| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-10-04 | |||
| 16:56:56 | opendevreview | ribaudr proposed openstack/nova master: Support rescuing an instance with shares (compute and API part) https://review.opendev.org/c/openstack/nova/+/860288 | |
| 18:41:32 | opendevreview | Merged openstack/os-vif stable/xena: Use TCP keepalives for ovsdb connections https://review.opendev.org/c/openstack/os-vif/+/841771 | |
| 18:59:57 | opendevreview | Merged openstack/os-vif stable/xena: only register tables used by os-vif https://review.opendev.org/c/openstack/os-vif/+/841772 | |
| #openstack-nova - 2022-10-05 | |||
| 04:58:20 | opendevreview | Merged openstack/nova master: Remove the periodic Centos 8 job https://review.opendev.org/c/openstack/nova/+/858272 | |
| 08:13:44 | opendevreview | ribaudr proposed openstack/nova master: Default Nova persistent objects without soft delete. https://review.opendev.org/c/openstack/nova/+/854355 | |
| 08:22:20 | Uggla | stephenfin, if you have some time can you have a look at https://review.opendev.org/c/openstack/openstacksdk/+/853949 | |
| 08:55:39 | Uggla | gibi, bauzas , Hi so if you can have a look at https://review.opendev.org/c/openstack/nova/+/854355. | |
| 08:56:11 | bauzas | Uggla: I'll try, a bit busy today due to some late stuff | |
| 08:57:03 | Uggla | gibi, if you are motivated, then you can continue the review of manila/virtiofs all patches are under this new topic bp/manila_shares_attachments_v2 | |
| 08:58:02 | Uggla | bauzas, I know, there is no hurry. | |
| 08:58:08 | opendevreview | Sahid Orentino Ferdjaoui proposed openstack/nova master: compute: enhance compute evacuate instance to support target state https://review.opendev.org/c/openstack/nova/+/858383 | |
| 08:58:08 | opendevreview | Sahid Orentino Ferdjaoui proposed openstack/nova master: api: extend evacuate instance to support target state https://review.opendev.org/c/openstack/nova/+/858384 | |
| 09:37:22 | bauzas | Uggla: so, I'll look at your patch, but I'll also like to have dansmith to dent it https://review.opendev.org/c/openstack/nova/+/854355 | |
| 09:46:14 | gibi | Uggla: yes. I would like to keep reviewing the manila work. It is a bit hard for me to balance my time at the moment. So I would like to ask you to be peristent asking for review :) I won't be offended if you keep pinging me all the time with the manial stuff :) | |
| 10:05:48 | gibi | dansmith: if you have time when you are up we need your ovo experties here https://review.opendev.org/c/openstack/nova/+/854355/3..4#message-90730ecb520b41d974a8d64c7f7cd0bb0f56356c | |
| 10:06:10 | gibi | Uggla: ^^ I see the problem but I don't see the imediate solution | |
| 10:06:22 | gibi | and I don't like duplicating created_at | |
| 10:07:07 | gibi | It might turn out that what I want is not possible with ovo | |
| 10:12:40 | sean-k-mooney | i feel like what you want should be possible | |
| 10:12:47 | sean-k-mooney | but i have not looked at the review properly | |
| 10:13:06 | sean-k-mooney | i might pull it locally quickly and play with it for a bit and see if we can drop those fields | |
| 11:33:04 | opendevreview | Stephen Finucane proposed openstack/nova master: conf: Add '[database] enable_soft_delete' flag https://review.opendev.org/c/openstack/nova/+/860401 | |
| 11:34:01 | stephenfin | I've been thinking about that for years 👆 Maybe it belongs in oslo.db but the idea is what's relevant. Would welcome input (particularly melwitt) | |
| 11:51:10 | sean-k-mooney | stephenfin: you were not in some of our interal team calls but we were discussign coudl we move to remvoing shaddow tables | |
| 11:51:59 | sean-k-mooney | soft delete i dont have a probalem with honestly | |
| 11:52:04 | sean-k-mooney | i dont like shadown tables | |
| 11:52:28 | sean-k-mooney | my only concern with the conf approch is it change api behavior | |
| 11:52:45 | sean-k-mooney | specificly for soft deleteing instance as we can use the restore api action to undelete them | |
| 11:53:20 | sean-k-mooney | there is no api impact to remvoe shadow tables however since once we have archive the rows there is no going back | |
| 11:53:28 | sean-k-mooney | so that seam less in vasive to me | |
| 11:54:29 | sean-k-mooney | stephenfin: but yes removing softdelete is someting to consider. we could for example use the soft delete timout as the config option instead | |
| 11:54:39 | sean-k-mooney | so if its not enable dthen just delete thing fully | |
| 11:55:08 | sean-k-mooney | if we had already disabeld shadow tabels or remvoe them i think we could do that | |
| 11:55:16 | stephenfin | sean-k-mooney: Aren't those different things? | |
| 11:55:32 | sean-k-mooney | there are 3 things | |
| 11:55:33 | stephenfin | soft-deleting an instance vs soft-deleteable tables | |
| 11:56:03 | sean-k-mooney | soft deleteing an instnace marks the row as deelte but does not actully delete it until the soft delete timeout expires | |
| 11:56:12 | stephenfin | I don't think it does | |
| 11:56:50 | sean-k-mooney | i woudl have to check but as far as im aware it at least updates teh state so that it nolonger shows up in insntace list | |
| 11:56:58 | sean-k-mooney | unless you add --deleted | |
| 11:57:18 | sean-k-mooney | if it truely is independed and not marking it as delete in the db | |
| 11:57:32 | stephenfin | https://github.com/openstack/nova/blob/master/nova/compute/manager.py#L3363-L3377 | |
| 11:57:44 | sean-k-mooney | then yes the 3 things (shadown tabels, soft deletable instance and soft deletable rows) would be independent | |
| 11:58:12 | stephenfin | it seems to be setting power_state, vm_state and task_state fields. It's not setting 'deleted' or 'deleted_at' | |
| 11:58:14 | sean-k-mooney | ok so its via a vm state | |
| 11:58:21 | stephenfin | yup | |
| 11:58:28 | stephenfin | which this doesn't touch | |
| 11:58:32 | sean-k-mooney | ok then ya no api impact so | |
| 11:58:44 | sean-k-mooney | add it as a ptg topic | |
| 11:59:53 | sean-k-mooney | so long term i like to remvoe nova-manage acive_delete_rows and make pruge act like purge in other project | |
| 12:00:12 | sean-k-mooney | in other projects purge it deletes the soft deleted rows | |
| 12:00:43 | sean-k-mooney | but if we can also remvoe those (optionally or eventually permently) i woudl be happy to do that too | |
| 12:01:36 | sean-k-mooney | stephenfin: i think i have mentioned this to you before but i woudl prefer if we mvoed to recommendign peole use https://github.com/ovh/osarchiver/ | |
| 12:02:54 | sean-k-mooney | so if they have audit usecases they woudl enable soft-delete but use osarchiver to move them to there archival storage and out of our db | |
| 12:03:37 | sean-k-mooney | if you can do that we dont need shadow tabels. if you dont have that usecase then you can disabel the soft-delete feature | |
| 12:04:02 | sean-k-mooney | and upstream in B or C we could defautl to disabling it. | |
| 12:05:58 | stephenfin | added to the agenda | |
| 12:08:09 | sean-k-mooney | cool im goign to summeries my toughts on the gerrit commit | |
| 12:16:55 | sean-k-mooney | ok done https://review.opendev.org/c/openstack/nova/+/860401/1#message-6ff04f04d84a33ff65f08e6c3956ea194a2a915c hopefully that makes sense | |
| 12:24:52 | stephenfin | lgtm | |
| 12:43:23 | sean-k-mooney | i like the idea of doing this in oslo too by the way but ya either could work | |
| 15:32:26 | stephenfin | Uggla: I figured out what was wrong with https://review.opendev.org/c/openstack/nova/+/854355/ Do you mind if I push my changes? | |
| 15:32:31 | stephenfin | I'm leaving a comment now too | |
| 15:33:14 | Uggla | stephenfin, cool, no pb. | |
| 15:33:48 | Uggla | stephenfin, btw did you have time to the openstacksdk patch ? | |
| 15:34:01 | stephenfin | Yeah, I think I +2d it | |
| 15:34:10 | Uggla | \o/ | |
| 15:40:44 | opendevreview | Stephen Finucane proposed openstack/nova master: objects: Add NovaSoftDeleteObject mixin https://review.opendev.org/c/openstack/nova/+/854355 | |
| 15:40:50 | stephenfin | Uggla: lmk what you think ^ | |
| 15:46:14 | Uggla | stephenfin, i'll have a look after our meeting. | |
| 15:47:23 | opendevreview | Stephen Finucane proposed openstack/nova master: rpc: Mark attributes as private https://review.opendev.org/c/openstack/nova/+/792803 | |
| 15:55:59 | melwitt | kashyap: the "abort live migration if monitoring fails" patch was to fail in a proper way when the error is encountered, there is another patch that needs review that will do the actual ignoring of the particular error https://review.opendev.org/c/openstack/nova/+/852002 | |
| 15:59:50 | melwitt | kashyap: there was an issue with the patch I had proposed to workaround it, so I abandoned it. ^ is the new one from another contributor | |
| 16:02:04 | Uggla | stephenfin, what you did in https://review.opendev.org/c/openstack/nova/+/854355/4..5, sounds good to me. Thank you. Now let's check if gibi, dansmith, bauzas agree. | |
| 16:03:11 | bauzas | I thought we said we haven't wanted to have API tables to be soft-deletable | |
| 16:03:36 | bauzas | but, we haven't said "yeah, we should deprecate the other tables" | |
| 16:05:16 | dansmith | yeah | |
| 16:05:28 | dansmith | I don't agree with the use of "deprecated" here | |
| 16:05:42 | dansmith | "not recommended for everything by default" makes sense | |
| 16:07:01 | bauzas | at least I'm afraid of saying "we deprecate instance record soft-deletion" | |
| 16:07:17 | dansmith | was there some decision to deprecate and actually remove this stuff? because I think I disagree with that | |
| 16:07:23 | dansmith | and if not, we should change the wording in the patch I think | |
| 16:11:34 | bauzas | Uggla: I looked at your patch | |
| 16:11:51 | kashyap | melwitt: Thanks for jogging my memory! I now recall | |
| 16:12:01 | bauzas | sounds quite good to me if you say 'we need to formally name which tables do softdelete" | |
| 16:12:16 | bauzas | which is what you code | |
| 16:12:19 | kashyap | melwitt: I thought this one from Brett has already merged...but apparently not yet. Is it waiting on something still? | |
| 16:12:43 | melwitt | kashyap: just needs a second reviewer | |
| 16:13:00 | kashyap | Ah, nod. I thought something else besides it. | |
| 16:13:02 | melwitt | I already +2ed it | |
| 16:13:09 | melwitt | nah | |
| 16:14:01 | kashyap | gibi: or any other core who's not Mel, can you please put this through? - https://review.opendev.org/c/openstack/nova/+/852002 | |
| 16:14:48 | kashyap | melwitt: Also thank you for - https://review.opendev.org/c/openstack/nova/+/859358/1 | |
| 16:15:44 | melwitt | :) | |
| 16:15:52 | kashyap | Often these unit tests take a ton of time (at least for me), and I keep duking around them | |
| 16:15:57 | gibi | kashyap: I added to my queue but no promises when I get to it | |
| 16:16:59 | kashyap | gibi: What? I thought you'd attach a promiese-to-be-executed-on-this-date to all your reviews! | |
| 16:17:20 | melwitt | kashyap: they take a ton of time for me, pretty much never goes smoothly 😆 | |
| 16:18:06 | kashyap | melwitt: Good to know; I feel particularly low when a unit test that I'm struggling with takes so long that a hen will develop teeth, but the test won't come out right. | |
| 16:19:59 | melwitt | kashyap: "hen develop teeth" haha I've never heard that before | |