| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-09-13 | |||
| 16:04:26 | sean-k-mooney | gibi i assume we will backprot the eventlet fix in oslo | |
| 16:04:28 | sean-k-mooney | to stable | |
| 16:04:41 | sean-k-mooney | we cant bump mins on stable | |
| 16:04:53 | sean-k-mooney | but we need to backport and ensure its allwoed by upper constratis | |
| 16:05:01 | sean-k-mooney | or we need to patch the lock in nova | |
| 16:05:05 | sean-k-mooney | and backprot that | |
| 16:05:13 | gibi | I assume we backport the fix in oslo | |
| 16:05:21 | sean-k-mooney | that would be my prefernce too | |
| 16:05:22 | gibi | but I did not have time to propose that | |
| 16:05:37 | bauzas | ok, at least I propose to set the bug to High | |
| 16:05:47 | bauzas | the gate isn't longer impact, our distros are tho | |
| 16:05:49 | sean-k-mooney | ack | |
| 16:05:54 | bauzas | impacted* | |
| 16:05:56 | sean-k-mooney | well its cloased for master | |
| 16:06:02 | sean-k-mooney | and high for the older brnaches | |
| 16:06:16 | bauzas | not in the Nova project | |
| 16:06:32 | bauzas | I can set a branch series for the bug | |
| 16:06:34 | sean-k-mooney | you want to keep it up for the min version bump | |
| 16:06:35 | bauzas | lemme do it | |
| 16:06:39 | bauzas | correct | |
| 16:06:39 | gibi | ack | |
| 16:06:40 | bauzas | so, | |
| 16:06:55 | bauzas | set to High, open a Yoga series to the bug report and set it to high too | |
| 16:07:06 | bauzas | and ask for the bump in master and the backport in Yoga | |
| 16:07:08 | sean-k-mooney | +1 | |
| 16:07:08 | bauzas | wfy ? | |
| 16:07:23 | bauzas | ok, let me do the paperwork | |
| 16:07:33 | sean-k-mooney | well backport in oslo for yoga but yes | |
| 16:07:48 | gibi | looks OK | |
| 16:08:31 | bauzas | done | |
| 16:08:39 | bauzas | moving on | |
| 16:08:57 | bauzas | #link https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New 5 new untriaged bugs (-4 since the last meeting) | |
| 16:09:05 | bauzas | kudos to Uggla for this excellent work | |
| 16:09:10 | bauzas | #link https://storyboard.openstack.org/#!/project/openstack/placement 26 open stories (+0 since the last meeting) in Storyboard for Placement | |
| 16:09:16 | bauzas | #info Add yourself in the team bug roster if you want to help https://etherpad.opendev.org/p/nova-bug-triage-roster | |
| 16:09:24 | bauzas | Uggla: do you want to discuss any bug ? | |
| 16:09:41 | Uggla | bauzas, yes | |
| 16:10:18 | Uggla | bauzas, https://bugs.launchpad.net/nova/+bug/1896617 seems valid for me. I have triaged it. | |
| 16:10:40 | opendevreview | Balazs Gibizer proposed openstack/nova master: Bump min oslo.concurrencty to >= 5.0.1 https://review.opendev.org/c/openstack/nova/+/857491 | |
| 16:10:49 | Uggla | But I would like your opinion | |
| 16:11:49 | bauzas | shit | |
| 16:12:07 | bauzas | yeah, if we set the perms | |
| 16:14:16 | sean-k-mooney | hum | |
| 16:14:25 | sean-k-mooney | so right now this sound like an rfe to me | |
| 16:15:06 | sean-k-mooney | or at least its a new restriction how what nova ia and is not allwoed to assume | |
| 16:15:24 | sean-k-mooney | normally we assume that libvirt and nova can share a group | |
| 16:16:37 | sean-k-mooney | we do not document that you can harden in this way | |
| 16:16:44 | sean-k-mooney | so to me its not explcitly supported | |
| 16:17:55 | bauzas | sorry, had a doorbell | |
| 16:18:00 | bauzas | I'm back | |
| 16:18:28 | bauzas | yeah, so maybe the bug is valid as a Wishlist ? | |
| 16:19:15 | bauzas | sean-k-mooney: agreed ? | |
| 16:19:57 | sean-k-mooney | i dont think its a bug | |
| 16:20:05 | sean-k-mooney | we could have it be whishlist yes | |
| 16:20:11 | sean-k-mooney | as a mini feature | |
| 16:20:14 | sean-k-mooney | im writign a responce | |
| 16:20:20 | sean-k-mooney | but i think the aswer is to | |
| 16:20:21 | bauzas | cool, can we move on then ? | |
| 16:20:31 | sean-k-mooney | chown nova:$libvirt-group | |
| 16:20:44 | sean-k-mooney | and then libvirt gets read via the group | |
| 16:20:49 | bauzas | if you really think this is a feature, we can set it to Invalid/Wishlist and ask for a blueprint to be filled | |
| 16:20:59 | bauzas | but that's tough for reporters | |
| 16:21:36 | bauzas | sean-k-mooney: yeah, ideally the group shall be libvirt but this sounds distro-oriented | |
| 16:21:39 | sean-k-mooney | ya i think so to as its is not a documented deployment model we support | |
| 16:21:50 | bauzas | problem is, we set perms | |
| 16:21:54 | sean-k-mooney | well it the "libvirt group" but that is disto dependet | |
| 16:21:56 | bauzas | so we're opiniated | |
| 16:22:06 | sean-k-mooney | we can have that be a config option | |
| 16:22:13 | bauzas | sean-k-mooney: yeah that's my point, this is distro-specific | |
| 16:22:31 | bauzas | I honestly don't like us to be opinionated about the perms we shall set | |
| 16:22:43 | bauzas | so, yeah a config knob seems maybe the best | |
| 16:22:51 | bauzas | or... I dunno | |
| 16:23:22 | bauzas | shall we just assume the tempdir is readable by anyone ? | |
| 16:23:28 | sean-k-mooney | we have some requirment around this for vhost-user by the way | |
| 16:23:37 | bauzas | not by *anyone* | |
| 16:24:09 | bauzas | but at least should we just document this is distro-specific and us having a requirement that the directory you set in nova.conf has right perms accordingly? | |
| 16:24:30 | bauzas | and remove this chmod in the code | |
| 16:24:45 | bauzas | this would require an upgrade relnote so a spec tho | |
| 16:24:45 | sean-k-mooney | well maybe not | |
| 16:25:03 | sean-k-mooney | i would have to look at the code but we can take this offline after the meeting | |
| 16:25:08 | bauzas | I don't like the idea of nova managing the underlying OS | |
| 16:25:15 | bauzas | cool | |
| 16:25:18 | Uggla | if I'm not wrong we are setting o+r, I don't really get the diff if we add o+g too. | |
| 16:25:48 | bauzas | Uggla: the point is that I don't see why Nova should do it | |
| 16:26:20 | sean-k-mooney | bauzas: | |
| 16:26:21 | Uggla | but it is already in the driver.py. | |
| 16:26:22 | bauzas | shalln't we just assume the directory are correctly readable by the required users ? | |
| 16:26:35 | sean-k-mooney | we assume tha tnova is in the libvirt group so that we can read the disk to upload the snapshot | |
| 16:26:58 | sean-k-mooney | in there case nova si not in the libvirt-qemu group i think | |
| 16:27:26 | sean-k-mooney | althoguh they apprently worked around this by addign libvirt-qemu user ot nova group | |
| 16:27:29 | sean-k-mooney | which is backwards | |
| 16:28:27 | bauzas | let's put this bug as Opinion then | |
| 16:28:42 | bauzas | this is half distro-specific, half a nova problem | |
| 16:29:10 | bauzas | and people wanting to discuss the resolution are more than happy to engage later on after the meeting | |
| 16:29:27 | bauzas | time flies and we're at the half of the meeting | |
| 16:29:39 | bauzas | agreed ? | |
| 16:29:54 | sean-k-mooney | ack | |
| 16:30:15 | Uggla | ack | |
| 16:31:08 | bauzas | moving on | |
| 16:31:09 | gibi | works for me | |