegas.m> 2. The 4.19 arm changes were reverted | 21:58 |
| kata-irc-bot | <jose.carlos.venegas.m> 3. The kernel was updated today - but not testing the PRs from master | 21:58 |
| kata-irc-bot | <jose.carlos.venegas.m> 1. Should be fixed now from https://github.com/kata-containers/tests/pull/1025 | 21:59 |
| kata-irc-bot | <jose.carlos.venegas.m> 3. It is fixed by https://github.com/kata-containers/packaging/pull/300 | 22:01 |
| kata-irc-bot | <eric.ernst> If we can figure out #2 then I'd be happier moving forward, not backwards. | 22:05 |
| kata-irc-bot | <eric.ernst> And we should make sure that 'kernel upgrade' process is straight forward for next time around. We should be updating this much more frequently, so want to make sure its painless. | 22:05 |
| kata-irc-bot | <jose.carlos.venegas.m> 2. It is fixed by https://github.com/kata-containers/packaging/pull/302 | 22:07 |
| kata-irc-bot | <salvador.fuentes> @eric.ernst we would still have the performance degradation, so I'll prefer to get back to 4.14.67, then we can merge the PR 302 from packaging to fix the process | 22:08 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst seems that is better to revert sebastien PR and have all in place | 22:08 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes will revert | 22:08 |
| kata-irc-bot | <eric.ernst> 1032 should go in, though | 22:08 |
| kata-irc-bot | <jose.carlos.venegas.m> then have all needed merged | 22:09 |
| kata-irc-bot | <jose.carlos.venegas.m> 1032? | 22:09 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst | 22:10 |
| kata-irc-bot | <eric.ernst> https://github.com/kata-containers/tests/pull/1032 | 22:10 |
| kata-irc-bot | <eric.ernst> this should be added regardless. | 22:10 |
| kata-irc-bot | <jose.carlos.venegas.m> yeah | 22:10 |
| kata-irc-bot | <jose.carlos.venegas.m> lets merge it | 22:10 |
| kata-irc-bot | <eric.ernst> otherwise this sounds pragmatic - need to unbreak CI and assess the footprint changes. | 22:11 |
| kata-irc-bot | <eric.ernst> I would like to know how the 4.19 configs were created, too. | 22:11 |
| kata-irc-bot | <salvador.fuentes> @eric.ernst @jose.carlos.venegas.m https://github.com/kata-containers/runtime/pull/1101 | 22:13 |
| kata-irc-bot | <eric.ernst> +1 | 22:13 |
| kata-irc-bot | <eric.ernst> though it shames me to still be on 4.14. | 22:13 |
| kata-irc-bot | <eric.ernst> let's do that assessment quickly. | 22:14 |
| kata-irc-bot | <jose.carlos.venegas.m> lets merge this first https://github.com/kata-containers/packaging/pull/300 | 22:17 |
| kata-irc-bot | <jose.carlos.venegas.m> and then test the other PR to verify that now the kernel is tested properly | 22:17 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes thx | 22:17 |
| kata-irc-bot | <jose.carlos.venegas.m> @salvador.fuentes nvm seems that packaging testing PR is not working we expect then that both fail metrics and arm | 22:24 |
| kata-irc-bot | <eric.ernst> @jose.carlos.venegas.m, @salvador.fuentes - please let me know where we end up with the kernel support. I just want to make sure we have someone tracking the changes. | 22:34 |
| kata-irc-bot | <eric.ernst> and review of the 4.19 (should we do .20?) config / size delta. | 22:35 |
| kata-irc-bot | <jose.carlos.venegas.m> working on that | 22:36 |
| kata-irc-bot | <eric.ernst> thx Carlos | 22:36 |
| kata-irc-bot | <jose.carlos.venegas.m> this is crazy there were a lot of issues that hit us at the same time | 22:37 |
| kata-irc-bot | <jose.carlos.venegas.m> but most of that is not proper testing I think :S | 22:37 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!