| *** mylinux has joined #kata-dev | 00:03 | |
| *** mylinux has quit IRC | 00:07 | |
| *** annabelleB has quit IRC | 00:11 | |
| *** justJanne has quit IRC | 00:27 | |
| *** justJanne has joined #kata-dev | 00:40 | |
| *** mylinux has joined #kata-dev | 00:43 | |
| *** mylinux has joined #kata-dev | 00:45 | |
| *** mylinux has quit IRC | 00:53 | |
| kata-dev-irc-bot | <eric.ernst> @bergwolf can you verify that the config file actually isn’t the same as resulting .config? | 01:54 |
|---|---|---|
| kata-dev-irc-bot | <eric.ernst> If it isn’t, then I agree with you, it should be. | 01:54 |
| kata-dev-irc-bot | <bergwolf> @eric.ernst I'm sure it isn't. As found out in the comment, `CONFIG_RD_GZIP` is not present in the config file. It doesn't say it's enabled or disabled. | 01:57 |
| kata-dev-irc-bot | <eric.ernst> So if you run make olddefconfig there’s a delta? | 01:58 |
| kata-dev-irc-bot | <eric.ernst> Btw, I met Xiaochen today who said you were old colleagues from EMC. Smart guy! | 01:59 |
| kata-dev-irc-bot | <bergwolf> I didn't try. `CONFIG_RD_GZIP` should be there in the resulting .config. If not, it is disabled and we lose initrd support. | 01:59 |
| kata-dev-irc-bot | <bergwolf> yeah, we also used to play tennis a lot together | 02:00 |
| kata-dev-irc-bot | <eric.ernst> Nice. Okay, let me check when on my workstation in another hour. | 02:00 |
| kata-dev-irc-bot | <eric.ernst> Be back in an hour... | 02:03 |
| *** mcastelino has quit IRC | 02:04 | |
| kata-dev-irc-bot | <eric.ernst> @bergwolf hey, so just took a closer look at https://raw.githubusercontent.com/devimc/kata-packaging/d1fb8e895778517eb12b88a973160a40ca426e96/kernel/configs/x86_kata_kvm_4.14.x_defconfig | 03:06 |
| kata-dev-irc-bot | <eric.ernst> I would rather see a full resulting config used, but need to look at how they are planning to use this before having too much of an opinion. | 03:06 |
| kata-dev-irc-bot | <eric.ernst> if you run olddfefconfig with what is in the PR, it does end up having RD_GZIP. Having said that, I prefer "WYSIWYG" for configs | 03:07 |
| kata-dev-irc-bot | <eric.ernst> A developer shouldn't need to 1) build it against kernel or 2) read the Kconfigs to determine what the resulting output could be. | 03:07 |
| kata-dev-irc-bot | <eric.ernst> Not sure why they wanted to use defconfig here. | 03:09 |
| kata-dev-irc-bot | <bergwolf> @eric.ernst yes, I agree. We'd better use the final .config since that is more clear about what we get actually get. | 03:09 |
| kata-dev-irc-bot | <eric.ernst> stepping back, we should probably define cadence for updating kernel, because perhaps they didn't want burden of always updating the .config? I'll check to see if they have this documented yet.... | 03:10 |
| kata-dev-irc-bot | <eric.ernst> I know there was debate keeping the defconfig in tree versus in packaging repo... i'll need to read more/understand how they got to this point. | 03:11 |
| kata-dev-irc-bot | <bergwolf> thanks @eric.ernst. | 03:11 |
| kata-dev-irc-bot | <bergwolf> I am fine with keeping config in the packaging repo as well. Just want to make sure we have a correct kernel config. | 03:12 |
| kata-dev-irc-bot | <eric.ernst> if we move away from how they have it (just defconfig, that is storing the non-defaults only), then there's a burden of updating packaging repository based on kernel version updates. I think that's why the PR manages it this way. | 03:14 |
| kata-dev-irc-bot | <eric.ernst> Which makes sense to me, but is a bit painful for quickly verifying. | 03:15 |
| kata-dev-irc-bot | <bergwolf> The defconfig also needs to be updated when there is a kernel version upgrade. That's why it's named x86_kata_kvm_4.14.x_defconfig, with a kernel version string there. If it's just stable release, there is little chance that a new Kconfig option would be merged. So it's the same situation for both defconfig and a full kernel config. | 03:17 |
| kata-dev-irc-bot | <eric.ernst> Most likely, anyway. Okay -- let's clarify in the PR what we're looking for, and I can work with him in realtime tomorrow if needed to get it updated. | 03:19 |
| kata-dev-irc-bot | <eric.ernst> the defconfig story will be a bit more interesting post release assuming we start moving forward with multiple architecture support, etc. | 03:20 |
| kata-dev-irc-bot | <eric.ernst> But, at least for now, we need to get this in so we can move forward with appropriate packaging. | 03:20 |
| kata-dev-irc-bot | <bergwolf> And kernel developers have a bias toward enabling their own kernel modules by default. So I'm afraid we might enable something we do not need by accident. | 03:20 |
| kata-dev-irc-bot | <eric.ernst> :slightly_smiling_face: Yes, agreed | 03:21 |
| kata-dev-irc-bot | <eric.ernst> "why wouldn't someone want my sweet new driver?" | 03:21 |
| kata-dev-irc-bot | <bergwolf> Yup ;) | 03:21 |
| *** mylinux has joined #kata-dev | 03:24 | |
| *** mylinux has quit IRC | 03:29 | |
| *** sjas_ has joined #kata-dev | 04:18 | |
| *** sjas has quit IRC | 04:21 | |
| *** jodh has joined #kata-dev | 05:58 | |
| *** jodh has quit IRC | 05:58 | |
| *** jodh has joined #kata-dev | 05:58 | |
| kata-dev-irc-bot | <sebastien.boeuf> @bergwolf hey | 06:05 |
| kata-dev-irc-bot | <sebastien.boeuf> what do you mean here: https://github.com/kata-containers/runtime/pull/256#discussion_r183922765 | 06:05 |
| kata-dev-irc-bot | <sebastien.boeuf> what is the "runtime root storage directory " | 06:06 |
| kata-dev-irc-bot | <sebastien.boeuf> ? | 06:06 |
| kata-dev-irc-bot | <bergwolf> If you run `kata-runtime --help`, you'll see the `--root` option | 06:06 |
| kata-dev-irc-bot | <bergwolf> that's where the runtime root storage is specified | 06:07 |
| kata-dev-irc-bot | <bergwolf> surely by default it's `/var/run/virtcontainers` but it can be changed | 06:07 |
| kata-dev-irc-bot | <sebastien.boeuf> okay then I guess we set this up from the Makefile then | 06:08 |
| kata-dev-irc-bot | <bergwolf> It's checked in the `runtimeFlags` in main.go | 06:10 |
| kata-dev-irc-bot | <sebastien.boeuf> yes but this variable comes from the Makefile | 06:11 |
| kata-dev-irc-bot | <sebastien.boeuf> `const defaultRootDirectory = "$(PKGRUNDIR)"` | 06:11 |
| *** sameo has joined #kata-dev | 06:12 | |
| kata-dev-irc-bot | <bergwolf> defaultRootDirectory is just the default value | 06:13 |
| kata-dev-irc-bot | <sebastien.boeuf> yeah if you follow the Makefile, the default value is `/var/run/kata-containers` | 06:13 |
| kata-dev-irc-bot | <sebastien.boeuf> yes sure | 06:13 |
| kata-dev-irc-bot | <bergwolf> kata cli should use what's passed in via `--root` if provided | 06:13 |
| kata-dev-irc-bot | <sebastien.boeuf> so I understand that we can definitely use --root | 06:13 |
| kata-dev-irc-bot | <sebastien.boeuf> yep, and by default it will be `/var/run/kata-containers` as the default is defined this way | 06:14 |
| kata-dev-irc-bot | <bergwolf> yup, so in your PR, you should use the `--root` value | 06:16 |
| kata-dev-irc-bot | <bergwolf> it can be the default or not | 06:16 |
| kata-dev-irc-bot | <sebastien.boeuf> yes I'll change that | 06:16 |
| kata-dev-irc-bot | <bergwolf> thanks | 06:16 |
| kata-dev-irc-bot | <james.o.hunt> https://github.com/clearcontainers/runtime/issues/565 | 06:24 |
| kata-dev-irc-bot | <bergwolf> huh | 06:39 |
| kata-dev-irc-bot | <bergwolf> Do we have a corresponding issue in kata runtime? | 06:40 |
| kata-dev-irc-bot | <sebastien.boeuf> yep, I have realized that nothing uses this flag | 06:53 |
| kata-dev-irc-bot | <sebastien.boeuf> there is nothing in the code for that | 06:53 |
| kata-dev-irc-bot | <sebastien.boeuf> @bergwolf maybe we can leave the PR as it is and update in a follow up PR introducing the support for --root | 06:54 |
| kata-dev-irc-bot | <bergwolf> @sebastien.boeuf fine with me | 06:55 |
| kata-dev-irc-bot | <sebastien.boeuf> @bergwolf I have just opened https://github.com/kata-containers/runtime/issues/257 | 06:57 |
| kata-dev-irc-bot | <sebastien.boeuf> we need to look at this as I agree this is gonna be useful in order to define a root directory where we can expect to store things related to the containers | 06:58 |
| *** dtk has joined #kata-dev | 07:14 | |
| *** sameo has quit IRC | 07:17 | |
| *** sameo has joined #kata-dev | 07:47 | |
| *** gwhaley has joined #kata-dev | 08:01 | |
| *** dtk has quit IRC | 08:02 | |
| *** dtk has joined #kata-dev | 08:02 | |
| stefanha | Is anyone involved with CRI-O + ClearContainers? The CRI-O website says it's tested and I wonder how easy it would be to do CRI-O + Kata instead. In theory it should work too since Kata is an OCI runtime. | 08:52 |
| kata-dev-irc-bot | <bergwolf> @james.o.hunt PTAL. https://github.com/kata-containers/agent/pull/211 | 08:53 |
| sameo | stefanha: It already works | 09:33 |
| sameo | stefanha: You can do CRI-O+Kata the same way you do CRI-O+Clear Containers. | 09:34 |
| stefanha | sameo: Excellent :) | 09:48 |
| *** mrbobbytables has quit IRC | 10:37 | |
| *** mrbobbytables has joined #kata-dev | 10:37 | |
| *** gwhaley has quit IRC | 10:59 | |
| *** cdent has joined #kata-dev | 11:04 | |
| *** justJanne has quit IRC | 11:38 | |
| *** justJanne has joined #kata-dev | 11:38 | |
| *** gwhaley has joined #kata-dev | 12:20 | |
| *** Jeffrey4l has quit IRC | 12:23 | |
| *** Jeffrey4l has joined #kata-dev | 12:26 | |
| *** mylinux has joined #kata-dev | 12:27 | |
| *** eernst has joined #kata-dev | 13:23 | |
| *** mylinux has quit IRC | 13:35 | |
| *** devimc has joined #kata-dev | 13:38 | |
| *** eernst has quit IRC | 13:38 | |
| *** mylinux has joined #kata-dev | 13:41 | |
| *** eernst has joined #kata-dev | 13:50 | |
| *** mylinux has quit IRC | 14:00 | |
| *** mylinux has joined #kata-dev | 14:08 | |
| *** gabyc_ has joined #kata-dev | 14:29 | |
| *** annabelleB has joined #kata-dev | 14:53 | |
| *** Jeffrey4l has quit IRC | 15:01 | |
| *** mylinux has quit IRC | 15:01 | |
| *** sjas_ has quit IRC | 15:01 | |
| *** devimc has quit IRC | 15:01 | |
| *** djinni has quit IRC | 15:01 | |
| *** stefanha has quit IRC | 15:01 | |
| *** EricAdamsZNC has quit IRC | 15:01 | |
| *** ChanServ has quit IRC | 15:01 | |
| *** devimc has joined #kata-dev | 15:01 | |
| *** djinni has joined #kata-dev | 15:01 | |
| *** stefanha has joined #kata-dev | 15:01 | |
| *** EricAdamsZNC has joined #kata-dev | 15:01 | |
| *** ChanServ has joined #kata-dev | 15:01 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 15:01 | |
| *** Jeffrey4l has joined #kata-dev | 15:02 | |
| *** mylinux has joined #kata-dev | 15:04 | |
| *** sjas_ has joined #kata-dev | 15:04 | |
| *** libregeekingkid[ has quit IRC | 15:04 | |
| *** mylinux has quit IRC | 15:17 | |
| kata-dev-irc-bot | <sebastien.boeuf> stefanha: let me know if you run into any issue ;) | 15:19 |
| *** mylinux_ has joined #kata-dev | 15:22 | |
| * gwhaley just realises that is likely the stefanha whose vsock-nfs patches I have just run up here locally :-) Hi :-) | 15:30 | |
| *** annabelleB has quit IRC | 15:31 | |
| *** mylinux_ has quit IRC | 15:32 | |
| *** annabelleB has joined #kata-dev | 15:53 | |
| *** devimc has quit IRC | 15:57 | |
| kata-dev-irc-bot | <eric.ernst> @julio.montes @jose.carlos.venegas.m - can we talk about https://github.com/kata-containers/packaging/pull/18 | 15:57 |
| kata-dev-irc-bot | <eric.ernst> @bergwolf and I talked last night -- I think it'd be better to maintain a resulting config rathe than a minimal defconfig. | 15:57 |
| *** mcastelino has joined #kata-dev | 15:59 | |
| *** jodh has quit IRC | 16:00 | |
| *** bgmccollum_ has joined #kata-dev | 16:09 | |
| *** bgmccollum has quit IRC | 16:10 | |
| *** libregeekingkid[ has joined #kata-dev | 16:11 | |
| *** devimc has joined #kata-dev | 16:16 | |
| *** mylinux has joined #kata-dev | 16:19 | |
| *** mylinux has quit IRC | 16:25 | |
| *** annabelleB has quit IRC | 16:31 | |
| *** sameo has quit IRC | 16:35 | |
| kata-dev-irc-bot | <julio.montes> @eric.ernst defconf files are easier to maintain and read | 16:38 |
| kata-dev-irc-bot | <julio.montes> also we can have a base defconfig with all command options | 16:38 |
| *** annabelleB has joined #kata-dev | 16:38 | |
| kata-dev-irc-bot | <julio.montes> *common | 16:39 |
| kata-dev-irc-bot | <sebastien.boeuf> @julio.montes the concern with `defconfig` is that they give a minimal set of things, and we don't control which what could be included by default | 16:41 |
| kata-dev-irc-bot | <julio.montes> @sebastien.boeuf ok fair enough, I'll change it | 16:48 |
| kata-dev-irc-bot | <eric.ernst> Yeah, I think its just easier to have a "WYSISYG" | 16:50 |
| kata-dev-irc-bot | <eric.ernst> I don't want to necessarily dig into Kconfig or pull kernel to see the resulting options. | 16:50 |
| kata-dev-irc-bot | <sebastien.boeuf> what does this mean "WYSISYG" ? | 16:50 |
| kata-dev-irc-bot | <eric.ernst> And, we'll need to maintain per release anyway already. I understand the original idea around just defconfig (makes it easier in some wayhs) | 16:50 |
| kata-dev-irc-bot | <eric.ernst> sorry, what you see is what you get. In actually missed a W | 16:51 |
| kata-dev-irc-bot | <sebastien.boeuf> ok thx for the acronym ;) | 16:51 |
| *** gwhaley has quit IRC | 16:55 | |
| kata-dev-irc-bot | <eric.ernst> thanks for the update, @julio.montes | 17:05 |
| *** mylinux has joined #kata-dev | 17:11 | |
| *** eernst has quit IRC | 18:02 | |
| *** annabelleB has quit IRC | 18:13 | |
| *** sjas_ is now known as sjas | 18:32 | |
| *** mylinux has quit IRC | 18:43 | |
| *** diga has joined #kata-dev | 18:50 | |
| *** FL1SK has quit IRC | 18:58 | |
| *** FL1SK has joined #kata-dev | 19:14 | |
| *** diga has quit IRC | 19:19 | |
| *** eernst has joined #kata-dev | 19:42 | |
| *** eernst has quit IRC | 19:46 | |
| *** eernst has joined #kata-dev | 19:59 | |
| *** annabelleB has joined #kata-dev | 20:09 | |
| *** FL1SK has quit IRC | 20:16 | |
| *** annabelleB has quit IRC | 20:17 | |
| *** annabelleB has joined #kata-dev | 20:22 | |
| *** nirmoy has quit IRC | 20:42 | |
| *** cdent has quit IRC | 21:08 | |
| *** annabelleB has quit IRC | 21:26 | |
| *** annabelleB has joined #kata-dev | 21:33 | |
| *** gabyc_ has quit IRC | 21:45 | |
| *** fuentess has quit IRC | 21:57 | |
| *** devimc has quit IRC | 22:12 | |
| *** FL1SK has joined #kata-dev | 22:26 | |
| *** eocardon has quit IRC | 23:23 | |
| *** annabelleB has quit IRC | 23:35 | |
| *** annabelleB has joined #kata-dev | 23:47 | |
| *** annabelleB_ has joined #kata-dev | 23:51 | |
| *** annabelleB has quit IRC | 23:53 | |
| *** annabelleB_ is now known as annabelleB | 23:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!