tonyb | hehe | 00:06 |
---|---|---|
frickler | what if we disabled CI for the EM branches completely? maybe call it "extended stasis" then instead. not sure if zuul would support that currently, but might be added as a feature | 06:50 |
frickler | that would solve all the issues I'm having with them from an opendev pov, also reduce moral load on teams to keep things working | 06:51 |
fungi | zuul does have a way to specify branches not to load configuration from, though i don't know if maintaining that list would be a lot of additional work either | 11:31 |
frickler | I've only seen this for config-projects "In the case of an untrusted-project, the configuration from every branch is included ..." https://zuul-ci.org/docs/zuul/latest/project-config.html#configuration-loading | 11:42 |
fungi | ahh, yeah and with some drivers, like being able to ignore unprotected branches in github | 11:43 |
frickler | also it might be necessary to also exclude globally defined jobs | 11:43 |
*** blarnath is now known as d34dh0r53 | 12:42 | |
*** ralonsoh__ is now known as ralonsoh | 14:26 | |
fungi | tonyb: josh gave you a thumbs-up on the proposed forum session about revisiting extended maintenance | 15:22 |
tonyb | Just replying now :) | 15:23 |
clarkb | The way gerrit acls are currently configured you need zuul to merge things. You might be able to address that via gerrit acls or just noop jobs for everything EM | 15:23 |
fungi | clarkb: i think the bigger concern is not ending up with a growing pile of zuul config errors from rotting configuration on abandoned branches | 15:28 |
fungi | and how to tell zuul not to look at the configs on those branches | 15:28 |
clarkb | if you had a config that was just check noop gate noop the chances of that bitrotting are very close to zero | 15:30 |
clarkb | still possible but far less likely | 15:30 |
fungi | yeah, but it does require someone to set that, and also to exclude those project+branches from globally-added jobs and templates | 15:42 |
clarkb | ya I'm just calling out that you hvae to do that anyway. Something has to be changed to accomplish that end state goal when the branch is in that stage of life. Whether that is modifying zuul configs or updatinggerrit acls or both | 15:43 |
rosmaita | tonyb: do yo have a day/time yet for the EM Forum session (want to make sure I don't schedule anything in conflict) | 15:44 |
tonyb | Looks like it'll be in the same room right after the existing cinder session | 15:45 |
tonyb | rosmaita: I'll keep you posted | 15:45 |
rosmaita | very convenient! | 15:45 |
rosmaita | thanks | 15:45 |
tonyb | rosmaita: I added a note to the existing etherpad | 15:45 |
tonyb | and once it's on the schedule I'll post to openstack-discuss | 15:46 |
rosmaita | sounds good | 15:46 |
dansmith | knikolla: I assume tc meeting is canceled next week? not sure I've seen a notification about that if so | 16:30 |
knikolla | dansmith: ah, good catch! yes, no meeting next week. | 16:34 |
gmann | fungi: yes, that is what I was proposing to end the testing (keep unit testing if ok) in EM will solve most of the issues | 16:45 |
gmann | did we get the new forum session on this ? or we continue as part of cinder sessions or ptg ? | 16:46 |
diablo_rojo_phone | There's a forum session for it now gmann | 17:10 |
fungi | gmann: same room as the cinder ptg session about it, immediately after | 17:15 |
fungi | come for the first, stay for the second. it's a two-for-one sale | 17:16 |
frickler | knikolla: fyi the https://zuul.opendev.org/t/openstack/config-errors page just received a facelift | 17:17 |
JayF | oh, that's *nice* | 17:19 |
gmann | great, I will check the schedule if no conflict. hope this late entry does not raise concern as we saw some objection earlier. | 17:19 |
knikolla | frickler: that looks fantastic! Will be exciting to see it shrink! | 17:37 |
frickler | knikolla: sadly the total count is missing for the moment, but we were under 200 just earlier | 17:39 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!