| *** yingjun has joined #kata-dev | 00:30 | |
| *** zerocoolback has joined #kata-dev | 01:00 | |
| *** Jeffrey4l_ has joined #kata-dev | 02:59 | |
| *** Jeffrey4l_ has quit IRC | 03:22 | |
| *** l4yerffej_ has joined #kata-dev | 03:22 | |
| *** l4yerffej_ has quit IRC | 03:23 | |
| *** l4yerffej has joined #kata-dev | 03:23 | |
| *** yingjun has quit IRC | 04:10 | |
| *** sjas_ has joined #kata-dev | 04:23 | |
| *** sjas has quit IRC | 04:26 | |
| *** yingjun has joined #kata-dev | 04:32 | |
| *** yingjun has quit IRC | 04:35 | |
| *** sameo has joined #kata-dev | 08:45 | |
| *** sjas_ is now known as sjas | 09:38 | |
| *** sameo has quit IRC | 10:01 | |
| *** yingjun has joined #kata-dev | 11:59 | |
| *** devimc has joined #kata-dev | 12:17 | |
| *** fuentess has joined #kata-dev | 12:43 | |
| *** annabelleB has joined #kata-dev | 13:53 | |
| kata-dev-irc-bot | <anne> Reminder that there's an Arch Committee meeting in one hour. Agenda: https://etherpad.openstack.org/p/katacontainers-2018-architecture-committee-mtgs | 13:56 |
|---|---|---|
| *** annabelleB has quit IRC | 14:10 | |
| *** annabelleB has joined #kata-dev | 14:12 | |
| *** zerocoolback has quit IRC | 14:18 | |
| *** yingjun has quit IRC | 14:25 | |
| *** devimc has quit IRC | 14:33 | |
| *** eernst has joined #kata-dev | 14:38 | |
| *** devimc has joined #kata-dev | 14:46 | |
| kata-dev-irc-bot | <sebastien.boeuf> @bergwolf it's read only | 14:47 |
| kata-dev-irc-bot | <sebastien.boeuf> @bergwolf not sure what makes it RO though. I'll ask @julio.montes and @manohar.r.castelino about that | 14:48 |
| kata-dev-irc-bot | <eric.ernst> @laijs - are you presenting vm templating today? | 14:52 |
| kata-dev-irc-bot | <laijs> yes | 14:54 |
| kata-dev-irc-bot | <laijs> there is no slides | 14:54 |
| *** spotz_ has joined #kata-dev | 14:55 | |
| kata-dev-irc-bot | <laijs> I’m sorry that I sent an updated to the qemu community without cc-ing you @eric.ernst | 14:56 |
| *** spotz_ is now known as spotz | 14:59 | |
| *** zerocoolback has joined #kata-dev | 15:01 | |
| *** fuentess has quit IRC | 15:20 | |
| *** fuentess has joined #kata-dev | 15:20 | |
| *** Jeffrey4l has left #kata-dev | 15:26 | |
| kata-dev-irc-bot | <jesse.butler> i'm having a hard time staying connected to today's call (network issues at co-working space). does anyone have details on what might be required for bare metal resources? i can check into availability on oracle's bare metal service. | 15:47 |
| kata-dev-irc-bot | <anne> @jesse.butler: @sebastien.boeuf and @graham.whaley will have reqs | 15:52 |
| kata-dev-irc-bot | <jesse.butler> and hopefully i'll have a better network for the next meeting... but slack/email can work for now. @eric.ernst may know as well. i can check into it when i have a bit of info. | 15:54 |
| *** zerocoolback has quit IRC | 15:57 | |
| *** eernst has quit IRC | 15:57 | |
| *** zerocoolback has joined #kata-dev | 15:58 | |
| *** eernst has joined #kata-dev | 15:58 | |
| *** zerocoolback has quit IRC | 16:03 | |
| kata-dev-irc-bot | <eric.ernst> @jesse.butler Thanks! The basic need is bare-metal; nothing very specific regarding hardware requirements - just need consistent results between runs (same hardware, no noisy neighbors on the node). | 16:08 |
| kata-dev-irc-bot | <eric.ernst> at least, this is how I understand it as far as requirements. Nested VT may also be useful if we plan to run both baremetal and within VM (since this is also an important use case) | 16:08 |
| kata-dev-irc-bot | <jesse.butler> any ideas on resource specifics (compute capacity, storage requirements, etc)? | 16:14 |
| kata-dev-irc-bot | <sebastien.boeuf> @jesse.butler thanks for looking into this. As @eric.ernst mentioned, the only specific thing that comes to mind is the capability of nesting VMs. About CPU/RAM/DISK, IIRC we don't need something fancy with 128cores and 1To of RAM, but @graham.whaley is the one who really used it and who designed the system, so I'll let him answer on that (he'll be back tomorrow). | 16:21 |
| kata-dev-irc-bot | <jesse.butler> nod, ok. this, ironically, might be the issue... bare metal instances aren't very small in general. ours (oracle's) even less so.. our smallest instance might just be too big to donate. but, i will check into it (already have fired off a note). i'll get back when i know more. | 16:22 |
| kata-dev-irc-bot | <jesse.butler> follow-up question - maybe for @graham.whaley - would resources be useful if not committed 100% of the time? kind of breaks down the CI/CD model to have service windows, but it might be more doable on our end, given the sizing issue | 16:23 |
| kata-dev-irc-bot | <jesse.butler> also worth mentioning from what i could gather from today's call, our v12n team are looking at getting some patches in upstream that allow us to just use qemu with kata. | 16:28 |
| kata-dev-irc-bot | <jesse.butler> i think that was something that came up today, but maybe just in passing (or maybe i totally misunderstood it?) | 16:28 |
| kata-dev-irc-bot | <jesse.butler> but if anyone's interested, that is something we are working on here. internally we're running with our qemu build (not lite) | 16:29 |
| kata-dev-irc-bot | <sebastien.boeuf> Yeah, that's why we still want to test with a vanilla version of qemu, to make sure we don't advertise numbers that we can only reach with a specific qemu version (lite) | 16:31 |
| *** fiddletwix has joined #kata-dev | 16:33 | |
| kata-dev-irc-bot | <eric.ernst> FYI - metric's presentation (cp/paste of what Seb presented, only with open permissions): https://docs.google.com/presentation/d/1TEvU0nfjODn8bn1vF2s2EOTw8PJ14wq54skk9t3c8NA/edit?usp=sharing | 16:44 |
| kata-dev-irc-bot | <anne> And recording of the call: https://zoom.us/recording/play/CTg6V-Gy4RB_UolsiwcG90iAs_LZYrTEcVt3if42IQYv_CqV3scSqnhijQVzIJ2n | 17:16 |
| *** mcastelino has joined #kata-dev | 17:22 | |
| *** mcastelino has quit IRC | 17:23 | |
| *** mcastelino has joined #kata-dev | 17:23 | |
| *** justJanne has quit IRC | 17:37 | |
| *** justJanne has joined #kata-dev | 17:37 | |
| *** jlbutler has quit IRC | 20:19 | |
| *** jlbutler has joined #kata-dev | 20:19 | |
| *** fiddletwix has quit IRC | 20:22 | |
| *** annabelleB has quit IRC | 20:25 | |
| *** annabelleB has joined #kata-dev | 20:40 | |
| kata-dev-irc-bot | <maran.wilson> I sent an email to the kata-dev mailing list just now to elaborate on what @jesse.butler mentioned regarding our ongoing efforts to upstream a solution that would allow Qemu to rapidly start KVM guests without firmware, via booting directly to the uncompressed Linux kernel. | 20:46 |
| kata-dev-irc-bot | <eric.ernst> @maran.wilson this is great. | 21:01 |
| kata-dev-irc-bot | <eric.ernst> thanks for sharing. | 21:01 |
| kata-dev-irc-bot | <eric.ernst> I assume you've done some testing of guest kernel and qemu patches ; any early numbers with respect to relative boot time and/or footprint? | 21:02 |
| kata-dev-irc-bot | <eric.ernst> not having to maintain a "lite" set of changes is very attractive. | 21:03 |
| kata-dev-irc-bot | <maran.wilson> No problem. We have a working proof of concept, but no perf data yet. | 21:03 |
| kata-dev-irc-bot | <eric.ernst> fair enough. | 21:03 |
| kata-dev-irc-bot | <maran.wilson> Working on making some changes to the Qemu design a bit to match what the maintainers seem to want. | 21:04 |
| kata-dev-irc-bot | <maran.wilson> so might as well get that part done and then start really looking into the performance and full integration into kata containers | 21:04 |
| kata-dev-irc-bot | <eric.ernst> Sure. | 21:07 |
| kata-dev-irc-bot | <sebastien.boeuf> that sounds great :slightly_smiling_face: | 21:07 |
| kata-dev-irc-bot | <eric.ernst> We have some QEMU folks over here that I'm going to make sure start tracking some of the upstream PRs. | 21:07 |
| kata-dev-irc-bot | <maran.wilson> Awesome. And I wouldn't be surprised if we end up reaching out to you guys for ideas when it comes to tying it all together with elements of the pc-lite qemu patches your folks have worked on in the past. | 21:10 |
| kata-dev-irc-bot | <eric.ernst> K, don't hesitate. I know some of our work has made it upstream already. | 21:11 |
| *** devimc has quit IRC | 21:14 | |
| *** jbryce has quit IRC | 21:18 | |
| *** jbryce has joined #kata-dev | 21:18 | |
| *** sameo has joined #kata-dev | 21:22 | |
| *** annabelleB has quit IRC | 21:34 | |
| *** justJanne has quit IRC | 21:35 | |
| *** justJanne has joined #kata-dev | 21:35 | |
| *** eernst has quit IRC | 21:35 | |
| *** eernst has joined #kata-dev | 21:36 | |
| *** eernst has quit IRC | 21:37 | |
| *** eernst has joined #kata-dev | 21:37 | |
| *** eernst has quit IRC | 21:38 | |
| *** eernst has joined #kata-dev | 21:38 | |
| *** justJanne has quit IRC | 21:42 | |
| *** justJanne has joined #kata-dev | 21:43 | |
| *** justJanne has quit IRC | 21:44 | |
| *** justJanne has joined #kata-dev | 21:45 | |
| *** eernst has quit IRC | 22:05 | |
| *** eernst has joined #kata-dev | 22:06 | |
| *** eernst has quit IRC | 22:06 | |
| *** eernst_ has joined #kata-dev | 22:07 | |
| *** eernst_ has quit IRC | 22:07 | |
| *** eernst has joined #kata-dev | 22:15 | |
| *** fuentess1 has joined #kata-dev | 22:26 | |
| *** annabelleB has joined #kata-dev | 22:41 | |
| *** jlbutler has quit IRC | 22:49 | |
| *** eernst has quit IRC | 23:11 | |
| *** eernst has joined #kata-dev | 23:14 | |
| *** eernst has quit IRC | 23:18 | |
| *** eernst has joined #kata-dev | 23:19 | |
| *** eernst has quit IRC | 23:20 | |
| *** eernst has joined #kata-dev | 23:20 | |
| *** eernst has quit IRC | 23:24 | |
| *** sameo has quit IRC | 23:38 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!