| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-06-29 | |||
| 14:17:44 | stephenfin | Yeah /o\ I considered changing it when rebasing and decided not to for some reason. Happy to change if it's not just me | |
| 14:18:50 | dansmith | I stared at "false is not 4" for a few moments when I broke the test on purpose... :) | |
| 14:19:12 | stephenfin | bauzas: Mentioned this Friday but it was a bit late. Care to take a look at https://review.opendev.org/#/c/706013/ when you've time? | |
| 14:20:54 | bauzas | stephenfin: sure I can try | |
| 14:22:13 | bauzas | stephenfin: humpf, I think you can't do this https://review.opendev.org/#/c/706013/6/nova/objects/migration.py | |
| 14:22:20 | bauzas | dansmith: ^ | |
| 14:22:44 | bauzas | stephenfin: if you want to change an object field, you can't just change its type directly | |
| 14:23:05 | bauzas | you need to provide another field and, | |
| 14:23:06 | stephenfin | bauzas: dansmith looked at it in the past. It's kosher. The serialized objects look identical, and the validation works as it did | |
| 14:23:18 | bauzas | you need to depracate the other | |
| 14:23:23 | bauzas | deprecate* | |
| 14:23:35 | bauzas | hmmm, ok | |
| 14:24:01 | stephenfin | There would be an issue if I was changing from e.g. StringField to ObjectField or IntegerField, but MigrationTypeField is an EnumField | |
| 14:25:03 | dansmith | bauzas: I haven't looked at what he's offering, but in the past, if we've converted the field type from string to enum and the enum has every possible historical value in it, we've allowed it | |
| 14:25:36 | dansmith | i.e. as long as it won't break existing clients.. the field type doesn't go over the wire, just the assumption that it's de-serializable by the type on the remote side | |
| 14:25:37 | bauzas | well, now I understand | |
| 14:25:46 | bauzas | yeah, stephenfin explained it | |
| 14:26:03 | dansmith | bauzas: I know, but if you're me you wouldn't take stephenfin's word for it, so.. :D | |
| 14:26:20 | dansmith | hence, I assume, the name drop above | |
| 14:26:29 | bauzas | because when deserializing the new object, then the old compute could still be able to create its object | |
| 14:26:48 | bauzas | dansmith: haha | |
| 14:26:51 | dansmith | bauzas: yeah | |
| 14:26:56 | stephenfin | dansmith: correct. Validation through association | |
| 14:27:06 | dansmith | bauzas: I think we've done worse things than string->enum even :) | |
| 14:27:06 | bauzas | okay, I'll provide a comment then | |
| 14:27:15 | bauzas | just to make sure people understand why I'm accepting it | |
| 14:27:31 | dansmith | does it have a code comment about the type changin? | |
| 14:27:40 | dansmith | if not, you could/should -1 probably and demand it | |
| 14:27:58 | dansmith | just so people know that older clients could be less strict | |
| 14:28:06 | stephenfin | The type isn't changing | |
| 14:28:19 | dansmith | it's going from string to enum right? | |
| 14:28:32 | stephenfin | Nope. It was an EnumField with 4 allowed values. It's still an EnumField with four values | |
| 14:28:35 | stephenfin | https://review.opendev.org/#/c/706013/6/nova/objects/migration.py | |
| 14:29:00 | stephenfin | I've just making it a custom enum field so I have constants I can reference | |
| 14:29:01 | dansmith | oh, even less of a thing then, nevermind | |
| 14:29:19 | dansmith | yeah that's definitely not visible to RPC, so whatever | |
| 14:30:46 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/train: Guard against missing image cache directory https://review.opendev.org/738455 | |
| 14:31:15 | bauzas | stephenfin: dansmith: yeah, it's just changing the EnumField to be a specific one | |
| 14:31:40 | bauzas | but, tbc, I don't want to see other changes just changing types without thinking about this | |
| 14:32:30 | bauzas | stephenfin: fwiw, I'd also have loved if you could have cut this change in two and not adding two properties as well in the same | |
| 14:32:42 | bauzas | given this wasn't needed | |
| 14:33:14 | stephenfin | It wouldn't really help though | |
| 14:33:37 | bauzas | anyway, reviewing it | |
| 14:33:56 | stephenfin | You'd save about four lines in the first patch (the properties), and then you'd have to review a second patch that's reworks virtually everything you changed in the first one | |
| 14:54:36 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: DNM: Test the state of VMware NSX 3pp CI https://review.opendev.org/734114 | |
| 15:17:45 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/train: Guard against missing image cache directory https://review.opendev.org/738455 | |
| 17:21:18 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Warn at controller start if there are older than N-1 computes https://review.opendev.org/738482 | |
| 17:38:08 | openstackgerrit | Merged openstack/nova master: objects: Add MigrationTypeField https://review.opendev.org/706013 | |
| #openstack-nova - 2020-06-30 | |||
| 04:15:12 | openstack | bug 1862633 in OpenStack Compute (nova) "unshelve leak allocation if update port fails" [Medium,Fix released] https://launchpad.net/bugs/1862633 - Assigned to Balazs Gibizer (balazs-gibizer) | |
| 04:15:12 | openstackgerrit | Merged openstack/nova stable/queens: Reproduce bug 1862633 https://review.opendev.org/729539 | |
| 06:04:18 | openstackgerrit | Alexandre Arents proposed openstack/nova master: Limit the number of concurrent snapshots https://review.opendev.org/736169 | |
| 07:05:35 | aarents | Good morning Nova, | |
| 07:05:50 | aarents | basic gerrit workflow question , should I rebase this https://review.opendev.org/#/c/734776/ , because I repushed nits on https://review.opendev.org/#/c/736169 (first one is on top of 2nd) | |
| 07:14:18 | bauzas | good morning nova | |
| 07:14:21 | bauzas | aarents: /me looks | |
| 07:15:01 | bauzas | aarents: indeed but I can +W it | |
| 07:15:12 | bauzas | since stephenfin and dansmith accepted it | |
| 07:16:26 | aarents | bauzas: thanks, doing | |
| 07:16:36 | bauzas | aarents: ping me when you're done | |
| 07:16:51 | aarents | k | |
| 07:17:00 | gibi | aarents: yeah, in general you don't have to be affraid loosing +2 or +A. if it was there we will be put it back easily | |
| 07:17:43 | aarents | great | |
| 07:22:54 | openstackgerrit | Alexandre Arents proposed openstack/nova master: Snapshot: offload glance upload in a native thread https://review.opendev.org/734776 | |
| 08:52:07 | openstackgerrit | Wenping Song proposed openstack/nova master: Add instance project_id for cyborg arq https://review.opendev.org/738428 | |
| 10:29:49 | openstackgerrit | Huaqiang Wang proposed openstack/nova master: test_hardware: code formating for multiline construct https://review.opendev.org/728793 | |
| 11:08:36 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Update keypairs when saving an instance https://review.opendev.org/683043 | |
| 11:08:37 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Reset changes for nested fields https://review.opendev.org/738583 | |
| 12:00:14 | openstackgerrit | Jinsheng Zhang proposed openstack/nova master: Add API to support multiple boot-volumes with boot order selection https://review.opendev.org/737913 | |
| 12:01:16 | gibi | dansmith: I need your guidance in https://review.opendev.org/#/c/738482 (see comments inline) | |
| 13:03:30 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Rework how we check for neutron extensions https://review.opendev.org/705792 | |
| 13:05:21 | sean-k-mooney2 | stephenfin: ^ why | |
| 13:06:17 | sean-k-mooney2 | more or less that seams to just be renames but you have also mixed in some parmater changes | |
| 13:09:41 | stephenfin | the only parameter change is that I use the extension alias to do extension lookup rather than the name | |
| 13:09:54 | stephenfin | the alias is a slug, so that seems more correct IMO | |
| 13:10:17 | sean-k-mooney2 | you changed to useing the named args instead of postiional too but that is correct | |
| 13:11:38 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Use neutronclient's port binding APIs https://review.opendev.org/706295 | |
| 13:11:42 | sean-k-mooney2 | also most people wont be familar with the term slugs its a term more or less exclucivly used in web dev form old newpaper printing | |
| 13:12:05 | sean-k-mooney2 | that said i agree that using the stevedor alias is better then the name | |
| 13:12:31 | sean-k-mooney2 | the entry point alias is required to be stable the name is not | |
| 13:14:24 | sean-k-mooney2 | strictly speaking i think you could make the argument the extenion names could be translatable | |
| 13:31:17 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Don't use generic 'Field' container https://review.opendev.org/738239 | |
| 13:31:18 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Use imports instead of type aliases https://review.opendev.org/738018 | |
| 13:31:18 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Remove unnecessary type aliases, exceptions https://review.opendev.org/738240 | |
| 13:31:19 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Remove wrappers around ovo mixins https://review.opendev.org/738019 | |
| 13:31:20 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Add type hints https://review.opendev.org/738020 | |
| 13:36:48 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Add type hints https://review.opendev.org/738020 | |
| 13:43:59 | tacco | hi there.. anyone knows where the volumes_attached: information in "openstack server show" comes from? | |
| 13:44:31 | tacco | because i deleted a volume long time ago.. but it still exists in that particular place.. and thats why i cannot resize this instance.. but the volume was successfull deleted long time ago. | |
| 13:44:36 | tacco | any idea? | |
| 13:44:43 | tacco | tryed to fix this in the DB but no chance for nor. | |
| 13:44:50 | tacco | s/nor/now/ | |
| 13:49:40 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to 'nova.pci' https://review.opendev.org/676209 | |
| 13:49:41 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to top-level modules https://review.opendev.org/705658 | |
| 13:49:41 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add nova.cmd, nova.conf, nova.console https://review.opendev.org/705657 | |
| 13:53:26 | stephenfin | tacco: That's this code https://github.com/openstack/nova/blob/master/nova/api/openstack/compute/views/servers.py#L714-L733 | |
| 13:59:34 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Remove custom comparison methods https://review.opendev.org/472285 | |
| 14:03:37 | tacco | stephenfin: thanks a lot will have a look | |
| 14:19:58 | tacco | cannot find the right way.. trough.. all i can find is in cinder that the volume is marked deleted and the volume does not exists anymore. But in the volumes_attached it still apeares. So strange.. | |
| 14:22:14 | tacco | is there any way to force detatch a volume from the instance by nova? | |
| 14:50:13 | gibi | tacco: look into the block_device_mapping table of the nova cell database | |