| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-11-27 | |||
| 13:46:43 | elod | ACK | |
| 14:55:47 | stephenfin | lyarwood, elod: In theory, we could now +2 +W all stable backports and they won't be able to merge because of dansmith's cherry-pick test, right? | |
| 14:56:08 | elod | stephenfin: why? O.o | |
| 14:56:44 | elod | how do you mean? | |
| 14:56:58 | stephenfin | take this change (thanks for the reviews, btw :)) https://review.opendev.org/c/openstack/nova/+/764455 | |
| 14:57:20 | stephenfin | that's for stable/train but I've got (identical) cherry-picks all the way back to stable/queens | |
| 14:58:00 | stephenfin | so in theory, if you were reviewing one, you could in theory review the others at the same time, and they'd have to merge in series because the pep8 job would fail otherwise | |
| 14:58:12 | stephenfin | *they'd have to merge in order | |
| 14:58:33 | elod | oh, so you mean we could +2+W in advance back till queens in this case | |
| 14:58:52 | stephenfin | Yup, exactly. That's much clearer :) | |
| 14:59:33 | elod | yeah, I guess that should work :] | |
| 15:00:32 | elod | I'm thinking now on cases where "somehowe-something-changes" but we don't realize and just recheck and patch merges... | |
| 15:02:12 | elod | hmmm. in case of a change in newer branch, the hash changes, so in that case it won't be a problem, on older branches the patches simply does not merge due to the same cherry-pick check | |
| 15:02:21 | elod | so that should work too | |
| 15:03:13 | stephenfin | yup, that what I'm thinking anyway | |
| 15:04:06 | elod | we just have to be careful that the hash of patch on branch N+1 is really there in patch on branch N | |
| 15:04:37 | stephenfin | well I assume the test does that | |
| 15:04:52 | stephenfin | what we need to validate is that the hash corresponds to what we think it does (as opposed to an arbitrary patch) | |
| 15:05:17 | elod | no, it just checks whether the listed hashes are merged somewhere | |
| 15:05:41 | gmann | lyarwood: humm that seems like image deleted with server seems, debugging | |
| 15:05:58 | stephenfin | oh, so you could in theory cherry-pick staight from master to e.g. stable/queens and it would pass | |
| 15:06:03 | stephenfin | okay, yeah, that's one to watch for | |
| 15:06:18 | stephenfin | but once we do that...! :D | |
| 15:09:32 | elod | :) well, we could try that... ;) | |
| 15:11:16 | openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Revert "zuul: Skip test_attach_scsi_disk_with_config_drive in nova-ceph-multistore" https://review.opendev.org/c/openstack/nova/+/764311 | |
| 15:48:00 | lyarwood | /o\ sorry ssh timed out and I didn't notice | |
| 15:48:29 | lyarwood | gmann: yeah that's odd | |
| 15:49:18 | gmann | lyarwood: or we are not passing the image_id for delete. testing changes in above revert, it should work now | |
| 15:50:00 | gmann | now i am passing image_id and image should be found and deleted correctly | |
| 15:51:38 | lyarwood | gmann: kk | |
| 15:52:09 | lyarwood | stephenfin: https://review.opendev.org/c/openstack/nova/+/764246 - would you mind taking a swing at that before you call it a day? | |
| 15:52:31 | stephenfin | sure | |
| 15:52:48 | lyarwood | https://review.opendev.org/c/openstack/nova/+/743220 is also trival | |
| 15:53:38 | gmann | gibi: as placement is under Nova now, this enable to notify the placement changes on nova channel too - https://review.opendev.org/c/openstack/project-config/+/764469 | |
| 15:54:07 | gmann | as next step we can probably merge both channel (delete placement channel basically ) | |
| 15:54:54 | gmann | while pushing few policy patches in placement i realized there are not bot to nova channel | |
| 15:56:26 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: zuul: Replace nova-live-migration with zuulv3 jobs https://review.opendev.org/c/openstack/nova/+/764432 | |
| 15:57:28 | lyarwood | elod / bauzas ; okay the ceph change is now squashed into that change so we hopefully have a passing gate again on stable/victoria | |
| 15:57:41 | bauzas | all good | |
| 15:57:48 | lyarwood | actually pep8 will fail until https://review.opendev.org/c/openstack/nova/+/762623 lands FWIW | |
| 15:58:09 | lyarwood | so bauzas / gibi / stephenfin ; ^ reviews on that would be appreciated | |
| 15:58:15 | bauzas | ack | |
| 15:58:49 | gibi | gmann: ohh, so the governance patch has been merged. cool. yes I agree that we shoudl redirect the placement channel traffic to nova | |
| 16:01:12 | gmann | gibi: ok, I will send email to request infra admin to do that. it can be done via some redirect way I think. | |
| 16:01:25 | gibi | gmann: thanks! | |
| 16:07:13 | gmann | gibi: sent, may be you can confirm there and I will follow up with infra team - http://lists.openstack.org/pipermail/openstack-discuss/2020-November/019092.html | |
| 16:07:27 | gibi | gmann: ack | |
| 16:28:39 | stephenfin | lyarwood: very soft -1 but I think it's worth thinking about before we proceed https://review.opendev.org/c/openstack/nova/+/764246 | |
| 16:28:58 | stephenfin | I'm pretty sure we've discussed this exact question before, in fact | |
| 16:30:00 | lyarwood | stephenfin: ack looking | |
| 16:31:04 | lyarwood | stephenfin: yeah we have, I might turn that into a hacking check of some kind | |
| 16:32:31 | stephenfin | I'm thinking the same thing. A modification of H904 perhaps | |
| 16:33:53 | gibi | stephenfin: good catch on non the delayed LOG interpolation. | |
| 16:34:18 | openstackgerrit | Lee Yarwood proposed openstack/nova master: libvirt: Skip encryption metadata lookups if secret already exists on host https://review.opendev.org/c/openstack/nova/+/764246 | |
| 16:35:32 | lyarwood | stephenfin: ack yeah we should add f-strings to that check | |
| 16:49:06 | lyarwood | lol so just found a wonderful gerrit bug | |
| 16:49:24 | lyarwood | if you're writing a comment don't leave your cursor anywhere outside of the textbox | |
| 16:49:59 | lyarwood | otherwise while text does end up in the box the ui will also react to any shortcuts you happen to be hitting at the sametime | |
| 16:50:07 | lyarwood | on google-chrome at least | |
| 16:52:07 | sean-k-mooney | ya that sometimes happened before | |
| 16:52:14 | sean-k-mooney | i never really figured out why | |
| 16:52:24 | sean-k-mooney | its not new to the new version | |
| 16:52:37 | sean-k-mooney | you coudl work around it by disabling fast rendering | |
| 16:52:38 | lyarwood | ah kk, first time I've seen it | |
| 16:52:46 | sean-k-mooney | but that is not an option now | |
| 17:15:49 | openstackgerrit | Elod Illes proposed openstack/nova stable/pike: Validate id as integer for os-aggregates https://review.opendev.org/c/openstack/nova/+/764479 | |
| 17:42:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: utils: Sanitize purely numeric TLD in hostnames https://review.opendev.org/c/openstack/nova/+/764482 | |
| 17:43:18 | stephenfin | artom: sean-k-mooney: Have at it ^ | |
| 17:43:33 | stephenfin | I think that's the most sensible solution of the poor selection available to us | |
| 17:44:02 | stephenfin | :) | |
| 17:44:31 | artom | stephenfin, yeah, a new spec/API is too much work for something like this | |
| 17:44:32 | lyarwood | incoming | |
| 17:46:23 | openstackgerrit | Lee Yarwood proposed openstack/nova master: Add encryption fields to BlockDeviceMapping object and table https://review.opendev.org/c/openstack/nova/+/760453 | |
| 17:46:24 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP virt: Add ephemeral encryption flag https://review.opendev.org/c/openstack/nova/+/760455 | |
| 17:46:24 | openstackgerrit | Lee Yarwood proposed openstack/nova master: image_meta: Add ephemeral encryption properties https://review.opendev.org/c/openstack/nova/+/760454 | |
| 17:46:25 | openstackgerrit | Lee Yarwood proposed openstack/nova master: func: Allow compute_driver to be set and used by _IntegratedTestBase https://review.opendev.org/c/openstack/nova/+/764484 | |
| 17:46:25 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP scheduler: Add an ephemeral encryption pre filter https://review.opendev.org/c/openstack/nova/+/760456 | |
| 17:46:26 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP compute: Update bdms with ephemeral encryption details when requested https://review.opendev.org/c/openstack/nova/+/764486 | |
| 17:46:26 | openstackgerrit | Lee Yarwood proposed openstack/nova master: BlockDeviceMapping: Add is_local property https://review.opendev.org/c/openstack/nova/+/764485 | |
| 17:56:38 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Improve docs for '_filter_*' functions https://review.opendev.org/c/openstack/nova/+/764376 | |
| 17:56:39 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Improve testing of 'nova.pci.request' module https://review.opendev.org/c/openstack/nova/+/764446 | |
| 17:56:39 | openstackgerrit | Stephen Finucane proposed openstack/nova master: pci: Add logging for filtering https://review.opendev.org/c/openstack/nova/+/764377 | |
| 18:01:08 | sean-k-mooney | stephenfin: we should jsut block anything with a . in it | |
| 18:04:54 | sean-k-mooney | stephenfin: if you want to allow FQDNS you shoudl write a spec and add a new microverion. this has implecatoin for cloud init and anything that assumes the hostname is a host name soe im not ok with adding support for hostnames as a bug fix | |
| 18:05:08 | sean-k-mooney | *support for FQDNs | |
| 18:13:20 | artom | Yeah, if it were up to me, I'd add a new microversion that did 2 things: 1. remove all the sanitization that we do, and just fail with a 400 if we get a bad hostname and 2. decouples VM names from dns_name, adding a separate argument for the latter | |
| 18:13:37 | artom | Piling on more obscure magic doesn't seem like the way forward | |
| 18:16:49 | artom | Wait, is instance.hostname exposed anywhere? | |
| 18:16:56 | artom | IOW, what do we actually use it for? | |
| 18:17:05 | sean-k-mooney | we use it for cloud init | |
| 18:17:10 | sean-k-mooney | its in the metadata | |
| 18:17:29 | sean-k-mooney | how your vm end up being called the same as the server name | |
| 18:17:54 | sean-k-mooney | but the vm also has a display name which can be modifed after and that is what we show in the cli/ui | |
| 18:18:03 | sean-k-mooney | the hostname cannot be changed | |
| 18:18:13 | sean-k-mooney | at lease i dont think it can | |
| 18:19:58 | artom | Ah, so the hostname isn't API-visible | |
| 18:20:05 | artom | OK, this mellows me a bit | |
| 18:24:07 | sean-k-mooney | artom: it is | |
| 18:24:17 | sean-k-mooney | its in the server details reponce | |
| 18:24:23 | sean-k-mooney | i think its just readonly once set | |