Friday, 2020-09-04

*** egernst has joined #kata-dev00:09
*** egernst has quit IRC00:23
*** fuentess has quit IRC01:39
*** dinechin_ has joined #kata-dev02:06
*** dinechin has quit IRC02:09
*** dinechin_ is now known as dinechin02:09
kata-irc-bot1<dgibson> can anyone tell me how to update the dependencies in the vendor/ subdirectory to a new version of their upstream source?04:12
*** bpradipt has joined #kata-dev05:13
*** bpradipt has quit IRC05:22
*** sameo has quit IRC05:23
*** sameo has joined #kata-dev05:26
*** pcaruana has joined #kata-dev06:04
*** pcaruana has quit IRC06:06
*** pcaruana has joined #kata-dev06:06
*** pcaruana has quit IRC06:15
*** pcaruana has joined #kata-dev06:28
*** dklyle has quit IRC06:37
*** fgiudici has joined #kata-dev06:53
kata-irc-bot1<lining2020x> The kata container using vsock boots up much slower than using kata-proxy.  Please have a look at:https://github.com/kata-containers/runtime/issues/292907:15
kata-irc-bot1<liam.merwick> @lining2020x What kernel version are you running on the host? There was a fix in Linux 5.7 that I think fixes this - this is the corresponding patch https://github.com/kata-containers/packaging/blob/master/kernel/patches/5.4.x/0002-net-virtio_vsock-Fix-race-condition-between-bind-and.patch I'm looking for the Kata issue that tracked it.08:36
kata-irc-bot1<liam.merwick> Looking at the kata-collect-data.sh output in #2929 it looks like you are running 4.18. I added a note to #2929 with a pointer to the path above and a link to https://github.com/kata-containers/runtime/issues/191708:51
kata-irc-bot1<lining2020x> @liam.merwick Cool !I test it on centos8 with a 4.18 kernel.09:08
*** fgiudici has quit IRC10:46
*** devimc has joined #kata-dev11:35
*** fuentess has joined #kata-dev12:13
kata-irc-bot1<salvador.fuentes> this guide should help: https://github.com/kata-containers/community/blob/master/VENDORING.md12:22
*** dinechin has quit IRC12:46
*** dinechin has joined #kata-dev12:46
*** bpradipt has joined #kata-dev13:00
*** crobinso has joined #kata-dev13:51
*** devimc has quit IRC14:02
*** devimc has joined #kata-dev14:02
*** dklyle has joined #kata-dev14:37
*** th0din has quit IRC14:45
*** sameo has quit IRC14:55
*** dinechin___ has joined #kata-dev17:08
*** sameo has joined #kata-dev17:09
*** dinechin has quit IRC17:10
*** dinechin___ is now known as dinechin17:10
*** crobinso has quit IRC17:40
*** th0din has joined #kata-dev18:03
*** bpradipt has quit IRC19:03
*** crobinso has joined #kata-dev20:34
*** devimc has quit IRC20:52
*** sameo has quit IRC21:29
*** fuentess has quit IRC21:41
kata-irc-bot1<eric.ernst> @chen.bo you know of the CLH API provides vcpu to guest cpu mapping when resizing the number of VCPUs?22:25
kata-irc-bot1<eric.ernst> I need a way to maintain the mapping between host threads and guest cpus22:26
kata-irc-bot1<chen.bo> @eric.ernst Let me check.22:26
kata-irc-bot1<eric.ernst> Thx. I just saw a general http response. There should be a way to map each thread.22:32
kata-irc-bot1<chen.bo> Right. I also checked that, the openAPI yaml in CLH. I see it only accepts the number of vcpus for resizing, nothing else. So it looks like not supporting what you want.22:33
kata-irc-bot1<chen.bo> So you want to ping a certain vcpu to a certain host thread?22:34
kata-irc-bot1<eric.ernst> Yep22:37
kata-irc-bot1<eric.ernst> Perhaps the threads are uniquely named22:38
kata-irc-bot1<eric.ernst> Iirc that’s what firecracker did22:38
kata-irc-bot1<chen.bo> The worker threads for each vcpu in clh are named as `vcpu0/vcpu1` , etc.22:41
kata-irc-bot1<eric.ernst> sweeeeet23:14
*** crobinso has quit IRC23:53

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!