| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-11-22 | |||
| 16:37:25 | bauzas | I discussed with jsanemet about his efforts and I proposed him to present himself and discuss about how we could paper the privsep effort during the meeting | |
| 16:37:40 | bauzas | the spec thing was more a wonder we had | |
| 16:37:52 | bauzas | I don't honestly feel it requires one spec | |
| 16:38:06 | bauzas | but we tho need to consider documenting the new context usages | |
| 16:38:12 | bauzas | hence the etherpad | |
| 16:38:52 | jsanemet | hello | |
| 16:38:53 | bauzas | I don't see any upgrade impact, or any other impact which would require us to accept a spedc | |
| 16:39:38 | bauzas | anyone disagree with that plan ? | |
| 16:40:21 | dansmith | what is the effort? just finishing some things? | |
| 16:40:28 | dansmith | or something more complex? | |
| 16:40:29 | bauzas | dansmith: https://etherpad.opendev.org/p/nova-privsep-review | |
| 16:40:42 | bauzas | dansmith: this is about fine-graining the privsep context to the right callers | |
| 16:40:43 | sean-k-mooney | there are several | |
| 16:40:58 | sean-k-mooney | this looks like the first low haning fruit | |
| 16:40:59 | dansmith | yeah so splitting into different contexts? | |
| 16:41:06 | sean-k-mooney | yep | |
| 16:41:08 | bauzas | as a first step yes | |
| 16:41:21 | dansmith | I dunno, I think that's going to be a bunch of review of which calls need which contexts right? | |
| 16:41:31 | bauzas | that's my concern | |
| 16:41:35 | bauzas | I don't think we need a spec | |
| 16:41:35 | dansmith | some things might seem simple but require both net and sys, etc | |
| 16:41:38 | bauzas | but, | |
| 16:42:01 | dansmith | I guess I'm not sure why no spec | |
| 16:42:03 | bauzas | I'd appreciate if we would agree on the contexts and the callers not exacly by reviewing the patches | |
| 16:42:07 | dansmith | this etherpad is half a spec already | |
| 16:42:15 | sean-k-mooney | there proably should be a spec to define the scope | |
| 16:42:21 | dansmith | agree | |
| 16:42:27 | bauzas | there, we have our answer | |
| 16:42:37 | bauzas | we can use the spec format to agree on the split | |
| 16:42:57 | sean-k-mooney | sure | |
| 16:43:30 | bauzas | jsanemet: you're ok with the outcome ? | |
| 16:43:39 | bauzas | I can help you on the process-y thingie | |
| 16:44:05 | jsanemet | sure | |
| 16:44:22 | jsanemet | then, i will translate the etherpad into a spec and we can discuss it further | |
| 16:44:26 | jsanemet | seems good to me | |
| 16:44:33 | bauzas | from a paperwork pov, you may also need to supersede the rfe bug by creating a launchpad blueprint | |
| 16:44:48 | bauzas | but that's an easy peasy | |
| 16:45:01 | bauzas | anyway, looks like we have the direction | |
| 16:45:06 | jsanemet | yes, i am a little green regarding what documentation is required so any help will be appreciated | |
| 16:45:39 | bauzas | jsanemet: as a reminder, you have a spec template https://specs.openstack.org/openstack/nova-specs/specs/2023.1/template.html | |
| 16:46:10 | bauzas | I guess that's it for today | |
| 16:46:11 | jsanemet | all right, i will save that | |
| 16:46:15 | jsanemet | thanks a lot | |
| 16:46:36 | bauzas | ok, folks, any other item to raise before we call the wrap ? | |
| 16:46:58 | bauzas | looks not | |
| 16:47:01 | bauzas | perfect timing | |
| 16:47:05 | bauzas | thanks all | |
| 16:47:08 | opendevmeet | Log: https://meetings.opendev.org/meetings/nova/2022/nova.2022-11-22-16.01.log.html | |
| 16:47:08 | opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/nova/2022/nova.2022-11-22-16.01.txt | |
| 16:47:08 | opendevmeet | Minutes: https://meetings.opendev.org/meetings/nova/2022/nova.2022-11-22-16.01.html | |
| 16:47:08 | opendevmeet | Meeting ended Tue Nov 22 16:47:08 2022 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | |
| 16:47:08 | bauzas | #endmeeting | |
| 16:47:12 | gibi | :) | |
| 16:47:13 | elodilles | thanks bauzas o/ | |
| 16:47:26 | gibi | thanks folks | |
| 16:47:37 | jsanemet | thanks o/ | |
| 17:36:21 | gmann | sean-k-mooney: bauzas: replied on focal new job https://review.opendev.org/c/openstack/nova/+/861111/5/.zuul.yaml#653 | |
| 17:36:52 | gmann | basically grenade job run only smoke test not all the tests. having a integrated job running on focal cover it well | |
| 17:37:42 | sean-k-mooney | gmann: im not sure we want full coverage since we are only supprotign it for upgrade but we could i guess | |
| 17:38:04 | sean-k-mooney | we run more then smoke in the multinode grenade job by the way | |
| 17:38:12 | sean-k-mooney | liek cold/live migration | |
| 17:38:14 | gmann | that was point there in PTG, to make sure we run the complete tests | |
| 17:38:43 | sean-k-mooney | that kind fo feels like overkill to me | |
| 17:38:52 | gmann | that is why we added Focal i testing runtime too https://governance.openstack.org/tc/reference/runtimes/2023.1.html | |
| 17:38:59 | sean-k-mooney | we never did that for moving form 18.04->20.04 did we | |
| 17:39:20 | gmann | sean-k-mooney: this is new change we discussed in TC | |
| 17:39:36 | gmann | https://review.opendev.org/c/openstack/governance/+/860599 | |
| 17:39:46 | sean-k-mooney | ok but if its just for upgrade im not sure we shoudl do more then run multinode grenade | |
| 17:40:49 | gmann | multinode grenade does not run complete tests, if testing runtime have requirement of testing Focal I feel we should run integrated tests on that not just grenade small set of tests | |
| 17:41:09 | sean-k-mooney | i disagree | |
| 17:41:22 | sean-k-mooney | your are correct on the tests that are run https://github.com/openstack/nova/blob/master/.zuul.yaml#L512 | |
| 17:41:36 | sean-k-mooney | """In this release, we are adding the testing of Ubuntu new version 22.04. For smooth upgrade, we will continue the minimum testing for previously supported Ubuntu version in this release.""" | |
| 17:42:04 | sean-k-mooney | it specificaly say we will continue the minium testing for previous supported ubuntu verions | |
| 17:42:08 | sean-k-mooney | that is not tempest-full | |
| 17:42:31 | gmann | minimum testing means at least single job running the tests and we knew grenade job is there but adding this explicitly to make sure things does not get broken on old distro | |
| 17:42:43 | gmann | not all the jobs | |
| 17:43:04 | gmann | "run all jobs on latest distro, have at least one job to run on old distro" | |
| 17:43:28 | gmann | https://review.opendev.org/c/openstack/governance/+/860599/5/reference/project-testing-interface.rst#191 | |
| 17:43:40 | sean-k-mooney | are we going to have to do the same for debian 11 when 12 comes out | |
| 17:44:19 | gmann | we should do but we will see where all we are running debian jobs and there we should make sure there same | |
| 17:44:23 | gmann | the same | |
| 17:44:47 | sean-k-mooney | we dont have ci capsity for that | |
| 17:45:28 | sean-k-mooney | fine lets just stick with this for now but i dont think we can reasonable run all LTS distros for 2 releases | |
| 17:45:40 | sean-k-mooney | in the B cycle we shoudl drop the job | |
| 17:45:40 | gmann | debian jobs are not run on projects side (there might be few) but say if devstack run it then we should do both when we change the debian version in testing | |
| 17:46:04 | sean-k-mooney | and only test with 22.04 | |
| 17:46:12 | gmann | sean-k-mooney: true, that will be removed in next cycle of change in distro version, so for Focal yes in B | |
| 17:46:51 | gmann | it is only the cycle changing the testing to new version will make sure old version is also tested. | |
| 17:48:26 | gmann | we can do a lot of upgrade testing with distro things but we cannot do as per our CI capacity so this single job for changing version cycle is good way to accommodate something and make sure upgrade will be more smooth | |
| 18:18:27 | opendevreview | Merged openstack/nova stable/wallaby: [stable-only] Use os-brick from source in wallaby https://review.opendev.org/c/openstack/nova/+/865134 | |
| 19:44:57 | opendevreview | Ghanshyam proposed openstack/nova master: Update gate jobs as per the 2023.1 cycle testing runtime https://review.opendev.org/c/openstack/nova/+/861111 | |
| 19:47:10 | opendevreview | Ghanshyam proposed openstack/osc-placement master: Update gate jobs as per the 2023.1 cycle testing runtime https://review.opendev.org/c/openstack/osc-placement/+/861470 | |
| 19:48:15 | opendevreview | Ghanshyam proposed openstack/os-traits master: Update python classifier for python 3.10 https://review.opendev.org/c/openstack/os-traits/+/861466 | |
| 19:48:30 | opendevreview | Ghanshyam proposed openstack/python-novaclient master: Update python classifier for python 3.10 https://review.opendev.org/c/openstack/python-novaclient/+/861469 | |
| 19:54:31 | gmann | sean-k-mooney: os-vif functional sudo job failing on Jammy, I think you mentioned you have some idea to fix that?, https://zuul.opendev.org/t/openstack/build/b40811601a614064a480a02e368c2f72 | |
| 20:41:40 | sean-k-mooney | gmann: i looked at it breifly but did not try to fix it yet | |
| 20:42:15 | sean-k-mooney | gmann: debian used to have a pathced version of distutils that did weired things with the python path | |
| 20:42:35 | sean-k-mooney | the funtional tests use sudo and we use -E to preserve teh env | |
| 20:42:59 | sean-k-mooney | there is a weired interaction between sudo and virutalenv and -E | |
| 20:43:12 | sean-k-mooney | so my guess was that has changed on 22.04 | |