Wednesday, 2019-11-06

kata-irc-bot3<aprice> Apologies for being annoying and asking this, but when can we expect release notes and a static build for 1.9.1/1.10.0-alpha0 - https://github.com/kata-containers/runtime/releases00:24
*** eernst has joined #kata-dev00:41
*** eernst has quit IRC01:03
*** sameo has quit IRC01:03
*** irclogbot_1 has quit IRC05:34
*** irclogbot_1 has joined #kata-dev05:37
*** pcaruana has joined #kata-dev05:59
*** sameo has joined #kata-dev07:20
*** lpetrut has joined #kata-dev07:22
*** lpetrut has quit IRC07:33
*** lpetrut has joined #kata-dev08:14
*** lpetrut has quit IRC08:26
*** sgarzare has joined #kata-dev08:52
*** gwhaley has joined #kata-dev09:01
*** davidgiluk has joined #kata-dev09:08
*** sameo has quit IRC10:30
*** sameo has joined #kata-dev10:35
*** lpetrut has joined #kata-dev11:52
*** gwhaley has quit IRC12:04
*** gwhaley has joined #kata-dev12:53
gwhaleyyankcrime: hey there - so, I see there is a Kata I/O performance talk at the London OpenInfra Nov 28th! - posting link for @all to see: https://www.meetup.com/Openstack-London/events/266230014/12:59
gwhaleybrtknr: so, that'll be your talk then? ... heh heh, do we have to wait to the 28th to see the results ? ;-)13:01
yankcrimegwhaley: yes! sorry, should've thought to announce it in here as well13:04
yankcrimegwhaley: welcome back by the way, and we need to get you along to a future open infra meetup for another kata talk...13:05
*** sgarzare has quit IRC14:03
*** devimc has joined #kata-dev14:06
*** sgarzare has joined #kata-dev14:08
*** canyounot has joined #kata-dev14:13
*** fuentess has joined #kata-dev14:35
*** devimc has quit IRC14:35
*** devimc has joined #kata-dev14:47
*** igordc has joined #kata-dev15:50
*** lpetrut has quit IRC16:03
brtknrhi gwhaley, its all in the open, mainly based on the blog article I posted back  in May and small progress since to include Virtiofs16:53
brtknrBut I have only really be able to run the tests for read mode16:53
gwhaleybrtknr: thx -do you have a link to the blog to hand?16:54
gwhaleyI'm not sure I will be able to make the talk though :-(16:55
*** sameo has quit IRC16:58
gwhaleybrtknr: aha, this blog: https://www.stackhpc.com/kata-io-1.html ? :-)16:58
brtknrthats the one16:59
brtknrgwhaley: btw, welcome back! hope you had a nice caravaning trip?16:59
gwhaley:-) loooong campervan trip around europe.... many wine regions, lots and lots of sleep... thx!17:00
*** devimc has quit IRC17:11
*** devimc has joined #kata-dev17:12
davidgilukgwhaley: I was curious, were you at Lyon and if so was it just the longest possible route you could find to it?17:12
gwhaleydavidgiluk: heh heh - I got near Lyon-ish, but, noooo - I got back to the UK a little before the conference, and deliberately didn't want to mix work and holidays (this time...) :-)17:15
davidgiluknod17:15
*** igordc has quit IRC17:16
*** igordc has joined #kata-dev17:17
stefanhabrtknr: Please let me know if there's any virtio-fs help you need prior to your talk.  Maybe I can attend too.17:20
gwhaleywhilst here stefanha - do you have any recent kata virtio-fs 'numbers' per chance? :-)17:23
* gwhaley waves to stefanha - hope you enjoyed Lyon ?17:23
*** igordc has quit IRC17:23
stefanhagwhaley: Lyon was great.  Nice city and the conference was good as well.17:26
stefanhagwhaley: In the talk we gave we reused vgoyal's numbers.  They aren't new.17:26
stefanhagwhaley: Is there any I/O pattern or workload you are particularly interested in?17:26
gwhaleystefanha: no details of a specific setup. I've not run any tests for a long time, so was looking for recent representative numbers.17:28
gwhaleyheard somebody not seeing the improvements over 9p that might be expected - I'm presuming they have some setup thing that is not understood, but wanted to have some baseline 'what you should expect' sort of numbers to hand.17:28
gwhaleythe blog link above says 2x-8x over 9p, which is a start :-)  thx17:29
brtknrstefanha: Yes I'd love to figure out why the randread, randwrite and read scenarios remain stuck17:33
brtknronly managed to generate results for write scenario so far17:33
brtknrusing virtiofs17:33
stefanhagwhaley: It depends on a lot on the workload.  I think for some benchmarks we didn't see a huge improvement - but there should be no degradation.  And for others (with DAX!) you get a huge speed-up.17:37
*** sameo has joined #kata-dev17:37
stefanhabrtknr: Okay, I'll be around to investigate that next week.17:37
stefanhabrtknr: Feel free to ping me we here or on #virtio-fs.17:37
brtknrAnyone know which version of kata has started using virtiofs newer than 0.317:38
brtknrI remember stefanha mention a hanging issue which has been fixed since 0.317:39
gwhaleyfuentess may remember .... I don't see @eric.ernst on (irc at least - maybe slack)17:40
kata-irc-bot3<eric.ernst> hi.17:42
kata-irc-bot3<eric.ernst> 1.8?17:42
gwhaleyhmm, v1.9.0 release notes say "Experimental support for virtio-fs v0.3 with qemu": https://github.com/kata-containers/runtime/releases/tag/1.9.017:46
gwhaleyso, I dunno...17:46
gwhaleyin the shortlog for the agent ov v1.9.0 it says "kata_agent: use virtio-fs 0.3+ mount options" - so,maybe v1.9.0 then?17:47
*** sgarzare has quit IRC17:48
*** EricAdamsZNC has joined #kata-dev17:51
*** sameo has quit IRC18:01
*** gwhaley has quit IRC18:03
kata-irc-bot3<eric.ernst> sorry, yes, you are correct.18:06
kata-irc-bot3<eric.ernst> (shouldn't multitask:)18:06
*** lpetrut has joined #kata-dev18:25
*** lpetrut has quit IRC18:26
*** igordc has joined #kata-dev18:28
*** sameo has joined #kata-dev19:18
*** canyounot has quit IRC19:23
*** canyounot has joined #kata-dev19:29
*** davidgiluk has quit IRC20:12
*** pcaruana has quit IRC21:27
*** igordc has quit IRC21:51
*** eernst has joined #kata-dev21:59
*** eernst has quit IRC22:09
*** canyounot has quit IRC22:13
*** devimc has quit IRC22:18
*** dklyle has quit IRC22:35
*** dklyle has joined #kata-dev22:36
*** dklyle has quit IRC22:36
*** dklyle has joined #kata-dev22:36
*** fuentess has quit IRC23:36

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!