| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-09-02 | |||
| 15:01:25 | sean-k-mooney | am do we want a quick DNM patch to hack in a full tempest run | |
| 15:01:44 | stephenfin | also, the version of QEMU and libvirt on Ubuntu 18.04 is too old | |
| 15:01:49 | sean-k-mooney | like we have done for numa before | |
| 15:01:56 | stephenfin | I had to use Fedora 31 for testing this locally | |
| 15:01:57 | lyarwood | *cough* Fedora *cough* | |
| 15:02:00 | sean-k-mooney | stephenfin: we are ment to be on ubuntu 20.04 | |
| 15:02:03 | sean-k-mooney | before m3 | |
| 15:02:11 | lyarwood | yeah good luck with that | |
| 15:02:17 | sean-k-mooney | its a comuntiy goal | |
| 15:02:25 | sean-k-mooney | and we really really should have done it already | |
| 15:02:34 | stephenfin | there a issues with some storage-related tests | |
| 15:02:42 | stephenfin | lyarwood has the specifics, iirc | |
| 15:02:47 | sean-k-mooney | that also affects cento8 and fedora | |
| 15:02:57 | sean-k-mooney | since they use teh same version of libvirt and qemu | |
| 15:03:02 | stephenfin | lyarwood: do we have a Fedora job I can hack on? | |
| 15:03:13 | sean-k-mooney | at least centos 8 and ubuntu 20.04 do | |
| 15:03:21 | sean-k-mooney | fedora might be newer | |
| 15:04:15 | stephenfin | lyarwood, sean-k-mooney: oh, more fun, cirros doesn't support TPM so we'd have to use different guest images too | |
| 15:04:15 | lyarwood | well I could at least reproduce it with bionic and the UCA | |
| 15:04:47 | lyarwood | stephenfin: >< okay well this is getting harder, I just wanted to ensure the file copying part worked tbh | |
| 15:05:24 | stephenfin | lyarwood: I can reproduce locally with DevStack master on Fedora 31, if that would help? Just go paste the results somewhere. | |
| 15:05:48 | stephenfin | Happy to get whitebox tests going post-M3 then. We'll need them at some point anyway | |
| 15:06:32 | lyarwood | stephenfin: yeah sure it's not worth holding this up for | |
| 15:08:07 | lyarwood | stephenfin: the -1 was for the commit nit btw not this testing | |
| 15:10:03 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add support for resize and cold migration of emulated TPM files https://review.opendev.org/639934 | |
| 15:10:20 | stephenfin | lyarwood: ta-da ^ | |
| 15:11:45 | jsuchome | lyarwood: adapted https://review.opendev.org/#/c/748212/ | |
| 15:12:51 | gibi | sean-k-mooney: as https://review.opendev.org/#/c/744816/ merged, will you propose an os-vif relase before the lib freeze? | |
| 15:13:14 | sean-k-mooney | am yes although i want to review two other patches today | |
| 15:13:25 | sean-k-mooney | ill propose a patch by eod | |
| 15:20:52 | stephenfin | gibi, sean-k-mooney: I can do it. Major or minor release, given the change in behavior? | |
| 15:20:59 | stephenfin | I'm tempted to say major | |
| 15:21:34 | sean-k-mooney | its minor we have not broke backwards compatblity | |
| 15:21:44 | gibi | sean-k-mooney, stephenfin: It is OK to have it done tomorrow, no rush. | |
| 15:22:18 | stephenfin | sean-k-mooney: cool | |
| 15:22:20 | sean-k-mooney | https://review.opendev.org/#/c/746673/ and https://review.opendev.org/#/c/745895/ are the two i want to re review | |
| 15:22:41 | stephenfin | gibi: It was a 1 minute job (I timed it :)) https://review.opendev.org/749535 | |
| 15:22:49 | stephenfin | releasing-as-code++ | |
| 15:22:51 | gibi | stephenfin: because you are fast :D | |
| 15:23:09 | sean-k-mooney | ya there is a tool you can run | |
| 15:23:17 | sean-k-mooney | that generate the patch for you | |
| 15:23:24 | gibi | stephenfin: thanks anyhow | |
| 15:23:27 | stephenfin | I know hberaud had one alright. Didn't know there was a generic one | |
| 15:23:39 | sean-k-mooney | its in the repo | |
| 15:23:43 | sean-k-mooney | you just use tox | |
| 15:23:47 | gibi | sean-k-mooney: hit me up with any of those patches if need a push on them before the freeze | |
| 15:24:09 | sean-k-mooney | i need to see if the previous comments have been adressed | |
| 15:24:27 | sean-k-mooney | it is a bug fix so we could backported anyway | |
| 15:25:55 | stephenfin | gibi: Could you carry your +2 from https://review.opendev.org/#/c/741995/10 to https://review.opendev.org/#/c/744958/8 ? They should be identical, commit message aside | |
| 15:25:58 | stephenfin | I had them separate to avoid intertwining two series, but that's not an issue now as the base patches for both have merged | |
| 15:26:16 | stephenfin | I'll abandon https://review.opendev.org/#/c/741995/10 now | |
| 15:27:20 | gibi | done | |
| 15:27:34 | stephenfin | thanks | |
| 15:27:35 | gibi | thanks for that ping, I forget about these duplications | |
| 15:29:03 | stephenfin | and while I have you (last one, promise) - this has two changes since your last +2: a change to some mocks to mock out the new "does this user/group exist" check, and a commit message nit from lyarwood https://review.opendev.org/#/c/639934/ | |
| 15:29:14 | stephenfin | if you could revisit and close out the vTPM series \o/ | |
| 15:29:17 | gibi | sure | |
| 15:29:19 | gibi | on it | |
| 15:29:28 | gibi | closing features before the FF week is nice :D | |
| 15:30:21 | gibi | and I'm planning for asking favors on the SRIOV series soon (as it is getting close to be ready)... | |
| 15:30:47 | stephenfin | fire 'em at me | |
| 15:31:17 | gibi | stephenfin: I will, working on functional tests now | |
| 15:33:12 | gibi | +2 | |
| 15:34:11 | sean-k-mooney | gibi: i havent reviewd the WIP patch fully although i tested it. are you ready to remove the WIP status from it | |
| 15:34:37 | sean-k-mooney | i do think its close too | |
| 15:35:07 | gibi | sean-k-mooney: from code perspective it is ready (and my next push will remove the WIP tag) from test perspective I will push the unit test coverage today, but the functional test needs another day | |
| 15:35:34 | sean-k-mooney | ok thanks ill start reviewing the code so. | |
| 15:35:56 | gibi | cool, thanks | |
| 15:47:43 | openstackgerrit | Merged openstack/nova-specs master: Correct some spell errors https://review.opendev.org/749449 | |
| 15:56:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: libvirt: Remove '[vnc] keymap', '[spice] keymap' options https://review.opendev.org/749056 | |
| 16:02:36 | openstackgerrit | Merged openstack/nova master: trivial: Use constants for libvirt version checks https://review.opendev.org/740921 | |
| 16:21:48 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Support SRIOV interface attach and detach https://review.opendev.org/740995 | |
| 16:22:11 | gibi | sean-k-mooney, stephenfin: unit test coverage is ready, and there is the first passing functional test as well ^^ | |
| 16:22:22 | gibi | I will continue with the functional tests tomorrow | |
| 16:22:49 | stephenfin | great | |
| 16:23:00 | stephenfin | can review in the AM too | |
| 16:40:04 | sean-k-mooney | stephenfin: speaking fo funcitonal tests what is the state of your sriov functional tests | |
| 16:40:08 | sean-k-mooney | are those merged | |
| 16:40:29 | sean-k-mooney | gibi: did you base your functional tests on stephens work? | |
| 16:40:52 | sean-k-mooney | gibi: i guess you are not doing live migration tests | |
| 16:40:55 | sean-k-mooney | so dont need to | |
| 16:41:31 | sean-k-mooney | it is conflicting with stephens patches however | |
| 18:15:54 | openstack | Launchpad bug 1893263 in OpenStack Compute (nova) "Cannot create instance with multiqueue image and vif_type=tap (calico)" [Undecided,In progress] - Assigned to Rodrigo Barbieri (rodrigo-barbieri2010) | |
| 18:15:54 | ganso | Hi nova folks! I know everyone is busy at the moment, but I'd like to ask for some input in https://bugs.launchpad.net/nova/+bug/1893263 , it is a blocker bug that I have already pushed a fix, would like someone's blessing on it. | |
| 18:40:25 | sean-k-mooney | ganso: you should be checking the vif_model | |
| 18:41:11 | sean-k-mooney | you are also checking the image property via the system metadata table rather then form the image | |
| 18:41:42 | sean-k-mooney | but yes that is where you would fix this but just not how you should fix this | |
| 18:42:21 | ganso | sean-k-mooney: you mean check vif_model to confirm the vif_type before setting multiqueue? | |
| 18:42:40 | sean-k-mooney | no check the vif_model to confirm its virtio | |
| 18:43:19 | sean-k-mooney | if you set hw_vif_model=e1000 then you cant use multiqueue | |
| 18:43:27 | sean-k-mooney | since its a virtio feature | |
| 18:43:35 | ganso | sean-k-mooney: oh I see! | |
| 18:44:21 | ganso | regarding the system metadata. I don't have access to the image in the vif.plug() method (or I am missing something) | |
| 18:45:22 | sean-k-mooney | you have acess via the instance | |
| 18:45:32 | sean-k-mooney | instance.image_meta | |
| 18:45:40 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/objects/instance.py#L250 | |
| 18:46:27 | ganso | oh I see! when I printed those fields I didn't get that, probably due to object property lazy loading | |
| 18:46:40 | ganso | sean-k-mooney: thanks, I will address those and push a new patch! | |
| 19:07:28 | sean-k-mooney | and yes the image_meta is a property on the instance its not stored in the instnace table directly so the proerty makes sure we correctly build an instance of the class meaning we get validation of the data | |
| 19:07:52 | sean-k-mooney | which is preferebale over parsing the raw data in the system metadta table | |