| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-09-17 | |||
| 13:31:58 | bauzas | but you could technically stick with diff (with -y, side-to-side) and just grep a few things | |
| 13:31:59 | stephenfin | sean-k-mooney: They won't accept it | |
| 13:32:00 | sean-k-mooney | i rarely use it but when i do it works well | |
| 13:32:02 | stephenfin | They being infra | |
| 13:32:10 | sean-k-mooney | stephenfin: why not | |
| 13:32:26 | stephenfin | Something about git-review being a dead simple tool that shouldn't grow any additional functionality | |
| 13:32:32 | stephenfin | which I kind of get | |
| 13:32:36 | bauzas | meld works too, I mean litterally any decent editor that supports merge diffs | |
| 13:32:39 | sean-k-mooney | it already has -m | |
| 13:32:46 | stephenfin | it probably shouldn't have that | |
| 13:32:48 | bauzas | -m has a very different behaviour | |
| 13:33:02 | sean-k-mooney | if you dont do the rebase in -m its the same | |
| 13:33:17 | bauzas | the 'if' makes the difference | |
| 13:33:32 | sean-k-mooney | as you said -m rebase both to master to elimitate the need to diff diffs | |
| 13:33:50 | sean-k-mooney | but both are trying to do the same thing | |
| 13:34:04 | bauzas | correct but then you have to handle merge conflicts if they occur | |
| 13:34:04 | sean-k-mooney | elimiate chagne that are not part of your actual patch | |
| 13:34:18 | bauzas | that's unnecessary work on my opinion | |
| 13:34:28 | sean-k-mooney | yes whcih is why i dont like how -m works | |
| 13:34:31 | sean-k-mooney | yep | |
| 13:35:16 | bauzas | anyway | |
| 13:35:25 | bauzas | just saying | |
| 13:36:02 | bauzas | this is a problem I've seen from 6 years ago, I bet someone would have come up with some stuff in git review if infra was happy with | |
| 13:36:12 | mriedem | gibi: ok comments in https://review.opendev.org/#/c/666857/ | |
| 13:36:45 | gibi | mriedem: looking | |
| 13:37:06 | stephenfin | sean-k-mooney: does glance have documentation for flavor extra specs? | |
| 13:37:09 | bauzas | snap, mriedem beated me | |
| 13:37:24 | mriedem | stephenfin: flavors != glance | |
| 13:37:31 | mriedem | you mean image properties? | |
| 13:37:40 | mriedem | https://docs.openstack.org/glance/latest/admin/useful-image-properties.html | |
| 13:37:50 | sean-k-mooney | stephenfin: it has metadefs for them | |
| 13:38:08 | stephenfin | mriedem: Nope, flavor extra specs. I know it's not their domain but sean-k-mooney had mentioned some about flavor definitions in glance before | |
| 13:38:09 | mriedem | it has metadefs for image properties | |
| 13:38:11 | mriedem | not flavor extra specs | |
| 13:38:11 | sean-k-mooney | and it has the image properties that mriedem linked | |
| 13:38:21 | sean-k-mooney | yes they do | |
| 13:38:53 | sean-k-mooney | the metadefs cover flavors image and host agrrates metatdata | |
| 13:39:17 | sean-k-mooney | https://github.com/openstack/glance/blob/master/etc/metadefs/compute-libvirt.json#L9-L14 | |
| 13:39:53 | sean-k-mooney | its where all the documenation in glance and heat comes form for those values | |
| 13:40:14 | sean-k-mooney | *horizon and heat | |
| 13:40:22 | mriedem | huh | |
| 13:40:40 | mriedem | i found https://github.com/openstack/glance/blob/master/etc/metadefs/compute-trust.json which we removed that filter from nova years ago...so it's good to know the information about this stuff is in glance and is stale | |
| 13:41:07 | sean-k-mooney | horizon and heat both autogenerat there ui from the glance metadef api | |
| 13:41:30 | sean-k-mooney | yes i have been slowly trying to update it every now and then | |
| 13:41:45 | sean-k-mooney | i would like to bring it fully up to date envtually | |
| 13:41:50 | stephenfin | IMO those definitions should live in nova and be thrown across the wall on a regular cadence to glance | |
| 13:42:11 | sean-k-mooney | they should live in a seperate project that is not part of either | |
| 13:42:16 | sean-k-mooney | but use by all services | |
| 13:42:52 | stephenfin | or we could just expose it via an API that glance could consume? | |
| 13:42:58 | sean-k-mooney | glance currentlly is our metadata defintion store/repositroy | |
| 13:43:10 | sean-k-mooney | glance already exposes this via an api | |
| 13:43:13 | sean-k-mooney | that others consume | |
| 13:43:22 | sean-k-mooney | we coudl move to all service exposing the same api | |
| 13:43:36 | openstackgerrit | Eric Fried proposed openstack/nova master: DNM: Try to repro bug 1823251 with mysql logs https://review.opendev.org/678051 | |
| 13:43:37 | openstack | bug 1823251 in OpenStack Compute (nova) "Spike in TestNovaMigrationsMySQL.test_walk_versions/test_innodb_tables failures since April 1 2019 on limestone-regionone" [High,Confirmed] https://launchpad.net/bugs/1823251 | |
| 13:43:57 | sean-k-mooney | e.g. have a /metadefs on each service but really this shoudl be its own serivice or stay in glance | |
| 13:45:08 | sean-k-mooney | haveing all serivice provide an identical api would only really work if we had a lib with the api defintion and each service just kept the data in the repo | |
| 13:48:02 | stephenfin | idk, clearly having our definitions live in another service is an issue | |
| 13:48:33 | sean-k-mooney | well we define image properties that belong to glance in our ovos | |
| 13:48:37 | stephenfin | and I'm not sure using an external library is an option since that means only official extra specs can be used | |
| 13:48:50 | sean-k-mooney | if you take that route we shoudl not be doing that either | |
| 13:49:09 | sean-k-mooney | stephenfin: no it doesnt | |
| 13:49:30 | sean-k-mooney | the metadefs can be exetended by operators or distors | |
| 13:49:47 | sean-k-mooney | you simply add entires to the files | |
| 13:49:56 | sean-k-mooney | or new files | |
| 13:50:33 | stephenfin | actually, we already figured this out | |
| 13:50:36 | sean-k-mooney | so we can take the same data driven approch or we can define a custom_ prefix or seveler other thigns | |
| 13:50:48 | stephenfin | stevedore | |
| 13:50:52 | sean-k-mooney | in the spec we figured out how to use stevedore | |
| 13:50:54 | sean-k-mooney | ya | |
| 13:51:03 | stephenfin | yeah, glance can just hook into that like we do | |
| 13:51:05 | stephenfin | easy as | |
| 13:51:17 | sean-k-mooney | not really | |
| 13:51:42 | sean-k-mooney | this is an api that has existed for 6+ years | |
| 13:51:54 | sean-k-mooney | and other pojects depend on it | |
| 13:52:01 | stephenfin | they can keep the API | |
| 13:52:18 | sean-k-mooney | yes but unless we can expose our flavor form it | |
| 13:53:03 | sean-k-mooney | *flavor extra specs via the new mechanisum at the metadefs endpoin im not sure going our own way is good | |
| 13:53:25 | stephenfin | It should be easy do | |
| 13:53:43 | sean-k-mooney | stevador only works if everythign is installed in teh same place | |
| 13:53:51 | sean-k-mooney | in contienr it wont work | |
| 13:54:11 | stephenfin | If they can pull our definitions in, they just need to transform them so they match whatever we had before | |
| 13:54:28 | stephenfin | then we provide a tool to dump in their JSON metadef style | |
| 13:54:35 | stephenfin | and push a patch to Glance | |
| 13:54:41 | sean-k-mooney | ya bout you would have have to install nova in the glance contienr | |
| 13:54:49 | sean-k-mooney | thats not good | |
| 13:55:01 | stephenfin | better than what we have at the momenty | |
| 13:55:06 | sean-k-mooney | havign a nova tool to generate teh metadefs and copy it could work | |
| 13:55:16 | stephenfin | oh, I thought you meant the tool | |
| 13:55:45 | sean-k-mooney | no i ment stevadore | |
| 13:56:05 | sean-k-mooney | stevador was a suggestion for other project to add validtors | |
| 13:56:11 | sean-k-mooney | to nova | |
| 13:56:20 | sean-k-mooney | not to share data to other projects | |
| 13:56:20 | stephenfin | Yup, gotcha. My bad | |
| 13:56:25 | mriedem | belmoreira: this might have an impact to compute start times when using the ironic driver https://review.opendev.org/#/c/666857/ | |
| 13:56:43 | mriedem | belmoreira: so if you have suggestions about what definitely not to do there it would be appreciated | |
| 13:56:54 | sean-k-mooney | it wouldnt be nova-mange but a nova-metadefs tool or something could be used to do an export i guess | |
| 13:57:16 | mriedem | belmoreira: specifically https://review.opendev.org/#/c/666857/9/nova/compute/manager.py@1329 | |
| 14:01:32 | belmoreira | mriedem thanks for ping me | |
| 14:02:38 | openstackgerrit | Eric Fried proposed openstack/nova master: Only allow one scheduler service in tests https://review.opendev.org/682486 | |