| *** ailan__ has joined #kata-dev | 05:54 | |
| *** jodh has joined #kata-dev | 06:30 | |
| *** sameo has joined #kata-dev | 06:56 | |
| *** pcaruana has quit IRC | 07:09 | |
| *** tmhoang has quit IRC | 07:09 | |
| *** th0din has quit IRC | 07:09 | |
| *** dklyle has quit IRC | 07:09 | |
| *** sameo has quit IRC | 07:09 | |
| *** amorenoz has quit IRC | 07:09 | |
| *** Jeffrey4l has quit IRC | 07:10 | |
| *** PrinzElvis has quit IRC | 07:10 | |
| *** fidencio has quit IRC | 07:10 | |
| *** jodh has quit IRC | 07:10 | |
| *** sjas has quit IRC | 07:10 | |
| *** sbrivio has quit IRC | 07:10 | |
| *** ChanServ has quit IRC | 07:10 | |
| *** brtknr has quit IRC | 07:10 | |
| *** stackedsax has quit IRC | 07:10 | |
| *** mugsie has quit IRC | 07:10 | |
| *** kata-irc-bot has quit IRC | 07:10 | |
| *** EricAdamsZNC has quit IRC | 07:10 | |
| *** sileht has quit IRC | 07:10 | |
| *** jugs has quit IRC | 07:10 | |
| *** irclogbot_3 has quit IRC | 07:10 | |
| *** noahm has quit IRC | 07:10 | |
| *** iamweswilson has quit IRC | 07:10 | |
| *** jbryce has quit IRC | 07:10 | |
| *** tobberydberg has quit IRC | 07:10 | |
| *** sgarzare has joined #kata-dev | 07:16 | |
| *** sameo has joined #kata-dev | 07:16 | |
| *** jodh has joined #kata-dev | 07:16 | |
| *** Jeffrey4l has joined #kata-dev | 07:16 | |
| *** dklyle has joined #kata-dev | 07:16 | |
| *** sbrivio has joined #kata-dev | 07:16 | |
| *** pcaruana has joined #kata-dev | 07:16 | |
| *** tmhoang has joined #kata-dev | 07:16 | |
| *** th0din has joined #kata-dev | 07:16 | |
| *** irclogbot_3 has joined #kata-dev | 07:16 | |
| *** amorenoz has joined #kata-dev | 07:16 | |
| *** brtknr has joined #kata-dev | 07:16 | |
| *** jugs has joined #kata-dev | 07:16 | |
| *** PrinzElvis has joined #kata-dev | 07:16 | |
| *** fidencio has joined #kata-dev | 07:16 | |
| *** ChanServ has joined #kata-dev | 07:16 | |
| *** stackedsax has joined #kata-dev | 07:16 | |
| *** sileht has joined #kata-dev | 07:16 | |
| *** mugsie has joined #kata-dev | 07:16 | |
| *** sjas has joined #kata-dev | 07:16 | |
| *** tobberydberg has joined #kata-dev | 07:16 | |
| *** kata-irc-bot has joined #kata-dev | 07:16 | |
| *** EricAdamsZNC has joined #kata-dev | 07:16 | |
| *** iamweswilson has joined #kata-dev | 07:16 | |
| *** jbryce has joined #kata-dev | 07:16 | |
| *** tepper.freenode.net sets mode: +o ChanServ | 07:16 | |
| *** davidgiluk has joined #kata-dev | 08:06 | |
| *** gwhaley has joined #kata-dev | 08:08 | |
| *** hashar has joined #kata-dev | 09:08 | |
| *** th0din has quit IRC | 09:53 | |
| *** th0din has joined #kata-dev | 10:05 | |
| *** jugs has quit IRC | 10:12 | |
| *** jugs has joined #kata-dev | 10:12 | |
| *** david-lyle has joined #kata-dev | 10:27 | |
| *** sameo_ has joined #kata-dev | 10:28 | |
| *** dklyle has quit IRC | 10:29 | |
| *** hashar has quit IRC | 10:30 | |
| *** sameo has quit IRC | 10:30 | |
| *** david-lyle has quit IRC | 10:37 | |
| *** pcaruana has quit IRC | 10:49 | |
| *** pcaruana has joined #kata-dev | 10:49 | |
| *** gwhaley has quit IRC | 11:02 | |
| *** pcaruana has quit IRC | 11:02 | |
| *** pcaruana has joined #kata-dev | 11:02 | |
| *** ailan__ has quit IRC | 11:48 | |
| *** sameo_ has quit IRC | 11:50 | |
| *** ailan has joined #kata-dev | 11:51 | |
| *** hashar has joined #kata-dev | 11:51 | |
| *** sameo_ has joined #kata-dev | 12:19 | |
| *** devimc has joined #kata-dev | 12:34 | |
| *** crobinso has joined #kata-dev | 12:39 | |
| *** gwhaley has joined #kata-dev | 12:53 | |
| *** pcaruana has quit IRC | 13:22 | |
| *** pcaruana has joined #kata-dev | 13:34 | |
| *** pcaruana2 has joined #kata-dev | 14:02 | |
| *** th0din has quit IRC | 14:08 | |
| *** th0din has joined #kata-dev | 14:19 | |
| *** devimc has quit IRC | 14:53 | |
| *** pcaruana2 has quit IRC | 14:53 | |
| *** devimc has joined #kata-dev | 14:53 | |
| *** pcaruana has quit IRC | 14:54 | |
| *** vgoyal has joined #kata-dev | 14:55 | |
| *** dklyle has joined #kata-dev | 15:00 | |
| *** jodh has quit IRC | 15:02 | |
| *** pcaruana has joined #kata-dev | 15:06 | |
| *** noahm has joined #kata-dev | 15:37 | |
| * davidgiluk notices -m 2048M,slots=10,maxmem=97387M maxmem sizes like that are going to confuse the heck out of a qemu | 15:45 | |
| davidgiluk | or something | 15:45 |
|---|---|---|
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m, so, would you mind to gimme a quick explanation about what is the "kata-deploy"? | 15:49 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio hey, sure, I is a way to install kata but we do via a container, the container installer requiere access some paths in your system, to reconfigure docker or a kubelet node | 15:51 |
| kata-irc-bot | <jose.carlos.venegas.m> https://github.com/kata-containers/packaging/tree/master/kata-deploy#docker-quick-start | 15:51 |
| kata-irc-bot | <fidencio> How much does it mess with your system? And what's the compatibility with different systems? | 15:52 |
| kata-irc-bot | <jose.carlos.venegas.m> The binaries it installs are the static binaries we publish on every kata release. | 15:53 |
| kata-irc-bot | <jose.carlos.venegas.m> it only do some modifications in the docker or crio service to register the runtime | 15:53 |
| kata-irc-bot | <jose.carlos.venegas.m> binaries are installed under /opt/kata | 15:53 |
| kata-irc-bot | <fidencio> Okay, the dependencies are up to the user to install and ensure those match the minimum requirements | 15:54 |
| kata-irc-bot | <fidencio> It seems like a quite okay way to distribute and give people a way to try the project. | 15:55 |
| kata-irc-bot | <jose.carlos.venegas.m> yes, well we also install qemu under `/opt/kata/bin/` but they are free to overwrite the config using `/etc` config | 15:56 |
| kata-irc-bot | <jose.carlos.venegas.m> actually we install qemu, firecracker, cloud-hypervisor | 15:57 |
| kata-irc-bot | <jose.carlos.venegas.m> so they can try all what kata supports | 15:57 |
| kata-irc-bot | <amorenoz> @julio.montes, I'm looking into vfio passthrough but I'm unable tu have it plugged into a pcie-root-port. Just adding `pcie_root_port = 2` should do the job? | 15:58 |
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m, that's cool. I'd advertise this more than the distros packages. About the distros packages, which distros are officially packaging kata? | 15:59 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio https://github.com/kata-containers/packaging/blob/master/obs-packaging/distros_x86_64 | 16:00 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio for fedora 31, I think gaby tried to updated but at the end was decided that wanated to push more users to use official fedora packages | 16:02 |
| kata-irc-bot | <fidencio> yeah, for Fedora we can rely on the distro packages for sure. | 16:03 |
| kata-irc-bot | <jose.carlos.venegas.m> \o/ | 16:04 |
| kata-irc-bot | <fidencio> I wonder whether more distros would accept the packages if we take the same approach as Fedora and recommend the distros to use their own kernel | 16:04 |
| kata-irc-bot | <fidencio> Usually distributing another kernel, as the ones used by the guest, may go against the distros policy (that's our case with Fedora, that's the current case with RHEL / CentOS as well) | 16:05 |
| sbrivio | devimc, fidencio, about the issue i had yesterday after disabling vsocks, even using podman, i still needed kernel_params = "systemd.unified_cgroup_hierarchy=0" in the configuration | 16:08 |
| sbrivio | now everything works, i thought podman would work with cgroups v2 as well -- or maybe that's not even the issue | 16:08 |
| kata-irc-bot | <fidencio> It does work. I'm running it now with cgroupsv1 | 16:09 |
| kata-irc-bot | <fidencio> *err v2 | 16:09 |
| sbrivio | fidencio, yeah, i'm puzzled, it also used to work with a lighter kernel configuration -- i'll revert to that and try again | 16:09 |
| kata-irc-bot | <fidencio> Do you have `sandbox_cgroup_only=true` in your configuration file? | 16:10 |
| devimc | sbrivio: another workaround is init=/usr/bin/kata-agent | 16:10 |
| sbrivio | fidencio, i do | 16:10 |
| kata-irc-bot | <fidencio> what did you build manually? | 16:10 |
| sbrivio | devimc, oh, nice trick :) well, i don't need it now, but nice to know | 16:10 |
| sbrivio | fidencio, shim, agent, runtime, proxy, kernel | 16:11 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio I think those the main problems to integrate with any distro, kernel and more because distros do not want to have their own kernel just for kata and build and the image is not an option, what you have for post-install step is good to approach but I think the best way go could be just initrd but it lose the optimization for memory because of nvdimm | 16:11 |
| kata-irc-bot | <fidencio> What we do know is just using initrd | 16:11 |
| kata-irc-bot | <jose.carlos.venegas.m> nice and do the initrd at build stage in the package? | 16:12 |
| gwhaley | davidgiluk: good day! - so, wrt the RSS and RANDOMIZE - would a next step be for me to try and look at the actual maps - or, are we not really that bothered? | 16:12 |
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m Generate that either as a post-install (for Fedora) or at boot time, via a systemd service (in case of Fedora CoreOS) | 16:13 |
| kata-irc-bot | <fidencio> the whole thing is that the installation on Fedora CoreOS is done on a "chroot", which has no access to the initrd used in the host | 16:14 |
| kata-irc-bot | <fidencio> Those we delay the generating for when the machine reboots | 16:14 |
| kata-irc-bot | <fidencio> @sbrivio, if it was working with a light weight kernel and now it doesn't, wouldn't be the kernel built the issue? | 16:15 |
| kata-irc-bot | <jose.carlos.venegas.m> I see and you do as a post install step because need to add the host modules | 16:16 |
| sbrivio | fidencio, ideally, yes, but then i had to change the configuration quite a bit (most notably, disabled vsocks), so i'm not sure anymore, i have to go back all the way and retry | 16:16 |
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m, exactly! | 16:16 |
| kata-irc-bot | <fidencio> sbrivio, whenever you feel comfortable to do that, would be quite interesting to have a "debugging guest kernel" kind of blogpost ... explaining all the steps you took | 16:17 |
| kata-irc-bot | <fidencio> sbrivio, as, most likely, we'll need to reproduce your steps to debug other issues in the future | 16:18 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio one more question, how are you loading modules inside the guest? | 16:18 |
| kata-irc-bot | <jose.carlos.venegas.m> using systemd? | 16:19 |
| sbrivio | fidencio, yup, i'm taking notes. and besides, if we want to "support" using a debug-heavy kernel, i guess there's something i need to fix somewhere anyway | 16:19 |
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m, yes, we're relying on systemd for now. I've done some tests in the past using the agent, in order to test a cgroupsv2 patch provided by Julio, and I didn't face any obvious issue there. | 16:20 |
| kata-irc-bot | <fidencio> But for now I'd say we'd like to stick to using systemd, at least till we get more familiar with the project, as systemd is something we know how to debug :slightly_smiling_face: | 16:20 |
| davidgiluk | gwhaley: I'm not sure; the only way I can see is yes look at the maps, I'd *expect* the growth is in the area for the guest RAM - if it's somewhere else then we've got something odd going on | 16:21 |
| kata-irc-bot | <jose.carlos.venegas.m> :slightly_smiling_face: nice, that would be nice to remove any dependency to systemd at some point probably for kata 2.0 | 16:21 |
| kata-irc-bot | <jose.carlos.venegas.m> if we can prove that has same parity and debug level that you requiere | 16:21 |
| kata-irc-bot | <jose.carlos.venegas.m> that would simplify more the project | 16:21 |
| kata-irc-bot | <fidencio> Indeed. That's something we'll add to our list of things to test. There are bigger issues tho, for 2.0. | 16:22 |
| kata-irc-bot | <fidencio> As the agent depending on musl libc to build, for instance | 16:23 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio rust agent ? | 16:23 |
| kata-irc-bot | <fidencio> yep, the rust one! | 16:24 |
| kata-irc-bot | <jose.carlos.venegas.m> oh, and fedora does not use musl at all? | 16:24 |
| kata-irc-bot | <jose.carlos.venegas.m> btw how well is integrated the process with rust projects today fedora packaging guidelines ? | 16:25 |
| kata-irc-bot | <fidencio> nops, that's a hard blocker for us right now (musl) | 16:25 |
| kata-irc-bot | <fidencio> packaging rust projects shouldn't cause any issue for us | 16:26 |
| kata-irc-bot | <fidencio> but I admit it's been a long time I don't check the guidelines | 16:26 |
| kata-irc-bot | <fidencio> https://docs.fedoraproject.org/en-US/packaging-guidelines/Rust/ (valid for Rawhide only) | 16:27 |
| kata-irc-bot | <jose.carlos.venegas.m> thx | 16:28 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio about musl I think that could be an important before official tag kata 2.0, @archana.m.shinde @bergwolf do you think need to be track as part of https://github.com/kata-containers/kata-containers/projects/2 ? | 16:30 |
| kata-irc-bot | <fidencio> I do believe we should, indeed. And I'd say not only for Fedora, but also for other distros. | 16:31 |
| kata-irc-bot | <archana.m.shinde> catching up.. | 16:33 |
| kata-irc-bot | <archana.m.shinde> @jose.carlos.venegas.m Lets add it | 16:33 |
| kata-irc-bot | <archana.m.shinde> I havent followed rust too much, but the decision to go with musl was to build static binaries | 16:34 |
| kata-irc-bot | <archana.m.shinde> the idea is to replace the agent in place | 16:34 |
| kata-irc-bot | <archana.m.shinde> @fidencio Fedora does not use musl at all ? | 16:35 |
| kata-irc-bot | <archana.m.shinde> @jose.carlos.venegas.m lets create an issue for that, will help some clarity on blockers to moving to another too chain | 16:35 |
| kata-irc-bot | <fidencio> AFAIR (mind that I'd have to check it), musl is packaged for Fedora but I sincerely have no idea about its state | 16:36 |
| kata-irc-bot | <fidencio> The problem would start if we decide to use that on RHEL | 16:37 |
| kata-irc-bot | <fidencio> as musl is not present on RHEL, at all | 16:37 |
| kata-irc-bot | <fidencio> by RHEL I mean RHEL / CentOS | 16:37 |
| kata-irc-bot | <jose.carlos.venegas.m> @archana.m.shinde @fidencio let me open an issue for that | 16:38 |
| *** sgarzare has quit IRC | 16:38 | |
| kata-irc-bot | <archana.m.shinde> ack @fidencio | 16:40 |
| kata-irc-bot | <fidencio> @jose.carlos.venegas.m++, @archana.m.shinde++ | 16:48 |
| kata-irc-bot | <fidencio> We need a bot here ... :slightly_smiling_face: | 16:48 |
| kata-irc-bot | <archana.m.shinde> @sbrivio @fidencio Look forward to this. Will be quite helpful in seeing if we could add additional support in Kata for kernel debugging. | 16:57 |
| *** gwhaley has quit IRC | 17:05 | |
| *** david-lyle has joined #kata-dev | 17:45 | |
| *** dklyle has quit IRC | 17:46 | |
| *** david-lyle is now known as dklyle | 17:49 | |
| *** hashar has quit IRC | 18:27 | |
| *** crobinso has quit IRC | 18:45 | |
| *** davidgiluk has quit IRC | 19:22 | |
| *** hashar has joined #kata-dev | 19:50 | |
| *** sameo_ has quit IRC | 20:03 | |
| kata-irc-bot | <fidencio> So, we're releasing 1.11.0 Today? :slightly_smiling_face: | 20:14 |
| *** sameo_ has joined #kata-dev | 20:17 | |
| kata-irc-bot | <salvador.fuentes> @jose.carlos.venegas.m ^ | 20:22 |
| kata-irc-bot | <jose.carlos.venegas.m> @fidencio yeah we are starting release process today hope is smooth release process | 20:24 |
| kata-irc-bot | <fidencio> Let me go through the PRs then | 20:25 |
| kata-irc-bot | <jose.carlos.venegas.m> thank you! | 20:32 |
| kata-irc-bot | <fidencio> Okay, I've opened all the issues we've discussed in the meeting Today (and already closed some as those were already documented as a limitation) | 20:54 |
| kata-irc-bot | <fidencio> @wmoschet will take care of updating the tests upstream as time allows, as our plan is to *not* duplicate efforts an rely on upstream. Obviously, we may test things on Fedora first as we have more control there but, also as discussed in the meeting, we'll get those things back to upstream as soon as time allows! | 20:56 |
| *** hashar has quit IRC | 21:00 | |
| *** devimc has quit IRC | 21:09 | |
| *** ailan has quit IRC | 22:43 | |
| *** vgoyal has quit IRC | 23:00 | |
| *** sameo_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!