| *** openstack has joined #kata-dev | 07:29 | |
| *** ChanServ sets mode: +o openstack | 07:29 | |
| *** sgarzare has joined #kata-dev | 07:49 | |
| *** davidgiluk has joined #kata-dev | 08:07 | |
| kata-irc-bot1 | <bergwolf> @dgibson The plan is to maintain tests in a separate repository as is. | 08:39 |
|---|---|---|
| *** sgarzare_ has joined #kata-dev | 08:53 | |
| *** jodh_ has joined #kata-dev | 08:54 | |
| *** jodh has quit IRC | 08:55 | |
| *** sgarzare has quit IRC | 08:55 | |
| *** sgarzare__ has joined #kata-dev | 09:44 | |
| *** sgarzare_ has quit IRC | 09:46 | |
| *** th0din has quit IRC | 09:59 | |
| *** th0din has joined #kata-dev | 10:00 | |
| *** th0din has quit IRC | 10:11 | |
| *** th0din has joined #kata-dev | 10:17 | |
| kata-irc-bot1 | <dgibson> @bergwolf!?!?!?!?!?! | 11:03 |
| kata-irc-bot1 | <dgibson> @bergwolf the only conceivable reason I can think of for having the tests separate to begin with was to make common stuff across the already split for no good reason repos | 11:04 |
| kata-irc-bot1 | <bergwolf> sorry my bad. turned back to the original issue and saw that we have agreed to merge tests/ci later on. https://github.com/kata-containers/kata-containers/issues/172 | 11:37 |
| kata-irc-bot1 | <bergwolf> I don’t see a timeline for it to happen though | 11:38 |
| *** devimc has joined #kata-dev | 12:08 | |
| *** crobinso has joined #kata-dev | 12:55 | |
| *** devimc has quit IRC | 13:07 | |
| *** devimc has joined #kata-dev | 13:08 | |
| *** pcaruana has quit IRC | 13:35 | |
| *** pcaruana has joined #kata-dev | 14:01 | |
| *** sgarzare__ has quit IRC | 14:28 | |
| *** sgarzare has joined #kata-dev | 14:28 | |
| *** sgarzare_ has joined #kata-dev | 14:31 | |
| *** sgarzare has quit IRC | 14:32 | |
| *** sgarzare_ has quit IRC | 15:11 | |
| *** sgarzare has joined #kata-dev | 15:15 | |
| kata-irc-bot1 | <jose.carlos.venegas.m> @fidencio this is the comment I did for Vivek, just in case he miss the github notification, I appreciate if you could ping him | 15:45 |
| kata-irc-bot1 | <fidencio> @jose.carlos.venegas.m, sure! | 15:47 |
| kata-irc-bot1 | <fidencio> he replied even before I could ping him :slightly_smiling_face: | 15:50 |
| kata-irc-bot1 | <jose.carlos.venegas.m> xD | 15:53 |
| *** jodh_ has quit IRC | 16:20 | |
| *** devimc has quit IRC | 16:34 | |
| kata-irc-bot1 | <fidencio> @eric.ernst @aadam, just did a quick test here, seems that you cannot access a Google Document without a google account. The simple test was: • Create a document from my personal google account; • Try to open the "shareable" link to in a private mode browser; By the moment you do this, you'll be asked to log in or create your brand new google account. | 16:50 |
| kata-irc-bot1 | <aadam> I believe that's solvable, see option #2 in the following link: https://learning.linkedin.com/blog/tech-tips/share-a-google-doc-with-a-non-google-user | 16:55 |
| kata-irc-bot1 | <aadam> I think the issue is still if everyone in China can properly access google docs or not | 16:55 |
| kata-irc-bot1 | <fidencio> And if we have a strong reason to actually move | 16:58 |
| *** fgiudici has quit IRC | 17:03 | |
| *** devimc has joined #kata-dev | 17:35 | |
| *** sgarzare has quit IRC | 17:53 | |
| *** sgarzare has joined #kata-dev | 17:54 | |
| *** sgarzare_ has joined #kata-dev | 18:02 | |
| *** sgarzare has quit IRC | 18:02 | |
| *** sgarzare_ has quit IRC | 18:40 | |
| *** davidgiluk has quit IRC | 19:21 | |
| *** bpradipt has quit IRC | 19:27 | |
| fidencio | devimc: @amshinde, quick question, how much do we rely on gogo/protobuf? | 20:16 |
| fidencio | asking as it may hit kata sooner than later: https://github.com/cri-o/cri-o/issues/3991 | 20:17 |
| devimc | fidencio: go files generator in the agent | 20:20 |
| *** devimc has quit IRC | 20:55 | |
| *** sameo has quit IRC | 21:06 | |
| kata-irc-bot1 | <eric.ernst> Hey there @chen.bo @jose.carlos.venegas.m | 21:13 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @eric.ernst hey o/ | 21:13 |
| kata-irc-bot1 | <eric.ernst> I had the pleasure of testing Kata on a centos based distro, and ran into a lack of glibc when trying to run Cloud Hypervisor | 21:13 |
| kata-irc-bot1 | <eric.ernst> Have you run into this drawback of the static bin? /opt/kata/bin/cloud-hypervisor: /lib64/libc.so.6: version `GLIBC_2.18' not found (required by /opt/kata/bin/cloud-hypervisor) | 21:14 |
| kata-irc-bot1 | <eric.ernst> Do we package a non-static version of CLH as well? | 21:14 |
| kata-irc-bot1 | <jose.carlos.venegas.m> yes | 21:14 |
| kata-irc-bot1 | <eric.ernst> hahah | 21:14 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I did not give it a try | 21:14 |
| kata-irc-bot1 | <jose.carlos.venegas.m> haha | 21:14 |
| kata-irc-bot1 | <eric.ernst> ninja edit yyes to no :slightly_smiling_face: | 21:14 |
| kata-irc-bot1 | <eric.ernst> I was hoping for yes :slightly_smiling_face: | 21:14 |
| kata-irc-bot1 | <jose.carlos.venegas.m> It should be a quick fix not tried, | 21:15 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I can send a PR now | 21:15 |
| kata-irc-bot1 | <eric.ernst> tell me more. | 21:16 |
| kata-irc-bot1 | <jose.carlos.venegas.m> well at leaset for kata-deploy binaries | 21:16 |
| kata-irc-bot1 | <eric.ernst> If the RPM has it all, I'm okay. | 21:16 |
| kata-irc-bot1 | <eric.ernst> I wasn't sure if we packaged CLH. | 21:16 |
| kata-irc-bot1 | <jose.carlos.venegas.m> https://github.com/cloud-hypervisor/cloud-hypervisor/blob/master/scripts/dev_cli.sh | 21:16 |
| kata-irc-bot1 | <jose.carlos.venegas.m> no we dont have packages for clh today | 21:17 |
| kata-irc-bot1 | <jose.carlos.venegas.m> with that script and the flag --libc musl | 21:17 |
| kata-irc-bot1 | <eric.ernst> yes, ok. | 21:17 |
| kata-irc-bot1 | <jose.carlos.venegas.m> we should have a static binary | 21:17 |
| kata-irc-bot1 | <eric.ernst> we don't set that today when creating packages? | 21:18 |
| kata-irc-bot1 | <jose.carlos.venegas.m> packages = static tarball ? | 21:18 |
| kata-irc-bot1 | <jose.carlos.venegas.m> we dont | 21:18 |
| kata-irc-bot1 | <jose.carlos.venegas.m> and packages (rpms, debs) we dont have any for CH | 21:19 |
| kata-irc-bot1 | <eric.ernst> i wonder if we shoudl add it in the RPM. | 21:19 |
| kata-irc-bot1 | <eric.ernst> I meant to say tarball, sorry | 21:19 |
| kata-irc-bot1 | <eric.ernst> Really, we want it in both. | 21:19 |
| kata-irc-bot1 | <eric.ernst> but for static it sounds like --musl is needed. | 21:19 |
| kata-irc-bot1 | <jose.carlos.venegas.m> yep | 21:22 |
| kata-irc-bot1 | <jose.carlos.venegas.m> about have it in both, we have not planned to added until someone really want it | 21:23 |
| kata-irc-bot1 | <jose.carlos.venegas.m> well you really want it haha | 21:23 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I dont not how easy would be to setup rust on a deb, rpm env, with go used a static tarball with go compiler | 21:24 |
| kata-irc-bot1 | <jose.carlos.venegas.m> for rust the way to go should be `rustup` that fetch from internet | 21:25 |
| kata-irc-bot1 | <eric.ernst> ok. | 21:25 |
| kata-irc-bot1 | <eric.ernst> not sure if it makes sense for an RPM from CLH itself. | 21:25 |
| kata-irc-bot1 | <jose.carlos.venegas.m> but I prefer that folks use static tarball, makes our life eaiser | 21:25 |
| kata-irc-bot1 | <eric.ernst> same | 21:25 |
| kata-irc-bot1 | <jose.carlos.venegas.m> we have some blockers on clh CI until we get a better tag name for virtiofs kernel | 21:28 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I can send the --musl after that | 21:28 |
| kata-irc-bot1 | <archana.m.shinde> Hey @fidencio, I hear from @jose.carlos.venegas.m that we are still awaiting tags from Vivek | 21:34 |
| kata-irc-bot1 | <archana.m.shinde> currently the CI is broken due to this | 21:34 |
| kata-irc-bot1 | <archana.m.shinde> both cloud-hypervisor and qemu-virtiofs | 21:34 |
| kata-irc-bot1 | <archana.m.shinde> I dont see Vivek on slack, is there a way you can get to him quicker | 21:35 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @fidencio hey yep, I am not sure if he is still online | 21:36 |
| kata-irc-bot1 | <eric.ernst> @chen.bo @jose.carlos.venegas.m @archana.m.shinde should I expect macvtap doesn't work with CLH? | 21:38 |
| kata-irc-bot1 | <eric.ernst> I see tcfilter works, but macvtap does not. | 21:38 |
| kata-irc-bot1 | <eric.ernst> Any idea? | 21:38 |
| kata-irc-bot1 | <eric.ernst> docker: Error response from daemon: OCI runtime create failed: error: 500 reason: VmBoot(VmBoot(DeviceManager(CreateVirtioNet(OpenTap(TapOpen(ConfigureTap(Os { code: 22, kind: InvalidInput, message: "Invalid argument" }))))))): unknown. | 21:38 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I think do, I think so on early enabling I was not aware of support for macvitap or other advanced networking | 21:39 |
| kata-irc-bot1 | <eric.ernst> I don't necessarily thing its an issue. | 21:40 |
| kata-irc-bot1 | <eric.ernst> I just want to understand why it doesn't work. | 21:40 |
| kata-irc-bot1 | <eric.ernst> Is how we configure the tap device different? What's the limitation. | 21:40 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I dont know :( , it is a good question | 21:41 |
| kata-irc-bot1 | <eric.ernst> i wouldn't try it out if my kernel had tc-filter options :slightly_smiling_face: | 21:41 |
| kata-irc-bot1 | <eric.ernst> I vaguely recall an issue here and just 'moving on' -- would be good to make sure we have it understood. @sebastien.boeuf may be able to help here. | 21:41 |
| kata-irc-bot1 | <eric.ernst> or maybe @manohar.r.castelino | 21:41 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @archana.m.shinde agree | 21:42 |
| kata-irc-bot1 | <eric.ernst> what's the error handling part? | 21:43 |
| kata-irc-bot1 | <jose.carlos.venegas.m> have more context on errors | 21:44 |
| kata-irc-bot1 | <jose.carlos.venegas.m> even a chain that explain the error path | 21:44 |
| kata-irc-bot1 | <eric.ernst> coming *from* CLH | 21:45 |
| kata-irc-bot1 | <jose.carlos.venegas.m> from the runtime code, that wraps the logic related with it | 21:46 |
| kata-irc-bot1 | <eric.ernst> I see. Is there an open issue for this? | 21:47 |
| kata-irc-bot1 | <eric.ernst> We should do what we can to make CLH first-class. | 21:47 |
| kata-irc-bot1 | <archana.m.shinde> @eric.ernst Agree we need a static tarball that is correctly built for clh | 21:47 |
| kata-irc-bot1 | <eric.ernst> this in this instance --> the better error "handling" | 21:48 |
| kata-irc-bot1 | <archana.m.shinde> @jose.carlos.venegas.m @chen.bo I raised a request for that last week, we should prioritise that | 21:48 |
| kata-irc-bot1 | <archana.m.shinde> I have also seen a person expecting to run clh after installing packages | 21:49 |
| kata-irc-bot1 | <archana.m.shinde> for start lets document that it is not available in packages | 21:49 |
| kata-irc-bot1 | <archana.m.shinde> if its not a big task, its a nice to have to include CLH in pacakges | 21:50 |
| kata-irc-bot1 | <eric.ernst> Cool. Yeah. My initial suffering was around a non gnu libc distro. Was easy enough to work around. | 21:51 |
| kata-irc-bot1 | <eric.ernst> I would rather make sure at runtime the errors are as clear and helpful as feasible. | 21:51 |
| fidencio | @archana.m.shinde, @jose.carlos.venegas.m, I've sent him a message on IRC | 21:53 |
| kata-irc-bot1 | <archana.m.shinde> thanks @fidencio | 21:53 |
| kata-irc-bot1 | <archana.m.shinde> can you provide me his email address? | 21:53 |
| kata-irc-bot1 | <archana.m.shinde> I cant seem to find it | 21:54 |
| fidencio | @archana.m.shinde Vivek Goyal <vgoyal@redhat.com> | 21:54 |
| fidencio | do you guys a tag also on a released kernel or -rc is good enough? | 21:55 |
| fidencio | *need a tag | 21:55 |
| kata-irc-bot1 | <archana.m.shinde> if its not a big hassle to have a stable kernel version, I would much rather have a stable one | 21:57 |
| kata-irc-bot1 | <archana.m.shinde> @fidencio ^ | 21:58 |
| *** vgoyal has joined #kata-dev | 21:59 | |
| fidencio | vgoyal++ :-) | 21:59 |
| vgoyal | fidencio: :-) | 21:59 |
| *** crobinso has quit IRC | 21:59 | |
| fidencio | @archana.m.shinde, do you also need to change / do something on the qemu tag or only on the kernel one? | 22:00 |
| kata-irc-bot1 | <archana.m.shinde> no just kernel one | 22:00 |
| fidencio | oh, perfect then! | 22:00 |
| kata-irc-bot1 | <archana.m.shinde> david did provide a tag for qemu recently | 22:00 |
| fidencio | vgoyal: welcome to this bridge between slack and IRC, which makes things confused for both worlds :-) | 22:01 |
| kata-irc-bot1 | <archana.m.shinde> :wave: @vgoyal | 22:01 |
| kata-irc-bot1 | <jose.carlos.venegas.m> vgoyal: hey \o , just sent a mail :P | 22:01 |
| vgoyal | @jose.carlos.venegas.m, is tag kata-v5.6-april-09-2020 good? | 22:03 |
| vgoyal | I thought mentioning kernel version in tag will help recognize which kernel we are talking about | 22:04 |
| kata-irc-bot1 | <jose.carlos.venegas.m> vgoyal: yep, that would be helpful | 22:04 |
| kata-irc-bot1 | <archana.m.shinde> vgoyal : looks good | 22:04 |
| kata-irc-bot1 | <archana.m.shinde> and thanks! | 22:05 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @fidencio thanks I should have a IRC client we have to I dont see who is connected via IRC | 22:06 |
| kata-irc-bot1 | <jose.carlos.venegas.m> yay, I see a new tag | 22:07 |
| kata-irc-bot1 | Action: jose.carlos.venegas.m goes and create PRs | 22:07 |
| vgoyal | @archana.m.shinde @jose.carlos.venegas.m I created a new tag | 22:08 |
| vgoyal | kata-v5.6-april-09-2020 | 22:08 |
| kata-irc-bot1 | <jose.carlos.venegas.m> thanks :slightly_smiling_face: | 22:08 |
| vgoyal | Check it out and see if it works for you | 22:08 |
| kata-irc-bot1 | <jose.carlos.venegas.m> sure | 22:11 |
| kata-irc-bot1 | <chen.bo> @eric.ernst Can you please share how I can reproduce the error you had w/ `macvtap`? I can look into this and check whether it is a limitation from our clh driver or clh itself. | 22:19 |
| *** vgoyal has quit IRC | 22:20 | |
| kata-irc-bot1 | <eric.ernst> just update the clh tom. | 22:21 |
| kata-irc-bot1 | <eric.ernst> toml* | 22:21 |
| kata-irc-bot1 | <eric.ernst> open /opt/kata/share/defaults/kata-containers/configuration-qemu.toml | 22:21 |
| kata-irc-bot1 | <eric.ernst> that might be in /usr/share/defaults/kata-contaers...... | 22:21 |
| kata-irc-bot1 | <eric.ernst> then just s/tcfilter/macvtap. | 22:22 |
| kata-irc-bot1 | <chen.bo> ok ok. | 22:22 |
| kata-irc-bot1 | <chen.bo> Got it reproduced. Is this an important feature we want to enable to make clh first-class in kata? | 22:26 |
| kata-irc-bot1 | <archana.m.shinde> @chen.bo the default networking mode works with clh, but it will be good to have this feature | 22:29 |
| kata-irc-bot1 | <archana.m.shinde> since we claim support for macvtap | 22:30 |
| kata-irc-bot1 | <chen.bo> @archana.m.shinde ok. At least, let's open an issue to track it now. Shall we put it in kata-runtime or kata-containers? | 22:31 |
| kata-irc-bot1 | <chen.bo> @archana.m.shinde @jose.carlos.venegas.m I also just opened a PR to the packing repo to build a real static binary for clh. Please take a look here: https://github.com/kata-containers/packaging/pull/1099 | 23:03 |
| kata-irc-bot1 | <chen.bo> I am not familiar with how we cache the built cloud-hypervisor. Please let me know if this is enough to work with that. | 23:04 |
| kata-irc-bot1 | <chen.bo> @archana.m.shinde Also, if I got you correctly, you mentioned that you opened an issue for building clh statically. Can you give me a pointer where is that? I only linked the above PR to an issue I opened a while ago (end of April)... | 23:07 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @chen.bo thanks for open the PR, I think the cache scripts should work out of the box | 23:09 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I think we dont have jobs for clh in packaging | 23:09 |
| kata-irc-bot1 | <archana.m.shinde> @chen.bo This is the issue: https://github.com/kata-containers/packaging/issues/1089 | 23:09 |
| kata-irc-bot1 | <chen.bo> @jose.carlos.venegas.m of course. I should have asked you about the cache much earlier so that we would have the static build earlier. | 23:10 |
| kata-irc-bot1 | <chen.bo> Great to know it works out of box. | 23:11 |
| kata-irc-bot1 | <chen.bo> @archana.m.shinde Thanks. Looks like a duplicate of issue: https://github.com/kata-containers/packaging/issues/1033? | 23:11 |
| kata-irc-bot1 | <chen.bo> Do you want me to close 1089 manually w/ 1033 when the PR is landed? | 23:13 |
| kata-irc-bot1 | <jose.carlos.venegas.m> About testing probably you can open an a PR with depens-on to your PR | 23:13 |
| kata-irc-bot1 | <jose.carlos.venegas.m> ah also heads-up the CI for clh is broken | 23:13 |
| kata-irc-bot1 | <jose.carlos.venegas.m> should be fixed until https://github.com/kata-containers/runtime/pull/2843 | 23:13 |
| kata-irc-bot1 | <jose.carlos.venegas.m> @chen.bo ^ | 23:14 |
| kata-irc-bot1 | <chen.bo> @jose.carlos.venegas.m Do you mean I should add depend-on the https://github.com/kata-containers/runtime/pull/2843 ? | 23:14 |
| kata-irc-bot1 | <jose.carlos.venegas.m> I think it will not work this time | 23:15 |
| kata-irc-bot1 | <jose.carlos.venegas.m> should wait until is merged | 23:15 |
| kata-irc-bot1 | <jose.carlos.venegas.m> that PR already depends on another PR from packaging | 23:15 |
| kata-irc-bot1 | <chen.bo> ok. Do you mean the depend-on from packaging repo won't work? | 23:16 |
| kata-irc-bot1 | <chen.bo> I see.. | 23:16 |
| kata-irc-bot1 | <jose.carlos.venegas.m> you can try but not sure if will work :P | 23:16 |
| kata-irc-bot1 | <chen.bo> So we may have to wait for runtime/pull/2843 to be landed before we can land this PR? | 23:16 |
| kata-irc-bot1 | <jose.carlos.venegas.m> yes | 23:16 |
| kata-irc-bot1 | <chen.bo> ok. Let me comment on that in the PR as a record. Thanks a lot. | 23:17 |
| kata-irc-bot1 | <chen.bo> I am assuming the runtime PR is close to be landed? | 23:17 |
| kata-irc-bot1 | <jose.carlos.venegas.m> the CI is running now, I’ll be offline now, I’ll check later today if is OK the new kernel tag | 23:17 |
| kata-irc-bot1 | <jose.carlos.venegas.m> talk you you tomorrow folks :slightly_smiling_face: | 23:18 |
| kata-irc-bot1 | <chen.bo> Right.. The other packaging PR.. | 23:18 |
| kata-irc-bot1 | <chen.bo> Have a good night. | 23:18 |
| kata-irc-bot1 | <chen.bo> Thanks, man. | 23:18 |
| kata-irc-bot1 | <jose.carlos.venegas.m> thanks, you too :slightly_smiling_face: | 23:18 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!