Saturday, 2018-10-27

*** tosky has quit IRC00:02
*** jamesmcarthur has joined #openstack-tc00:09
*** jamesmcarthur has quit IRC00:13
*** diablo_rojo has joined #openstack-tc01:45
*** diablo_rojo has quit IRC02:07
*** jamesmcarthur has joined #openstack-tc02:25
*** jamesmcarthur has quit IRC02:29
*** diablo_rojo has joined #openstack-tc02:59
*** diablo_rojo has quit IRC03:20
*** diablo_rojo has joined #openstack-tc03:49
*** jaosorior has quit IRC05:20
*** diablo_rojo has quit IRC06:08
*** jaosorior has joined #openstack-tc08:38
*** e0ne has joined #openstack-tc08:52
*** e0ne_ has joined #openstack-tc08:56
*** e0ne has quit IRC08:58
*** jaosorior has quit IRC09:05
*** e0ne_ has quit IRC09:39
mnaserclarkb: in regards to that discussion, i think we need to work and help tripleo get more stable in whatever way possible, however, from a community pov, they (redhat) brings in a tremendous amount of resources inside openstack in terms of devs so i think it's a bit of a give and take11:22
mnaserclose to 8k commits in rocky so far, and way ahead anyone else (and i'm sure those aren't typo fixes either :p)11:22
mnaseralmost consistently ~4x more the #2 committing company11:23
clarkbmnaser: ya thats why I find the comparison with other deployment projects interesting. We need longer span of data before jumping to conclusions but if kolla and osa can test openstack deployments with far more resources tripleo likely can too. But also I really want to get away from the idea that adding eg airshipor stx is going yo suddenly precent nova from merging code11:25
clarkb*far fewer resources11:26
mnaserclarkb: yes i agree, their volume is very small.  but i also think that given the numbers, tripleo just has more contributors, we've (osa) had our fair share really bad gate issues but i guess given our velocity is slower, we don't make a big dent11:26
mnaseranother thing we've done is not have a shared queue too in the gat11:26
mnasergate*11:26
mnaserso we don't cause as much resets really if some role failed11:27
clarkbalsp osa seems to try and fix those problems :)11:27
mnaseryeah i think tripleo also does this thing where they abandon all changes, stop approving stuff and what not to help not affect too much but yeah maybe we need a bit of an awareness like11:28
mnaser"hey these are all shared resources so lets all be nice to each other"11:28
clarkba big part of the tripleo pain seems to be how big and complicate it is so understanding failures is hard. I know because Ive tried. The good news is tripleo has committed to working on making that and other stuff better11:28
clarkbansible rune mistral which runs a sible which runs puppet in docker or something11:29
mnaseryeah, it's a huge system so indeed it can take a really long time to fix things11:29
clarkbits not easy to unravel11:29
mnaserand i think the other hard thing is it's also not a quick deployment tool either so11:29
mnaseriterating to try and find fixes using the gate can take like... 2-3 attempts a day and you're done11:29
*** tosky has joined #openstack-tc12:04
clarkbanother factor worth noting is the openstack integrated gate has also been flaky. There is a non zero chance that those bugs are also causing problems for tripleo. So we shouldnt assume this is a tripleo only problem to solve12:11
*** jaosorior has joined #openstack-tc12:15
*** dklyle has quit IRC14:04
*** dklyle has joined #openstack-tc14:06
*** jamesmcarthur has joined #openstack-tc14:22
*** dklyle has quit IRC15:03
*** dklyle has joined #openstack-tc15:03
*** jamesmcarthur has quit IRC15:48
*** jamesmcarthur has joined #openstack-tc15:49
*** jamesmcarthur has quit IRC15:53
*** dklyle has quit IRC16:15
*** e0ne has joined #openstack-tc16:39
*** e0ne has quit IRC16:53
*** tosky has quit IRC22:52

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