Friday, 2026-11-13

rm_workamorin: Was this to me or someone else, I can’t tell heh 01:19
rm_workah prolly to Stephen01:33
rm_workSpeaking of stephenfin, you might be the one to give a second look at https://review.opendev.org/c/openstack/tooz/+/1007408 ?01:34
rm_workHuge performance gains / avoids some stalls, in my benchmarking01:35
rm_workWe’re using it internally because now we use Redis as our backend for oslo RPC with HTTP-RPC so we really tax it a bit 😅01:36
rm_workI think https://review.opendev.org/c/openstack/oslo.service/+/1012821 is close too, tkajinam hopefully you’re good with it now?01:44
opendevreviewMerged openstack/oslo.service master: Threading backend: fix more no_fork exit hangs  https://review.opendev.org/c/openstack/oslo.service/+/101282102:57
*** mhen_ is now known as mhen02:59
opendevreviewAdam Harwell proposed openstack/oslo.service stable/2026.2: Threading backend: reload on SIGHUP with no_fork  https://review.opendev.org/c/openstack/oslo.service/+/101553106:08
opendevreviewAdam Harwell proposed openstack/oslo.service stable/2026.2: Threading backend: fix more no_fork exit hangs  https://review.opendev.org/c/openstack/oslo.service/+/101553206:08
rm_work^^ so there’s the backports… I’m not sure if both of those are going to be acceptable as backports tho so I made them WIP and we can discuss?06:09
rm_workIt would be nice for me because I run 2026.2 in my deployment so I could drop the two manual patches, but also my CI is already handling it so it isn’t too bad06:10
rm_workBoth are on top of https://review.opendev.org/c/openstack/oslo.service/+/1013546 which I guess still needs a +2/+W as well06:14
rm_workClaude seems to think 1015531 might be at risk of being considered “change in behavior” even though the behavior it changes was a bug 🫠06:24
amorinrm_work , yes I was answering to stephenfin :) About your work on http driver, I tested long time ago, that a huge work as well, I wish I can find time to check this again soon!07:13
*** ralonsoh_ooo is now known as ralonsoh07:39
opendevreviewLuca Miccini proposed openstack/oslo.cache master: Honor socket_timeout and dead_retry in the bmemcached pool  https://review.opendev.org/c/openstack/oslo.cache/+/101554309:29
opendevreviewAdam Harwell proposed openstack/oslo.service stable/2026.2: Ensure spawn pool workers have logging configured  https://review.opendev.org/c/openstack/oslo.service/+/101554509:40
opendevreviewAdam Harwell proposed openstack/oslo.service stable/2026.2: Threading backend: reload on SIGHUP with no_fork  https://review.opendev.org/c/openstack/oslo.service/+/101553109:40
opendevreviewAdam Harwell proposed openstack/oslo.service stable/2026.2: Threading backend: fix more no_fork exit hangs  https://review.opendev.org/c/openstack/oslo.service/+/101553209:40
rm_work^^ needed that first one for tests to work… idk hopefully it is backportable too09:41
opendevreviewLuca Miccini proposed openstack/oslo.cache master: Honor socket_timeout and dead_retry in the bmemcached pool  https://review.opendev.org/c/openstack/oslo.cache/+/101554310:24
stephenfingmaan: I was coming here to suggest we talk about this synchronously (timezones, eh?) 😂10:54
stephenfingmaan: They don't need to go together, no. I'm happy to try just removing the dependencies for now (assuming hberaud and damani are too).10:55
stephenfingmaan: Okay, left a comment on the change. I'm happy to merge (and release) the rest of the stack. We can continue to discuss the change in backend and either merge or abandon that later11:11
rm_workWheeeee11:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: tooz service discovery  https://review.opendev.org/c/openstack/oslo.messaging/+/100762011:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: wire protocol and client  https://review.opendev.org/c/openstack/oslo.messaging/+/100762111:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: server side and entry point  https://review.opendev.org/c/openstack/oslo.messaging/+/100762211:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: fanout broadcaster  https://review.opendev.org/c/openstack/oslo.messaging/+/100762311:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: DevStack plugin, CI and docs  https://review.opendev.org/c/openstack/oslo.messaging/+/100762411:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: oslo.metrics counters  https://review.opendev.org/c/openstack/oslo.messaging/+/101111511:25
opendevreviewAdam Harwell proposed openstack/oslo.messaging master: HTTP-RPC driver: interop with the original driver  https://review.opendev.org/c/openstack/oslo.messaging/+/100762511:25
rm_workHad to rebase everything just to add a metrics job in patch 6 lol11:26
stephenfinrm_work: some follow-up requests on https://review.opendev.org/c/openstack/tooz/+/100740811:27
rm_workWill look thanks! 🙏11:28
rm_workUgh yeah I need to make a note that I should do a better job of humanizing some of those release notes. That particular one I read and understood so I mostly left it, but maybe I only understand it because I spent the time working on the fix, lol11:32
rm_workI’ll word up a better follow up (your text also looks right to me pretty much, but let me poke at it)11:33
rm_workThe type annotation thing I kind of threw up my hands on but maybe it’s possible to make it a bit better… have you poked the project folks about your patch in a while? I find you pretty much have to stay on top of people or they assume you don’t actually care and move on 🥲11:35
stephenfinrm_work: larger request on https://review.opendev.org/c/openstack/tooz/+/1007940 too11:45
stephenfinon the redis-py type annotation thing: I haven't pushed. Redis (i.e. the company, formerly Redis Labs) make enough money that they should be able to push on this stuff themselves 😅11:47
stephenfin(I have a much larger stack of changes locally too, but not point pushing that if I can get the existing smaller stack in. Anyway...)11:48
rm_workOn the casts, should I fix just the ones I added, or just fix all of them in this file12:26
rm_workI think I’ll just fix them all12:26
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Follow-ups to group reads without WATCH  https://review.opendev.org/c/openstack/tooz/+/101557312:32
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Follow-ups to group reads without WATCH  https://review.opendev.org/c/openstack/tooz/+/101557312:42
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Follow-ups to group reads without WATCH  https://review.opendev.org/c/openstack/tooz/+/101557312:49
rm_workOk I think it’s good now12:49
rm_workOh lol I should have looked at the other CR first, forgot it removed a bunch of these12:52
rm_workI’ll move it to the end12:55
opendevreviewMerged openstack/tooz master: redis: Read groups without WATCH  https://review.opendev.org/c/openstack/tooz/+/100740813:08
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Retry operations in one bounded layer  https://review.opendev.org/c/openstack/tooz/+/100794013:16
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Follow-ups to group reads without WATCH  https://review.opendev.org/c/openstack/tooz/+/101557313:16
rm_workOK here we go13:16
opendevreviewAdam Harwell proposed openstack/tooz master: redis: Recover after primary role changes  https://review.opendev.org/c/openstack/tooz/+/100388013:16
rm_workMoved them around a bit and did the decorator thing, though it hilariously immediately changes13:17
rm_workWill need a re-review from you on… we’ll all of them, but one of them lost a +2 because I had to move and tweak it13:19
rm_work*well, all of them13:19
gmaanstephenfin: works for me, checking the reviews. give me some time15:55
opendevreviewTakashi Kajinami proposed openstack/tooz master: Use FileNotFoundError to detect missing file  https://review.opendev.org/c/openstack/tooz/+/101537515:59
gmaanstephenfin: done, I will wait for Daniel  response otherwise will workflow next week16:04
gmaanand we can continue discussion on default backend independently. removing default installation first can be good and we might get projects input if it is ok for them and keeping eventlet as default at least till it is removed. we can proceed accordingly16:06
-opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline briefly while we reboot for system updates, but should return within a few minutes.17:42

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!