| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-09-16 | |||
| 09:40:14 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP zuul: Introduce nova-multinode-evacuate https://review.opendev.org/744883 | |
| 09:51:15 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP zuul: Introduce nova-multinode-evacuate https://review.opendev.org/744883 | |
| 09:56:30 | openstackgerrit | Merged openstack/nova master: Allow tap interface with multiqueue https://review.opendev.org/748533 | |
| 10:04:24 | brinzhang_ | stephenfin: about https://review.opendev.org/#/c/752133/1, you mean, we should change 500 to 400? | |
| 10:04:29 | stephenfin | yes | |
| 10:04:34 | stephenfin | gimme a sec, I have a patch | |
| 10:04:48 | brinzhang_ | yeah | |
| 10:14:29 | openstackgerrit | Stephen Finucane proposed openstack/nova master: api: Set min, maxItems for server_group.policies field https://review.opendev.org/752211 | |
| 10:14:29 | openstack | bug 1894966 in OpenStack Compute (nova) "Create servergroup failed with unexpected error" [Low,In progress] https://launchpad.net/bugs/1894966 - Assigned to Brin Zhang (zhangbailin) | |
| 10:14:29 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Add regression test for bug 1894966 https://review.opendev.org/752210 | |
| 10:14:35 | stephenfin | brinzhang_: gibi: ^ | |
| 10:14:46 | stephenfin | That approach makes more sense to me, at least | |
| 10:16:34 | gibi | stephenfin: as I stated above I'm OK with both way, just documenting, or changing the return code to 400. gmann felt it is better just to document it | |
| 10:18:04 | gibi | gmann: ^^ | |
| 10:18:50 | stephenfin | I consider this a better form of documentation | |
| 10:19:27 | stephenfin | In that the API will tell the user what they did wrong, as opposed to yielding some cryptic error that they must go Google for | |
| 10:19:31 | stephenfin | My 2c :) | |
| 10:19:48 | gibi | make sense | |
| 10:19:58 | gibi | I added gmann to the review | |
| 10:20:37 | brinzhang_ | it's ok, is it need to docs that the changes of this API? | |
| 10:22:36 | brinzhang_ | This may break the user's perception of existing create server_group API. | |
| 10:22:53 | brinzhang_ | but it's ok for me too. | |
| 10:30:05 | stephenfin | brinzhang_: Resolving a HTTP 5xx error does not require a microversion | |
| 10:30:25 | stephenfin | The idea being that a user shouldn't have to opt-in to non-broken behaviour | |
| 10:32:01 | openstackgerrit | Stephen Finucane proposed openstack/nova master: functional: Add and use 'GlanceFixture' https://review.opendev.org/745907 | |
| 10:32:01 | openstackgerrit | Stephen Finucane proposed openstack/nova master: functional: Enable real policy fixture by default https://review.opendev.org/745906 | |
| 10:32:02 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Remove '_FakeImageService' https://review.opendev.org/745908 | |
| 10:32:18 | stephenfin | bauzas: Fixed the failing test in that first patch ^ | |
| 10:32:36 | sean-k-mooney | its the scond branch in the flow chard https://docs.openstack.org/nova/pike/contributor/microversions.html | |
| 10:32:40 | brinzhang_ | stephenfin: I know, which way is ok forme | |
| 10:32:42 | stephenfin | I'm eager to get it merged because things keep changing under my feet /o\ | |
| 10:32:43 | bauzas | stephenfin: will look at those later today | |
| 10:32:48 | bauzas | did a bug scrub | |
| 10:32:50 | stephenfin | thanks | |
| 10:34:28 | brinzhang_ | From the requirements of the manual, changing the API response requires adding microversion | |
| 10:34:57 | gibi | bauzas: can I take the bug lock? | |
| 10:35:01 | gibi | thanks for scrubbing | |
| 10:38:59 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Remove 'NovaObjectDictCompat' from 'Migration' https://review.opendev.org/723572 | |
| 10:39:00 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Remove 'NovaObjectDictCompat' from 'InstancePCIRequest' https://review.opendev.org/723573 | |
| 11:47:17 | nightmare_unreal | what can be the cause of tempest failures ? is it related to user patch or infrastructure/zuul issue ? | |
| 11:50:21 | sean-k-mooney | nightmare_unreal: both | |
| 11:50:35 | sean-k-mooney | normlly user patches | |
| 11:50:44 | sean-k-mooney | the jobs can fail for infra reasons | |
| 11:50:53 | sean-k-mooney | but then that normlaly happens before tempest runs | |
| 11:50:59 | gibi | stephenfin: one small question / request in https://review.opendev.org/#/c/745907/9/nova/tests/functional/regressions/test_bug_1675570.py@131 | |
| 11:51:18 | sean-k-mooney | in rare ocation tempets test will fail because of infra if for example the iops of the disk are too low | |
| 11:51:57 | sean-k-mooney | nightmare_unreal: but normally tempest failures are because fo the use patch. unless its a POST_FAILURE which si almost always infra related | |
| 11:52:04 | nightmare_unreal | sean-k-mooney: in my case osc-tx-py36-tips, tempest-full-py3, osc-functional-devstack and osc-functional tips failed | |
| 11:52:42 | sean-k-mooney | nightmare_unreal: which patch? | |
| 11:52:49 | nightmare_unreal | https://review.opendev.org/#/c/742210/ | |
| 11:52:54 | nightmare_unreal | status is post_failure | |
| 11:53:03 | nightmare_unreal | not for tempest | |
| 11:53:04 | sean-k-mooney | ya so post_failure | |
| 11:53:07 | sean-k-mooney | is not your problem | |
| 11:53:21 | sean-k-mooney | that means it failed in one of the post jobs like uploading the logs | |
| 11:53:35 | sean-k-mooney | that is an infra issue and you can just recheck | |
| 11:53:38 | nightmare_unreal | Oh, thanks . will recheck help | |
| 11:53:42 | nightmare_unreal | ah okay | |
| 11:53:44 | sean-k-mooney | the tempst run is only 16m long | |
| 11:53:47 | nightmare_unreal | thanks :) | |
| 11:53:47 | gibi | stephenfin: I'm totally OK to do an hard coded uuid GlanceFixture.imageX reference replacemnet in a separate patch | |
| 11:53:48 | sean-k-mooney | let me check that | |
| 11:53:55 | sean-k-mooney | but that looks like it failed to deploy | |
| 11:54:30 | sean-k-mooney | ERROR: No matching distribution found for oslo.service===2.4.0 (from -c /opt/stack/requirements/upper-constraints.txt (line 306)) | |
| 11:54:47 | sean-k-mooney | nightmare_unreal: so the tempest job failed because it could not install oslo.service | |
| 11:54:56 | sean-k-mooney | that is likely because our mirrors where out of sync | |
| 11:55:03 | nightmare_unreal | oh got it | |
| 11:55:08 | sean-k-mooney | which is also an infa issue | |
| 11:55:24 | nightmare_unreal | I see | |
| 11:55:29 | sean-k-mooney | so i think we can recheck this but good to ask infra if the issue is fixed first | |
| 11:55:35 | nightmare_unreal | I guess I will wait for some time | |
| 11:55:37 | nightmare_unreal | and then recheck | |
| 11:55:46 | nightmare_unreal | okay | |
| 11:56:00 | sean-k-mooney | the zuul run was yesterday | |
| 11:56:04 | nightmare_unreal | i will ask in that channel | |
| 11:56:24 | sean-k-mooney | so i think it shoudl be ok gibi do you know if the mirror issue we hit yesterday is fixed | |
| 11:56:51 | frickler | nightmare_unreal: sean-k-mooney: that was an issue with pypi and should be fixed since around midnight | |
| 11:57:12 | nightmare_unreal | alright :) . I will recheck in that case | |
| 11:57:15 | nightmare_unreal | thanks everyone | |
| 11:57:25 | frickler | we only proxy and cache requests to pypi, no mirroring | |
| 11:57:44 | sean-k-mooney | frickler: ah the mirros are for the distro packages only | |
| 11:58:03 | sean-k-mooney | i tought we have devpi running on the afs file system too | |
| 11:58:54 | frickler | sean-k-mooney: yes, we stopped running bandersnatch when hit the 2T limit a long time ago. yesterday we learning that pypi's own mirror filled up 12T and got stuck with that a month ago | |
| 11:59:22 | sean-k-mooney | ah i see | |
| 11:59:45 | sean-k-mooney | 12TB im guessing it does not prune old verions regularly | |
| 12:00:39 | frickler | well for a full mirror they likely need to keep all versions | |
| 12:00:40 | sean-k-mooney | that or there is a lot of software reguarlly asking for old verions of stuff | |
| 12:00:50 | sean-k-mooney | ah right | |
| 12:01:06 | sean-k-mooney | ya i like devpi because it was a caching proxy for pypi | |
| 12:01:22 | sean-k-mooney | as well as a local pip comparitble data store for your own packages | |
| 12:01:34 | sean-k-mooney | frickler: are we just using squid or something now as a http cache | |
| 12:02:40 | sean-k-mooney | i havent redeployed my onw cache since i moved house and got gigabit but its still vaguly on my todo list after i redeploy my testing third party ci. | |
| 12:05:24 | frickler | we use apache2 because it does both the proxying and serve our mirrored repos at the same time and on the same host, just with different paths | |
| 12:10:05 | sean-k-mooney | ah i might have to look at how that is configured and maybe steal it for my own use at somepoint. having all of infras confic public as a referce is quite helpful that way | |
| 13:06:37 | bauzas | gibi: sorry missed your ping | |
| 13:06:40 | bauzas | you surely can | |
| 13:06:58 | bauzas | gibi: but I'll do some scrubbing too, so lemme know which ones you're looking | |
| 13:16:41 | openstack | Launchpad bug 1881944 in OpenStack Compute (nova) "nova-api returns empty block-device-mapping in metadata queries" [Undecided,New] | |
| 13:16:41 | gibi | bauzas: looking at https://bugs.launchpad.net/nova/+bug/1881944 right now and I think that will be my last for today | |
| 13:16:52 | gibi | nice work on the open bugs, thanks | |