Monday, 2025-04-28

hberaud[m]JayF: you mean when you use oslo.service?08:00
hberaud[m]Can you clarify your scenario08:00
hberaud[m]?08:00
gibiJayF: I'm trying to prevent accidental monkey_patching by https://review.opendev.org/c/openstack/nova/+/948311/1/nova/monkey_patch.py#11412:30
gibihberaud[m]: by the way, do we have any successful user of the oslo.service threading backend open patch? I try to use it locally and it feels broken in strange ways (after I winged some local patches to fix the obvious errors)12:31
gibisee my fresh comments in https://review.opendev.org/c/openstack/oslo.service/+/94572012:31
hberaud[m]I seen your comment. AFAIK we have no previous example of usages in real world. I also sent comments about problems I identified in this patch (https://review.opendev.org/c/openstack/oslo.service/+/945720/comment/e96ddcf0_3c2cc5cd/ , https://review.opendev.org/c/openstack/oslo.service/+/945720/comment/aed21ec8_eef31ee0/), but Daniel told me that the fix will coming soon. 12:36
hberaud[m]gibi: ^12:38
hberaud[m]damani: FYI ^12:39
gibihberaud[m]: OK. thanks12:41
hberaud[m]gibi: I cannot tell you more than that about oslo.service, hopefully Daniel will. AFAIK neutron and mistral delayed this part of the migration for when the new backend will be "more" ready. In all the case I think your approach is also helpful because it help to identify problems.12:44
gibiyeah I think giving feedback about the real use cases are needed to eventually get a working lib. 12:46
gibiI synced with neutron in the meantime and yeah I got the same info that they delayed this part of the migraton12:47
gibiNova has some code that needs to be verified with no monkey patching involved and instead of investing now deeply to transfrom the unit/functional test env for that I thought I will just spin up a nova-scheduler with full threading mode and trigger that piece of code. Hence my trials12:48
gibianyhow I can refocus for a while on other directions to give Daniel time to address the feedbacks12:49
hberaud[m]ack, thanks12:49
hberaud[m]Daniel just pushed a new patch set, but almost all the comments remains relevant for now12:50
hberaud[m]The latest PS should fix the Launcher topic13:13
gibiack13:15
gibiJayF, hberaud[m]: I had success with temporary fixes to of the oslo.service threading backend patch https://review.opendev.org/c/openstack/oslo.service/+/948310 The nova-scheduler is now able to start in native threading mode and scheduling VMs at least in a local devstack. CI hopefully give use some data points here https://review.opendev.org/c/openstack/nova/+/94831116:15
gibis/to of/top of/16:16
hberaud[m]Cool, good news16:16
JayFNice!16:34

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!