| *** tetsuro has joined #openstack-tc | 00:27 | |
| gmann | o/ | 01:32 |
|---|---|---|
| *** ricolin has joined #openstack-tc | 02:01 | |
| *** tetsuro has quit IRC | 02:20 | |
| *** redrobot has quit IRC | 03:42 | |
| *** camelCaser has quit IRC | 04:23 | |
| *** camelCaser has joined #openstack-tc | 04:26 | |
| *** iurygregory has joined #openstack-tc | 06:25 | |
| *** e0ne has joined #openstack-tc | 06:27 | |
| *** ralonsoh has joined #openstack-tc | 07:29 | |
| *** tosky has joined #openstack-tc | 07:39 | |
| *** markvoelker has quit IRC | 07:45 | |
| *** markvoelker has joined #openstack-tc | 07:47 | |
| *** markvoelker has quit IRC | 07:52 | |
| *** tosky has quit IRC | 08:04 | |
| *** tosky_ has joined #openstack-tc | 08:04 | |
| *** tosky_ is now known as tosky | 08:19 | |
| *** tetsuro has joined #openstack-tc | 08:24 | |
| *** tetsuro has quit IRC | 08:48 | |
| *** ricolin has quit IRC | 10:48 | |
| *** tkajinam has quit IRC | 10:59 | |
| *** e0ne has quit IRC | 12:33 | |
| *** e0ne has joined #openstack-tc | 12:37 | |
| *** rm_work has quit IRC | 12:38 | |
| *** e0ne has quit IRC | 12:39 | |
| *** mordred has quit IRC | 12:40 | |
| *** rm_work has joined #openstack-tc | 12:42 | |
| *** mordred has joined #openstack-tc | 12:48 | |
| knikolla | o/ | 13:28 |
| *** bnemec has joined #openstack-tc | 13:30 | |
| cloudnull | o/ | 13:34 |
| *** ricolin has joined #openstack-tc | 13:43 | |
| jungleboyj | o/ | 13:47 |
| njohnston | o/ | 13:53 |
| *** e0ne has joined #openstack-tc | 14:00 | |
| mnaser | tc-members: gouthamr has been waiting for reviews for 12 days without much response -- can everyone please look at https://review.opendev.org/#/c/740509/ ? | 14:06 |
| gouthamr | o/ ty mnaser :) | 14:13 |
| openstackgerrit | Luigi Toscano proposed openstack/governance master: V goals, Zuul v3 migration: update links and grenade https://review.opendev.org/741987 | 14:22 |
| *** e0ne has quit IRC | 14:22 | |
| *** e0ne has joined #openstack-tc | 14:23 | |
| *** Guest23136 has joined #openstack-tc | 14:47 | |
| *** Guest23136 is now known as redrobot | 14:49 | |
| *** KeithMnemonic has joined #openstack-tc | 14:59 | |
| *** cmurphy_afk is now known as cmurphy | 15:07 | |
| *** e0ne has quit IRC | 16:15 | |
| *** ricolin has quit IRC | 16:19 | |
| *** e0ne has joined #openstack-tc | 17:20 | |
| dansmith | knikolla: smcginnis mnaser gmann: thinking forward to the call tomorrow -- given the work in the last 48 hours to prove it's possible, do we want to make the "goal" actually to revert the devstack default of glance-as-standalone? | 17:43 |
| dansmith | meaning, shoot for WSGI mode being the default for glance like everything else, both in devstack but also their docs? | 17:43 |
| dansmith | I need to figure out whether they should have a wsgi-based job as the "alternative" in their tree, or whether the base devstack job should be wsgi and they have a standalone alternative job | 17:44 |
| dansmith | even though I just did all the work to make standalone the default, I think we actually should just revert back to defaulting to wsgi | 17:44 |
| dansmith | based on what I know now. | 17:44 |
| fungi | at least the standalone default is probably backportable while the switch to uwsgi as default may not be, so it's possible those patches were not entirely a waste | 17:46 |
| fungi | (i mean, aside from their obvious educational benefit and effective conversation starter) | 17:46 |
| dansmith | fungi: backporting standalone patches to what? older devstack? | 17:47 |
| fungi | right | 17:47 |
| fungi | assuming they're needed for ussuri or older | 17:47 |
| dansmith | I'm not sure why we'd do that.. they could already run in standalone mode, | 17:47 |
| dansmith | it was just that all of devstack had to run services as stanalone instead of just glance as a snowflake | 17:47 |
| fungi | ahh, yep | 17:48 |
| dansmith | I made it so glance can be controlled separately, and then we defaulted that to standalone | 17:48 |
| fungi | and i guess there's little interest in stable devstack jobs doing uwsgi for everything except glance | 17:49 |
| dansmith | yeah I wouldn't think so | 17:49 |
| dansmith | they currently have no tempest tests for the new features that couldn't previously run in wsgi mode, so there's not really any reason I don't think | 17:50 |
| fungi | got it | 17:50 |
| dansmith | one good reason to revert the default to all-wsgi, is grenade | 17:50 |
| dansmith | switching from wsgi to standalone in the middle of grenade sucks, | 17:51 |
| dansmith | so we just made grenade stay wsgi-to-wsgi, which previously meant no import tests there, | 17:51 |
| dansmith | but if we land the fixes to glance wsgi, then we stop having that disparity | 17:51 |
| fungi | right, probably need a full cycle to attack that so have non-wsgi jobs do ussuri-victoria and wsgi jobs doing victoria-w | 17:52 |
| dansmith | well, since we've been doing glance as wsgi in devstack for many cycles, | 17:53 |
| fungi | ooh, yep i guess it's already sort of "done" then | 17:53 |
| dansmith | I'm saying we go back to all-wsgi for devstack and grenade so that we just never concern ourselves with having to switch models mid-upgrade just so that we can run the newer tests | 17:53 |
| dansmith | and then they can have a standalone job in their tree for making sure the standalone thing continues to perform, but probably no need to run grenade that way, | 17:54 |
| fungi | fair | 17:54 |
| dansmith | and if so, standalone-to-standalone would work too | 17:54 |
| gmann | dansmith: let's wait for tomorrow call and then we can revert the devstack to uwsgi. | 17:58 |
| dansmith | gmann: for sure wait until after the call, and the wsgi stack isn't even merged yet.. I just mean I think the *goal* should be agreement that the default should be wsgi | 17:59 |
| gmann | yeah totally agree. | 18:01 |
| dansmith | FWIW, I think I have addressed _all_ the wsgi deficiencies that have been raised at this point: threads, subprocess exec, graceful shutdown, config reload | 18:02 |
| gmann | and we can also run new import tests on top of your wsgi fix patch. | 18:05 |
| *** e0ne has quit IRC | 18:15 | |
| *** ralonsoh has quit IRC | 18:45 | |
| *** bnemec has quit IRC | 19:03 | |
| *** e0ne has joined #openstack-tc | 19:17 | |
| *** e0ne has quit IRC | 19:47 | |
| *** e0ne has joined #openstack-tc | 20:04 | |
| *** e0ne has quit IRC | 20:12 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/governance master: Deprecate os_congress project https://review.opendev.org/742533 | 20:14 |
| *** bnemec has joined #openstack-tc | 21:04 | |
| *** bnemec has quit IRC | 21:11 | |
| *** bnemec has joined #openstack-tc | 21:18 | |
| gmann | gouthamr: replied on https://review.opendev.org/#/c/740509/ | 22:50 |
| gmann | am thinking manila should check resources during upgrade also which is what assert:supports-accessible-upgrade tag is, mainly during verify_noapi phase. | 22:51 |
| gmann | or any other job than grenade verify that? | 22:51 |
| gouthamr | gmann: hi, thanks for commenting on the patch, i'll take a look | 22:54 |
| gouthamr | gmann: nope, grenade's our only upgrade test in the upstream gate at least | 22:55 |
| gmann | gouthamr: ok, i think we need to have some checks in grenade's verify_noapi phase then. | 22:56 |
| tosky | gmann: but isn't it a separate issue than the zuulv3 porting? | 22:57 |
| tosky | oh, that's not the zuulv3 job | 22:57 |
| tosky | sorry | 22:57 |
| gouthamr | gmann: agreed... i'll see if we can get that in | 22:57 |
| * tosky back to his cave | 22:57 | |
| gmann | tosky: yeah, this is about tag verification | 22:58 |
| gmann | also few other projects like keystone also does not do that either we need to verify some created users there from DB or something otherwise say not applicable. | 22:58 |
| * gouthamr what timezone is this tosky guy in :O | 22:58 | |
| tosky | late :) | 22:58 |
| gmann | anyways those projects will be checked during tag cleanup task I am doing. | 22:58 |
| gmann | gouthamr: heh tosky is 24/7 guy :) | 22:59 |
| gouthamr | gmann: oh, are you auditing all the tags? api-interop was going to be my next one | 22:59 |
| gmann | gouthamr: yeah, api-interop is ok for manila as microversion is there for discovery and all | 23:00 |
| gmann | but there are many other tag like 'assert:supports-zero-downtime-upgrade' where we need some testing and then ask projects to adopt there is no project having this tag for now | 23:01 |
| gouthamr | gmann: ah true... | 23:01 |
| gmann | 'Zero downtime to API availability' not sure if that is valid for what all projects and need some common framework to test it | 23:02 |
| gmann | anyways tea time for me. but if you propose api-interop for manila, i think that will be quick. | 23:03 |
| gouthamr | gmann: thank you, will do and ping :) | 23:03 |
| *** tosky has quit IRC | 23:03 | |
| gmann | gouthamr: and sorry for late response on that, i kept it open for review and then got busy in writing glance tests. but mnaser's ping reminded me. | 23:05 |
| gouthamr | ack no problem at all :) | 23:05 |
| *** markvoelker has joined #openstack-tc | 23:27 | |
| *** markvoelker has quit IRC | 23:32 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!