| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-01-29 | |||
| 11:45:51 | gibi | kashyap: on it | |
| 11:46:21 | kashyap | gibi: Thank you. I'm still quietly thinking through if I'm forgetting to consider any other aspects | |
| 11:46:39 | kashyap | If you have any questions, can respond here, or on the change | |
| 11:49:06 | gibi | kashyap: If I trust the commit message the the change looks good as it does what the commit message describes | |
| 11:49:43 | kashyap | gibi: Heh, noted. FWIW, two libvirt CPU modelling devs also suggested the same | |
| 11:49:54 | kashyap | So I'm not doing this haphazardly; as this has live migration implications | |
| 11:49:57 | gibi | we have some livemigration tests but I guess there the two node is symmetric | |
| 11:50:10 | kashyap | Yeah, we don't have an intergration test for this scenario in the Gate | |
| 11:50:23 | kashyap | gibi: gibi: So, I'm waiting on Red Hat upgrades testing folks to report back w/ the result | |
| 11:50:37 | kashyap | I gave them them patch to give it a try | |
| 11:50:49 | gibi | kashyap: OK, so RedHat will do some extra testing on it | |
| 11:50:57 | gibi | that sounds good to me | |
| 11:50:58 | kashyap | gibi: Also if you have time, please read the "Root cause anlaysis": https://bugs.launchpad.net/nova/+bug/1913716 | |
| 11:50:59 | openstack | Launchpad bug 1913716 in OpenStack Compute (nova) "Live-migrating an instance from 'Queens' (CentOS-7) to 'Train' (CentOS-8) fails during libvirt's compareCPU() check" [Undecided,New] | |
| 11:51:02 | kashyap | gibi: Thank you! | |
| 11:51:18 | gibi | I will leave a +1 and wait for the result of the extra testing | |
| 11:51:19 | kashyap | I filed that bug report, based on yesterday's analysis w/ DanPB | |
| 11:51:24 | kashyap | gibi: Yep, understood | |
| 11:52:28 | kashyap | gibi: There is another occurrence in the code where we use compareCPU(). I need to carefully review the code again to see if that also can be removed, or replace it w/ the newer API. As noted in the commit message | |
| 11:52:42 | gibi | ack | |
| 11:52:44 | kashyap | But that's a separate item; not mixing up here | |
| 11:52:48 | gibi | cool | |
| 12:01:17 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: libvirt: Define and emit DeviceRemovedEvent and DeviceRemovalFailedEvent https://review.opendev.org/c/openstack/nova/+/749929 | |
| 12:02:43 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: libvirt: add AsyncDeviceDetachEventsHandler https://review.opendev.org/c/openstack/nova/+/772381 | |
| 12:02:43 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: libvirt: allow querying devices from the persistent domain https://review.opendev.org/c/openstack/nova/+/772383 | |
| 12:03:29 | gibi | lyarwood: fixed your comment in https://review.opendev.org/c/openstack/nova/+/749929 | |
| 12:05:27 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: libvirt: parse alias out from device config https://review.opendev.org/c/openstack/nova/+/772384 | |
| 12:06:07 | sean-k-mooney | stephenfin: by the way you dont need to update teh git review url | |
| 12:07:02 | sean-k-mooney | stephenfin: on fedora 33 the issue with git-review is that gerrit does not implement the negociation to enable the ssh clinet to use rsa-sha2 | |
| 12:07:05 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Replace blind retry with libvirt event waiting in detach https://review.opendev.org/c/openstack/nova/+/770246 | |
| 12:07:16 | stephenfin | sean-k-mooney: that wasn't kashyap's issue | |
| 12:07:17 | sean-k-mooney | and on fedora 33 sha1 is disable | |
| 12:07:29 | kashyap | Yeah, as stephenfin says that wasn't my issue | |
| 12:07:42 | sean-k-mooney | oh you had review.openstack.org | |
| 12:07:48 | sean-k-mooney | not review.opendev.org | |
| 12:08:05 | kashyap | Right; but that doesn't needed updating; it auto-redirects and did the right thing until yesterday | |
| 12:08:36 | kashyap | My issue was something dumber: out of sync remotes. Just a `fetch` fixed it | |
| 12:08:47 | sean-k-mooney | ah | |
| 12:09:08 | sean-k-mooney | you should still update the url i have fond that the redirects are flaky | |
| 12:09:39 | sean-k-mooney | it works for https | |
| 12:09:51 | sean-k-mooney | havent hand the same look using ssh | |
| 12:10:22 | kashyap | (I did update, though, while at it) | |
| 12:45:06 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Replace blind retry with libvirt event waiting in detach https://review.opendev.org/c/openstack/nova/+/770246 | |
| 12:47:43 | openstackgerrit | Stephen Finucane proposed openstack/nova master: libvirt: Remove dead code https://review.opendev.org/c/openstack/nova/+/772928 | |
| 12:47:44 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to 'nova.pci' https://review.opendev.org/c/openstack/nova/+/676209 | |
| 12:47:45 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add nova.cmd, nova.conf, nova.console https://review.opendev.org/c/openstack/nova/+/705657 | |
| 12:47:45 | openstackgerrit | Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to top-level modules https://review.opendev.org/c/openstack/nova/+/705658 | |
| 12:47:46 | openstackgerrit | Stephen Finucane proposed openstack/nova master: trivial: Clean manager.Manager, service.Service signatures https://review.opendev.org/c/openstack/nova/+/764806 | |
| 12:47:46 | openstackgerrit | Stephen Finucane proposed openstack/nova master: WIP: Expand type hints for nova.block_device https://review.opendev.org/c/openstack/nova/+/742170 | |
| 12:54:14 | openstackgerrit | Merged openstack/nova master: Enable mypy on scheduler/utils.py https://review.opendev.org/c/openstack/nova/+/769942 | |
| 13:16:54 | kashyap | gibi: Nice split of your wait-for-libvirt-events patches ... it's easier to review | |
| 13:40:26 | gibi | thanks | |
| 14:26:42 | kashyap | And thanks for updating the bug metadata for me | |
| 14:29:08 | gibi | no problem. I hope at some point the automatic update will work again with the new gerrit | |
| 15:54:46 | stephenfin | gibi, bauzas: This patch has been around for a while. If you've a few minutes spare next week, could you chuck it onto your review queues? https://review.opendev.org/c/openstack/nova/+/741663/ | |
| 15:54:55 | stephenfin | The other ones would be nice too, but that would be a good start :) | |
| 15:55:20 | gibi | ohh notifier, nice | |
| 15:55:24 | gibi | I will look at it | |
| 15:55:33 | stephenfin | thanks :) | |
| 15:59:01 | bauzas | stephenfin: ack, will try | |
| 16:03:15 | brtknr | what are ccommon reasons for nova and cinder states going out of sync? | |
| 16:13:39 | sean-k-mooney | sometimes tokens can expire cause the update of cinder to not happen correctly | |
| 16:13:51 | sean-k-mooney | service_tokens solves that problem | |
| 16:14:14 | sean-k-mooney | the other common case is where you hit a bug during a move operation or in the revert of a move operation | |
| 16:15:44 | sean-k-mooney | where move opartion bugs are not caused by token issues tempory entwork issue with rabbitmq or the cinder api can cause reverts or updates to be lost which if they are not retried can lead to them being out of sync | |
| 16:16:10 | sean-k-mooney | i think we retry cinder calls up to 3 times by default to mitigate the tempory api outage case | |
| 16:17:50 | sean-k-mooney | stephenfin: by the way im not worried about delineation i just find [libvirt]/virt_type more pleaseing to read then [libvirt]virt_type | |
| 16:18:07 | stephenfin | what about '[libvirt] virt_type' | |
| 16:18:31 | sean-k-mooney | that i could be ok with if it was in quotes | |
| 16:18:46 | sean-k-mooney | btu in a sentance without the qutoes or backticks i think that coudl be more confusin | |
| 16:18:56 | stephenfin | That's fair | |
| 16:19:03 | sean-k-mooney | so [libvirt] virt_type is a littel confusing | |
| 16:19:31 | stephenfin | The one good thing that did come out of that discussion is that I spotted a lot of examples in the docs that were not using the ':oslo.config:option:' role | |
| 16:19:36 | sean-k-mooney | but '[libvirt] virt_type' is fine | |
| 16:19:37 | stephenfin | So I'm going to fix those | |
| 16:20:12 | sean-k-mooney | ya so one thing about that though is we cant use that in the release notes right | |
| 16:20:22 | stephenfin | no, we can't | |
| 16:20:23 | sean-k-mooney | so i have always used the / there | |
| 16:20:30 | sean-k-mooney | and in specs | |
| 16:21:04 | sean-k-mooney | i agree though were we can use the olso way we proably should | |
| 16:21:30 | stephenfin | for sure | |
| 16:21:31 | sean-k-mooney | on a different topic | |
| 16:21:39 | sean-k-mooney | have we remvoed libvirt/xen support | |
| 16:21:43 | sean-k-mooney | or just xenapi | |
| 16:21:46 | stephenfin | we have, yes | |
| 16:21:50 | stephenfin | that merged last week | |
| 16:21:57 | stephenfin | libvirt/UML is gone too | |
| 16:21:59 | sean-k-mooney | ah ok that was my quetion on https://review.opendev.org/c/openstack/nova/+/772928 | |
| 16:22:06 | sean-k-mooney | cool ill upgrate to a +1 then | |
| 16:22:18 | stephenfin | Yup, the Change-Id I linked is the patch that removed it | |
| 16:22:26 | sean-k-mooney | oh you called it out in the commit message missed that | |
| 16:22:33 | stephenfin | Yup :-) | |
| 16:23:14 | sean-k-mooney | if a commit message is really short i dont always read them for short patchs | |
| 16:23:28 | sean-k-mooney | if the patch is anyway long i do but in this case it would have helped | |
| 16:23:47 | stephenfin | you mean you don't read the literary marvels that are my commit messages | |
| 16:23:50 | stephenfin | I'm disappointed :P | |
| 16:24:44 | sean-k-mooney | stephenfin: if you right an easy in them i alwasy do | |
| 16:25:00 | sean-k-mooney | but if its 4 lines long for a 32 line patch not normally | |
| 16:25:39 | sean-k-mooney | ill try and take a look at your mypy stuff next week by the way | |
| 16:26:00 | sean-k-mooney | but dont have time to do it today | |