| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst https://github.com/kata-containers/tests/pull/1033 | 00:00 |
|---|---|---|
| kata-irc-bot | <jose.carlos.venegas.m> one more fix, could you ack | 00:00 |
| kata-irc-bot | <eric.ernst> ack | 00:00 |
| kata-irc-bot | <jose.carlos.venegas.m> thank you | 00:02 |
| kata-irc-bot | <eric.ernst> no, thank you carlos!@ | 00:26 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst @salvador.fuentes https://github.com/kata-containers/runtime/pull/1101 | 01:17 |
| kata-irc-bot | <eric.ernst> i see its merged | 01:18 |
| kata-irc-bot | <eric.ernst> cool | 01:18 |
| kata-irc-bot | <eric.ernst> @manohar.r.castelino @jose.carlos.venegas.m: We should update packaging scripts to include firecracker | 01:18 |
| kata-irc-bot | <eric.ernst> I know this is something @manohar.r.castelino was working on script for - pretty sttraight forward | 01:19 |
| kata-irc-bot | <eric.ernst> but once this is done, I want to include a few extra things in our tarball now: 1. a fc specific config.toml 2. firecracker binary 3. sh, kata-fc, which allows us to reuse kata-runtime by providing indirection (ie, kata-runtime --config=kata-fc-config.toml) 4. Probably same indirection for creating kata-qemu? | 01:21 |
| kata-irc-bot | <manohar.r.castelino> @eric.ernst still working on it.. I ran into some interesting kernel panics when building it as static in a CL container | 01:21 |
| kata-irc-bot | <eric.ernst> ouch | 01:21 |
| kata-irc-bot | <eric.ernst> @krsna1729 - I think some of the work that's in our deploy.sh (specifically the indirection) should be part of the tarball. WDYT? | 01:24 |
| kata-irc-bot | <krsna1729> ya and deploy.sh can only do the cp of the wrapper, just like it copies other binaries | 01:33 |
| kata-irc-bot | <eric.ernst> Yah | 01:41 |
| kata-irc-bot | <manohar.r.castelino> I really give up.... just do it with the default container they provide with firecracker | 02:25 |
| kata-irc-bot | <manohar.r.castelino> Kernel panic debugging due to new CPU instructions is not something I need to learn right now | 02:25 |
| *** fuentess has quit IRC | 02:56 | |
| *** sameo has joined #kata-dev | 06:54 | |
| *** sameo has quit IRC | 07:04 | |
| *** LinuxMe has joined #kata-dev | 07:30 | |
| *** jodh has joined #kata-dev | 07:32 | |
| *** LinuxMe has quit IRC | 07:37 | |
| *** lpetrut has joined #kata-dev | 07:55 | |
| *** davidgiluk has joined #kata-dev | 09:03 | |
| *** sameo has joined #kata-dev | 09:17 | |
| *** LinuxMe has joined #kata-dev | 09:33 | |
| *** LinuxMe has quit IRC | 09:38 | |
| *** gwhaley has joined #kata-dev | 09:39 | |
| *** fuentess has joined #kata-dev | 13:00 | |
| *** stackedsax has quit IRC | 13:04 | |
| *** LinuxMe has joined #kata-dev | 14:16 | |
| *** simosx has joined #kata-dev | 14:19 | |
| *** LinuxMe has quit IRC | 14:20 | |
| *** eernst has quit IRC | 14:40 | |
| *** davidgiluk has quit IRC | 15:26 | |
| *** lpetrut has quit IRC | 15:37 | |
| *** simosx has quit IRC | 15:37 | |
| * gwhaley looks around for maybe @fungi... | 15:41 | |
| gwhaley | hi openstack, fungi, clarkb - quick zuul-ish question - where is the code for the github zuul plugin, presuming it is in a git repo. I'm struggling to locate it :-) | 15:42 |
| *** davidgiluk has joined #kata-dev | 15:43 | |
| *** fungi has joined #kata-dev | 15:48 | |
| fungi | gwhaley: assuming you mean the github app, i'll get it | 15:49 |
| gwhaley | fungi: yep, the github app. thx. I want to have a peek at how it feeds back the info to github (not the comment posting, but the app status), as the link points to the generic zuul entry point, and it would be really nice if it pointed at the job that passed/failed directly instead etc. | 15:52 |
| gwhaley | also, I'd just like to know to add it to my list of 'where is the zuul kata config/code', so I can try and move on with adding metrics CI via zuul, and contemplate the pullapprove replacement aspects. | 15:53 |
| fungi | yeah, i have to learn a bit about github apps real quick, to determine if they're actually software which even has source code, or just some arbitrary construct referring to a collection of github permissions and webhook urls | 15:54 |
| gwhaley | fungi: I think they have some code - that probot repo has a bunch of I think node.js stuff and some other stuff I'd never heard of - but, tbh, I too am uncertain how much or what lives 'inside github', and what if anything has to live outside on a server node somewhere. | 15:57 |
| clarkb | fungi: gwhaley aiui a github app is a registration of a webhook target. The webhook handling code lives in https://git.zuul-ci.org/cgit/zuul | 15:58 |
| fungi | gwhaley: so it seems it's the latter. github apps aren't software, they're just a collection of permissions and pointers. a running zuul service would connect to github and interact with projects under the guises of an "app" construct | 15:58 |
| fungi | gwhaley: so, yes, as clarkb says, the "source code" for the zuul github app would be the zuul service source code | 15:59 |
| fungi | https://developer.github.com/apps/about-apps/#about-github-apps summarized says "write software which interacts with the github api" | 15:59 |
| fungi | so specifically it'll be the github connection driver in the zuul source tree | 16:00 |
| fungi | https://git.zuul-ci.org/cgit/zuul/tree/zuul/driver/github | 16:01 |
| gwhaley | fungi: clarkb: thx :-) | 16:02 |
| fungi | no problem! | 16:03 |
| clarkb | also on the reporting side of things I think the idea for making that better is to use github's new api for that. What is it called it has its own tab on PRs | 16:04 |
| gwhaley | I think that is the 'github checks' maybe. I don't think we are using anything that drives those in kata yet (but sure, it'll probably be inevitable one day as old APIs will probably get deprecated) | 16:05 |
| kata-irc-bot | <thierry> gwhaley: looks like GitHub checks support counting multiple approvers now -- would that work for that corner of the validation? | 16:36 |
| kata-irc-bot | <graham.whaley> ttx: yes, I think the github multiple ack counting would cover us for 80% of our use case. What it does not get us afaik is: | 16:37 |
| kata-irc-bot | <graham.whaley> - it cannot do 'groups', so we would lose the 'also needs documentation acks' type function - we can live without that for now I think | 16:37 |
| kata-irc-bot | <graham.whaley> - it will count acks from any member (maybe commit/write enabled member) of the org, which is not quite how we have the acks set up at present (but nearly). We currently count acks by 'group member' (teams), not the whole org. tbh, we can probably live with that as well right now | 16:38 |
| kata-irc-bot | <thierry> gwhaley: long term i think Zuul would benefit from having some way to express complex github approval rules anyway, but that's unlikely to be ready by Feb 1 | 16:39 |
| kata-irc-bot | <thierry> gwhaley: but DCO-style signed-off approval is there already | 16:39 |
| kata-irc-bot | <thierry> I'll continue to dig deeper, but it sounds promising | 16:40 |
| kata-irc-bot | <graham.whaley> DCO style we can also enalbe in our checkcommits I think, which we already run in the CIs (SOB check is off by default). | 16:40 |
| fungi | it sounded like codeowners along with required reviews might solve the documentation ack requirement? i'm afraid i don't know quite what the expectations are there though | 16:40 |
| kata-irc-bot | <graham.whaley> So, as a plan, we can start by enabling the github ack checks and the checkcommits SoB checks - and that should only duplicate the checks we currently have, and not break anything I think... | 16:40 |
| kata-irc-bot | <graham.whaley> fungi: on github, or in Zuul (I think github maybe?) | 16:41 |
| kata-irc-bot | <thierry> gwhaley: the WIP check would still be missing? | 16:43 |
| fungi | codeowners is a github thing where you can set expected reviewers on parts of your tree, i think? still learning, but zuul has the ability to check whether a pr meets github's merge requirement expectations | 16:43 |
| *** eernst has joined #kata-dev | 16:48 | |
| kata-irc-bot | <thierry> Wip check if done by labels should be simple enough to include in the Zuul check | 16:49 |
| kata-irc-bot | <graham.whaley> WIP by label would be fine - right now we do by label or subject keyword, but by label is better than not at all ;) | 16:57 |
| *** jodh has quit IRC | 17:00 | |
| kata-irc-bot | <graham.whaley> ooh, that page says you need '5th gen core', and the NUCs on the small farm are i7-5557U - which maybe qualify! | 17:21 |
| kata-irc-bot | <graham.whaley> (oops, wrong window! ;) ) | 17:21 |
| kata-irc-bot | <graham.whaley> ttx (and fungi), do you have an example of the Zuul github SOB checking - a link? I had a surf/search but didn't find one, and could not see an obvious way from the zuul github driver docs. thx | 17:58 |
| *** gwhaley has quit IRC | 17:59 | |
| fungi | here's one... https://github.com/ansible-network/cloud_vpn/pull/60 | 18:09 |
| fungi | they call their signed-off-by job "dco-license" since it's confirming the patch author has signed off in agreement with the developer certificate of origin | 18:10 |
| fungi | the "softwarefactory" bot there is a zuul installation, in case that's not obvious | 18:11 |
| *** lpetrut has joined #kata-dev | 18:33 | |
| *** lpetrut has quit IRC | 18:37 | |
| kata-irc-bot | <eric.ernst> Okay - working on this now - hopefully finishing it will coincide with @manohar.r.castelino having firecracker build working in packaging repo | 18:54 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst ready to merge https://github.com/kata-containers/packaging/pull/300 | 19:08 |
| kata-irc-bot | <jose.carlos.venegas.m> after that we create PR to come back to 4.19 | 19:09 |
| kata-irc-bot | <eric.ernst> done! | 19:09 |
| kata-irc-bot | <eric.ernst> sweet. merged. Yeah, and with that PR, we can start to look @ potential metric regressions? | 19:10 |
| kata-irc-bot | <jose.carlos.venegas.m> @eric.ernst @manohar.r.castelino how it is going ? it this only for the static binary tarball or want to request to have package to fc team? | 19:11 |
| kata-irc-bot | <eric.ernst> At this point just the tarball | 19:12 |
| kata-irc-bot | <jose.carlos.venegas.m> yeah | 19:12 |
| kata-irc-bot | <jose.carlos.venegas.m> I saw that the boot time failing metric was a change in the format in the boot log | 19:12 |
| kata-irc-bot | <jose.carlos.venegas.m> now we need to confirm memory | 19:12 |
| kata-irc-bot | <manohar.r.castelino> @jose.carlos.venegas.m I am testing the binary I build right now h@ | 19:12 |
| kata-irc-bot | <jose.carlos.venegas.m> @manohar.r.castelino for static qemu we used a dockerfile | 19:13 |
| kata-irc-bot | <jose.carlos.venegas.m> probably the you want to go in a similar way | 19:14 |
| kata-irc-bot | <jose.carlos.venegas.m> thx for the heads up | 19:14 |
| kata-irc-bot | <eric.ernst> Yeah, thankfully FC folks already have a dockerfile in place too! | 19:14 |
| kata-irc-bot | <eric.ernst> I'm curious what the config changes are, and how we created the 4.19 config (defconfig v noconfig) | 19:24 |
| *** davidgiluk has quit IRC | 20:18 | |
| *** fuentess has quit IRC | 21:58 | |
| kata-irc-bot | <eric.ernst> @jose.carlos.venegas.m i'm having fun thinking about the best way to create a firecracker configuration.toml for kata | 22:37 |
| kata-irc-bot | <eric.ernst> ie, having an option for alternative 'defaults' | 22:38 |
| kata-irc-bot | <eric.ernst> @manohar.r.castelino we've only done this by hand so far. | 22:38 |
| kata-irc-bot | <manohar.r.castelino> slack is confusing me.. where is this thread? | 22:39 |
| kata-irc-bot | <manohar.r.castelino> it shows under kata dev.. but seems private | 22:40 |
| kata-irc-bot | <eric.ernst> :) Its under the original comment 21 hours ago, i guess. | 22:44 |
| *** stackedsax has joined #kata-dev | 22:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!