| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-10-19 | |||
| 02:58:39 | openstackgerrit | wangzhh proposed openstack/nova master: Ignore InstanceNotFound exception when host rejoin cluster https://review.opendev.org/757053 | |
| 07:27:11 | bauzas | good morning Nova | |
| 07:47:47 | xinranwang | gibi: Hi, Gibi. Could you and others please review the sriov nic support spec before PTG so we can discuss it during the meeting? https://review.opendev.org/#/c/742785/ | |
| 07:48:25 | xinranwang | Please feel free to leave the comment if any questions or suggestions. Thank you all in advance :) | |
| 07:54:43 | bauzas | review day for me indeed | |
| 08:10:31 | gibi | xinranwang: hi! I did review it two weeks ago and left comments | |
| 08:10:47 | gibi | bauzas: o/ | |
| 08:11:34 | xinranwang | gibi: oops, sorry for miss your comments. I will look into it soon. :) | |
| 08:30:20 | gibi | xinranwang: no worries | |
| 08:31:38 | stephenfin | bauzas: Seeing as I wasn't there Thursday afternoon, the suggestions was that we could use alembic for everything >= Train, allowing us to drop the unsupported/unloved sqlalchemy-migrate. However, that's a follow-up and probably needs a spec, for sure | |
| 08:32:13 | stephenfin | i.e. I wasn't suggesting introducing alembic as part of the compaction, just so we're clear :) | |
| 08:32:45 | bauzas | all good then | |
| 08:32:51 | bauzas | we can discuss it during next week | |
| 08:32:56 | stephenfin | yup | |
| 08:33:51 | gibi | stephenfin: I've approved the bp as per the meeting discussion | |
| 08:34:00 | stephenfin | gibi: Great! Thanks :) | |
| 08:34:03 | gibi | stephenfin: but left two comments in the bp whiteboard as requested | |
| 08:34:17 | stephenfin | Now to start generating schema diffs | |
| 08:34:24 | gibi | cool | |
| 08:34:50 | gibi | stephenfin: I spent some time on mypy and o.vo it is a tricky busines | |
| 08:35:10 | stephenfin | with the pyi approach? | |
| 08:35:13 | gibi | stephenfin: do you know a way to tell mypy that we have stub for a type but not for a whole file | |
| 08:35:16 | gibi | stephenfin: yes | |
| 08:35:31 | stephenfin | not off the top of my head, no | |
| 08:35:36 | gibi | stephenfin: for me it seems mypy handles pyi as per file :/ | |
| 08:36:09 | gibi | so if we generate pyi for the Instance o.vo in nova then mypy will not load the py file and does not check the member functions on Instance o.vo | |
| 08:37:10 | gibi | so now I looking at writing a plugin for mypy using the get_class_decorator_hook triggering on the oslo register decorator | |
| 08:37:53 | gibi | anyhow, it is interesting enough to spend some time on it :) | |
| 08:38:36 | stephenfin | I guess that's kind of greenfield? :) | |
| 08:38:46 | gibi | totally green :D | |
| 08:39:49 | gibi | this is how mypy sees an ovo definition http://paste.openstack.org/show/799163/ | |
| 08:40:51 | kashyap | stephenfin: Morning, what's the current Ubuntu LTS release, again? | |
| 08:41:02 | stephenfin | kashyap: In the gate? | |
| 08:41:08 | kashyap | stephenfin: Yeah, indeed | |
| 08:41:17 | kashyap | stephenfin: I'm trying to see what libvirt/QEMU version we test in the Gate | |
| 08:42:17 | stephenfin | I'm not sure if we're on Focal (20.04) yet. lyarwood is probably the guy to ask. We were on the version from the Bionic UCA (Cloud Archive) | |
| 08:42:37 | kashyap | (Okay, "Bionic Focal" is the latest LTS) | |
| 08:43:13 | stephenfin | Nope, Focal | |
| 08:43:15 | stephenfin | 20.04 | |
| 08:43:19 | stephenfin | Bionic is 18.04 | |
| 08:43:33 | stephenfin | (Focal Fossa and Bionic Beaver, fwiw) | |
| 08:43:45 | lyarwood | yup we should be using Focal in everything but the intergrated-compute job iirc | |
| 08:44:00 | lyarwood | https://review.opendev.org/#/c/755525/ will move that once the retry fix lands on stable/victoria | |
| 08:44:29 | kashyap | stephenfin: Err, I mean Bionic Beaver I meant. Sorry | |
| 08:44:45 | kashyap | Thanks, both. So whatever is in Focal, right? | |
| 08:44:53 | stephenfin | gibi: looks about right; I was attempting to work around it by doing something like https://review.opendev.org/#/c/738020/6/nova/objects/agent.py but that's just a hack, really | |
| 08:44:57 | lyarwood | https://review.opendev.org/#/c/757305/ - elod, bauzas, would you mind kicking this today? | |
| 08:45:13 | lyarwood | kashyap: yeah | |
| 08:45:46 | bauzas | lyarwood: can look, but I'd like to prioritize specs reviews | |
| 08:46:38 | lyarwood | kashyap: libvirt 6.0.0 and QEMU 4.2-3 | |
| 08:46:45 | lyarwood | bauzas: ack | |
| 08:47:03 | kashyap | lyarwood: Thank you; someone from the upstream libvirt/QEMU was asking the versions | |
| 08:47:32 | kashyap | lyarwood: And general "do you (OpenStack) folks do any grab-the-latest-release-of-libvirt/QEMU-to-catch-regressions kind of testing"? | |
| 08:47:37 | gibi | stephenfin: doesn't that field definitions conflict with the generated fields runtime? | |
| 08:47:55 | kashyap | lyarwood: The answer to the above is: we've got 'virt-preview' at most CI, but not testing from latest upstream releases, right? | |
| 08:48:10 | gibi | stephenfin: or these are no-op runtime? | |
| 08:48:25 | kashyap | (Although, I remember, some 4 years ago we wanted to allow testing tar releases of libvirt/QEMU in DevStack) | |
| 08:50:27 | lyarwood | kashyap: our virt-preview jobs are in the experimental queue at the moment, we might move them to the main queue as non-voting at some point this cycle if they continue to be stable enough | |
| 08:50:40 | kashyap | (Oh, right, experimental queue) | |
| 08:51:00 | lyarwood | kashyap: between them and our fedora jobs we should get enough coverage | |
| 08:51:02 | kashyap | lyarwood: What's the diff. between 'non-voting' and 'experimental'? One would think they're the same | |
| 08:51:12 | lyarwood | kashyap: we've already caught a number of issues early | |
| 08:51:18 | kashyap | (Right) | |
| 08:51:38 | lyarwood | kashyap: experimental is only used when someone leaves a `check experimental` comment | |
| 08:51:48 | lyarwood | kashyap: non-voting always runs automatically | |
| 08:51:53 | lyarwood | kashyap: but doesn't vote | |
| 08:51:58 | kashyap | lyarwood: Ah, thank you | |
| 08:57:02 | CeeMac | lyarwood: regarding the fix that merged recently https://review.opendev.org/#/c/754695/ | |
| 08:57:24 | CeeMac | is this compatible with rocky as is, or would it need a backport? | |
| 09:01:53 | lyarwood | CeeMac: I'm not sure if the cherry pick is clean back to stable/rocky directly from master | |
| 09:02:00 | lyarwood | CeeMac: I'll post stable backports now | |
| 09:02:08 | lyarwood | just sorting some admin stuff out downstream | |
| 09:03:31 | CeeMac | thanks lyarwood , no rush, whenever you have free time | |
| 09:05:15 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/victoria: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758730 | |
| 09:05:38 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/ussuri: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758732 | |
| 09:05:56 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/train: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758733 | |
| 09:06:25 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/stein: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758734 | |
| 09:07:00 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/rocky: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758735 | |
| 09:07:02 | lyarwood | CeeMac: ^ it's clean | |
| 09:07:22 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/queens: compute: Lock by instance.uuid lock during swap_volume https://review.opendev.org/758736 | |
| 09:07:39 | CeeMac | lyarwood: amazing! thanks again for your efforts, they're really appreciated | |
| 09:22:03 | kashyap | lyarwood: Any reason why I don't see the non-voting Fedora job mentioned in the list of CI jobs on your master change above? - https://review.opendev.org/#/c/754695/ | |
| 09:23:20 | lyarwood | kashyap: that's also experimental at the moment, sorry if I wasn't clear before | |
| 09:23:39 | lyarwood | kashyap: https://github.com/openstack/nova/blob/20ca7c4337adb58d9290521fb20803d789f2b574/.zuul.yaml#L530-L533 | |
| 09:23:46 | kashyap | lyarwood: Ah, okay; I misread as: "we have Fedora 'stable' non-voting; but 'virt-preview' experimental" | |
| 09:24:03 | lyarwood | yup my bad | |
| 09:24:28 | kashyap | lyarwood: No problem; the chat with you helped | |
| 09:24:52 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/victoria: Handle disabled CPU features to fix live migration failures https://review.opendev.org/758760 | |
| 09:25:20 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/ussuri: Handle disabled CPU features to fix live migration failures https://review.opendev.org/758761 | |
| 09:26:14 | openstackgerrit | Stephen Finucane proposed openstack/nova stable/train: Handle disabled CPU features to fix live migration failures https://review.opendev.org/758763 | |
| 09:27:00 | stephenfin | lyarwood: think you could close out https://review.opendev.org/#/q/topic:sriov-functional-tests+status:open today? sean-k-mooney has signed off on those changes to the NeutronFixture now | |
| 09:27:18 | stephenfin | or this week, it's not really urgent tbf :) | |
| 09:27:31 | lyarwood | yarp | |
| 09:27:35 | stephenfin | ta | |
| 09:27:41 | lyarwood | sorry keeps dropping off the end of my queue | |
| 09:40:02 | kashyap | gibi: stephenfin: Thx for that quick review of that BBC's fix: https://review.opendev.org/#/c/757577 | |
| 09:40:46 | kashyap | stephenfin: (Was on PTO last week, so I only saw it now). I'm mildly curious, will my +1 take effect in the end, while it is in "Needs Verified Label"? | |
| 09:41:00 | kashyap | ["it" == the change] | |
| 09:42:44 | stephenfin | kashyap: It'll register until it's merged, yeah | |
| 09:42:52 | kashyap | I see | |