Friday, 2020-09-25

kata-irc-bot<efoster> I'm guessing that the disk IO slide might be a bit out of date, though.00:38
kata-irc-bot<efoster> I've found disk IO performance to be 0-10% slower with virtio-fs.00:39
kata-irc-bot<eric.ernst> Comparing L2 versus L1?02:43
*** Rene__ has quit IRC06:52
*** Rene__ has joined #kata-general06:59
*** sameo has joined #kata-general07:00
*** sgarzare has joined #kata-general07:06
kata-irc-bot<caoruidong> https://github.com/kata-containers/documentation/blob/80702f6dd7ce211327d9f1ded4b11b120576c81b/Developer-Guide.md#set-up-a-debug-console I remember it has been a service08:08
*** sameo_ has joined #kata-general08:15
*** sameo has quit IRC08:18
*** sameo_ has quit IRC08:39
*** sameo_ has joined #kata-general08:57
*** sgarzare has quit IRC09:04
*** devimc has joined #kata-general12:23
*** fuentess has joined #kata-general13:26
*** pvdp66556 has quit IRC14:16
kata-irc-bot<efoster> Yep! Same scenario as the slide, Ds_v3 VM, `fio` benchmarking on a `runc` container vs a `kata` container14:49
*** devimc has quit IRC14:51
*** devimc has joined #kata-general14:52
kata-irc-bot<eric.ernst> cool.15:42
kata-irc-bot<eric.ernst> So, to clarify, you're saying its improved? 10% hit?15:43
kata-irc-bot<efoster> Ah, apologies. 10% hit. I *have* seen performance improvements on occasion, but that was with virtio-fs caching set to always, with a big chunk of memory set aside.15:51
kata-irc-bot<efoster> Hey folks! Quick question, does anyone know of possible workarounds to get some form of port-forwarding working when using CRI-O + Kata via shimv2? I've been thinking of having a daemonset that I could somehow jump through, but I'm not sure... Any thoughts?15:55
*** devimc has quit IRC20:57
*** sameo_ has quit IRC21:57
*** fuentess has quit IRC22:54
*** pvdp66556 has joined #kata-general23:05

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