*** hemna8 is now known as hemna | 07:38 | |
opendevreview | Iago Filipe proposed openstack/nova master: Adding unversioned notification publish mechanism https://review.opendev.org/c/openstack/nova/+/823200 | 11:45 |
---|---|---|
noonedeadpunk | frickler: oh, btw, is `folks` gender neutral? As according to google translate seems that yes, but not sure... | 12:47 |
noonedeadpunk | also question - what to do if you've catched `nova.exception.TooOldComputeService`? Yes, I know that upgrades through version are not supported and thus this should not happen if everything is done correctly. | 12:49 |
noonedeadpunk | But, if this has happened, and conductor being restarted, it goes to crash loop with TooOldComputeService. And even when you upgrade all computes - it doesn't return back, as computes needs conductor to report their new version, that can't spawn because computes are still "old" for it | 12:51 |
noonedeadpunk | so it feels like you require to manually bump version in cell1 but that is.... frustrating a bit I guess... | 12:51 |
noonedeadpunk | or I don't see how this could be healed, as it looks like chicken/egg to me | 12:52 |
noonedeadpunk | So the best I could find is to `update nova.services set version = 57;` .... | 13:12 |
frickler | noonedeadpunk: IIUC "folks" is neutral, yes. no idea about your exception, might be a bug not to be able to clear that | 13:18 |
noonedeadpunk | nah, it was more that I tried to do V -> X upgrade and then fix the result :D | 13:22 |
noonedeadpunk | So if conductor/scheduler being upgraded before computes, they just fail to startup with TooOldComputeService, as at the moment of start compute is indeed old. But when compute is being upgraded, it jsut has no way to report it's new state as everything crashed | 13:24 |
noonedeadpunk | Probably I should have run nova-status upgrade check before restart.... | 13:25 |
noonedeadpunk | but still I'm not sure if not having way to recover except db update is fine... | 13:28 |
noonedeadpunk | I wonder if compute should be just upgraded first... | 13:29 |
opendevreview | Gabriel Silva Trevisan proposed openstack/nova stable/ussuri: Reproduce bug 1953359 https://review.opendev.org/c/openstack/nova/+/822047 | 13:57 |
opendevreview | Gabriel Silva Trevisan proposed openstack/nova stable/ussuri: Extend the reproducer for 1953359 and 1952915 https://review.opendev.org/c/openstack/nova/+/822048 | 13:57 |
opendevreview | Gabriel Silva Trevisan proposed openstack/nova stable/ussuri: [rt] Apply migration context for incoming migrations https://review.opendev.org/c/openstack/nova/+/822050 | 13:57 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!