| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-04-21 | |||
| 09:55:43 | bauzas | mgoddard: since the virt driver stopped reporting the node, then the RT deleted the compute node record | |
| 09:55:51 | openstackgerrit | Lee Yarwood proposed openstack/nova master: docs: Add stable device rescue docs https://review.opendev.org/700837 | |
| 09:56:49 | lyarwood | stephenfin: ^ thanks btw | |
| 09:57:05 | mgoddard | bauzas: but the service should still exist, even without a node | |
| 09:57:31 | bauzas | mgoddard: we're making different assumptions in different pieces of code | |
| 09:57:55 | mgoddard | bauzas: I haven't seen that happen before | |
| 09:58:14 | bauzas | mgoddard: in the RT, we assume that it's reasonable to not have a CN record when the virt driver tells you no about it | |
| 09:58:35 | bauzas | mgoddard: but on the other hand, we assume that compute nodes exist if you delete a compute service | |
| 10:00:22 | mgoddard | bauzas: if that is true, should be easy to reproduce? | |
| 10:00:37 | mgoddard | and probably easy to test my hypothesis also | |
| 10:03:30 | bauzas | mgoddard: I've asked the reporter to come back with details | |
| 10:04:21 | openstackgerrit | Merged openstack/nova master: Use oslo policy flag to disable default change warning instead of all https://review.opendev.org/717884 | |
| 10:04:26 | bauzas | I could have some OSP crazy env that would help me to verify this hypothesis but to be frank I don't have the energy for it :) | |
| 10:05:05 | bauzas | mgoddard: like I said in my last comment of bug 1860312, I'd rather spend my energy on reviewing your change | |
| 10:05:05 | openstack | bug 1860312 in OpenStack Compute (nova) "compute service failed to delete" [Undecided,Incomplete] https://launchpad.net/bugs/1860312 | |
| 10:08:36 | bauzas | mgoddard: actually, the bug is pretty nasty and has a lot of other related bugs, do you think you'd be able to give it a good shape soon ? | |
| 10:08:54 | bauzas | to give *the change a good shape soon ? | |
| 10:10:38 | bauzas | gibi: FYI, ^ | |
| 10:11:29 | bauzas | we're too close to RC1 to consider those bugs as rc-potential but they're definitely worth backportable down the road to all our stable branches | |
| 10:12:08 | bauzas | actually, mriedem thought the same https://bugs.launchpad.net/nova/+bug/1841481 | |
| 10:12:08 | openstack | Launchpad bug 1841481 in OpenStack Compute (nova) "Race during ironic re-balance corrupts local RT ProviderTree and compute_nodes cache" [Medium,In progress] - Assigned to Matt Riedemann (mriedem) | |
| 10:12:41 | gibi | bauzas: ack, I agree that we should not try to blow up RC1 with this fix. | |
| 10:12:45 | bauzas | mgoddard: sorry for interupting you again, but we have around 4 open bugs for probably the same issue(s) or related | |
| 10:12:59 | bauzas | mgoddard: could we consider https://bugs.launchpad.net/nova/+bug/1853009 as the primary one ? | |
| 10:13:00 | openstack | Launchpad bug 1853009 in OpenStack Compute (nova) "Ironic node rebalance race can lead to missing compute nodes in DB" [High,In progress] - Assigned to Mark Goddard (mgoddard) | |
| 10:14:16 | mgoddard | bauzas: I'll get the patches into shape today | |
| 10:15:15 | bauzas | mgoddard: as I said, we're too short in time to deliver them for RC1, but I think this should be addressed as quicky as possible after the victoria branch | |
| 10:15:48 | mgoddard | bauzas: ack | |
| 10:22:05 | bauzas | mgoddard: fwiw, starting to review https://review.opendev.org/#/c/695012/4/nova/tests/functional/regressions/test_bug_1853009.py | |
| 10:23:04 | mgoddard | bauzas: cool. I don't expect the code to change significantly, just need to adapt to upstream changes since proposed | |
| 10:39:23 | gibi | bauzas: what do you think about https://bugs.launchpad.net/nova/+bug/1872385 it is currently marked as rc potential but it the fix is not ready. I can take the fix over this afternoon and make it complete or we can drop the rc potential tag as you also noted that this is not a clear regression for Ussuri as we have child RPs for couple of releases now | |
| 10:39:23 | openstack | Launchpad bug 1872385 in OpenStack Compute (nova) "delete sub resource provider in tree" [High,In progress] - Assigned to Wenping Song (wenping1) | |
| 10:47:09 | sean-k-mooney | gibi: i dont have it deploy but i have worked with dvd on https://bugs.launchpad.net/nova/+bug/1851545 downstream https://bugzilla.redhat.com/show_bug.cgi?id=1767797 | |
| 10:47:09 | openstack | Launchpad bug 1851545 in OpenStack Compute (nova) "Port update exception on nova unshelve for instance with PCI devices (part 2)" [Undecided,New] | |
| 10:47:10 | openstack | bugzilla.redhat.com bug 1767797 in openstack-nova "When unshelving an SR-IOV instance, the binding profile isn't reclaimed or rescheduled, and this might cause PCI-PT conflicts" [High,New] - Assigned to nova-maint | |
| 10:48:06 | sean-k-mooney | gibi: it looks like all my comments downstream are private but yes this is broken on queens i memtioned this seperatly before. we are not updating the port binding profile before generating the xml on unshelve | |
| 10:48:33 | sean-k-mooney | im not sure if it affect all release but its broken on queens | |
| 10:50:57 | sean-k-mooney | gibi: ill traige the upstream bug. i can proably try and replicate it later i just have not had time to work on it since it was first reported but now might be as good a time as any one i finish rebasing my current patches | |
| 11:11:07 | gibi | sean-k-mooney: thanks a lot! | |
| 11:29:57 | openstackgerrit | Merged openstack/nova master: Add missing white spaces between words in log messages https://review.opendev.org/720112 | |
| 12:03:16 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Revert "Temporarily skip TestNovaMigrationsMySQL" https://review.opendev.org/718630 | |
| 12:15:21 | gibi | bauzas and others: I have to jump on a meeting marathon for the next 2 and a half hours | |
| 12:19:47 | openstackgerrit | Merged openstack/nova-specs master: Adds spec infrastructure for Victoria https://review.opendev.org/710023 | |
| 12:48:35 | bauzas | gibi: ack thanks, and I'll review https://review.opendev.org/#/c/719163/ | |
| 13:05:10 | openstackgerrit | Merged openstack/nova master: docs: Add stable device rescue docs https://review.opendev.org/700837 | |
| 13:16:31 | kashyap | lyarwood: Since I've written it partially, and been vacillating on this for a while, I'll repropose this spec for V, as that is on TODO "forever" anyway. I know you've already got plenty on your plate | |
| 13:17:01 | kashyap | lyarwood: And I'll add you to the 'collaborators' list, that sounds okay? | |
| 13:17:32 | lyarwood | ack sure | |
| 13:20:01 | kashyap | lyarwood: Appreciate the virtual kick in the *** | |
| 13:20:38 | kashyap | A real one will also be accepted when circumstances allow in-person gathering ;-) | |
| 13:22:44 | gmann | gibi: stephenfin thanks. let me squash both policy doc patches. | |
| 13:43:37 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP: nova-next: Start testing the 'q35' machine type https://review.opendev.org/708701 | |
| 13:45:25 | lyarwood | kashyap: ^ added a virt-preview job, I think swap volume might only work once we switch to block copy | |
| 13:45:32 | lyarwood | kashyap: with q35 | |
| 13:46:59 | kashyap | lyarwood: Ah, yes. Indeed that's true | |
| 13:49:05 | lyarwood | that said I can't reproduce this outside of the tempest test | |
| 13:55:26 | lyarwood | yup np | |
| 14:20:26 | openstackgerrit | Mark Goddard proposed openstack/nova master: Add functional regression test for bug 1853009 https://review.opendev.org/695012 | |
| 14:20:26 | openstack | bug 1853009 in OpenStack Compute (nova) ussuri "Ironic node rebalance race can lead to missing compute nodes in DB" [High,In progress] https://launchpad.net/bugs/1853009 - Assigned to Mark Goddard (mgoddard) | |
| 14:20:26 | openstackgerrit | Mark Goddard proposed openstack/nova master: Clear rebalanced compute nodes from resource tracker https://review.opendev.org/695187 | |
| 14:20:26 | openstackgerrit | Mark Goddard proposed openstack/nova master: Invalidate provider tree when compute node disappears https://review.opendev.org/695188 | |
| 14:20:27 | openstackgerrit | Mark Goddard proposed openstack/nova master: Prevent deletion of a compute node belonging to another host https://review.opendev.org/694802 | |
| 14:20:27 | openstackgerrit | Mark Goddard proposed openstack/nova master: Fix inactive session error in compute node creation https://review.opendev.org/695189 | |
| 14:20:39 | mgoddard | bauzas: ^ | |
| 14:21:34 | mgoddard | unit test fix was simple, missing a fair=True argument to the synchronized decorator in RT, added upstream | |
| 14:25:21 | bauzas | mgoddard: ack, will review it later today | |
| 14:25:28 | mgoddard | thanks! | |
| 14:26:49 | bauzas | stephenfin: can you please help me understand what you mean by https://review.opendev.org/#/c/721548/1/releasenotes/notes/ussuri-prelude-4b96f1244cefcdf4.yaml@12 ? | |
| 14:27:07 | bauzas | stephenfin: looking at https://www.sphinx-doc.org/en/master/usage/restructuredtext/basics.html#hyperlinks I can't find what you mean | |
| 14:27:15 | stephenfin | bauzas: | |
| 14:27:17 | stephenfin | `Hello, world`__ | |
| 14:27:18 | stephenfin | ||
| 14:27:26 | stephenfin | __ https://example.com | |
| 14:27:52 | stephenfin | Or grep for '__' in releasenotes/notes | |
| 14:27:54 | bauzas | okay, but that's not something related to sphinx rST right? | |
| 14:28:13 | stephenfin | It's a general rST thing, not specific to Sphinx | |
| 14:28:46 | stephenfin | bauzas: https://docutils.sourceforge.io/docs/user/rst/quickref.html#indirect-hyperlink-targets | |
| 14:28:51 | bauzas | ok, /me goes looking at https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html | |
| 14:29:01 | bauzas | heh, jinxed | |
| 14:29:10 | openstackgerrit | Lee Yarwood proposed openstack/nova master: WIP: nova-next: Start testing the 'q35' machine type https://review.opendev.org/708701 | |
| 14:30:37 | bauzas | ah, OK, I understand better : "The reference name of the reference is not used to match the reference to its target. Instead, the order of anonymous hyperlink references and targets within the document is significant: the first anonymous reference will link to the first anonymous target. The number of anonymous hyperlink references in a document must match the number of anonymous targets. " | |
| 14:31:42 | stephenfin | Yup. Named links must be unique in their document so they chance of having conflicts in release notes is reasonably high, given the volume of them we have | |
| 14:32:07 | stephenfin | *so the chance | |
| 14:47:37 | stephenfin | dansmith: Am I wrong, or should this not happen? https://bugs.launchpad.net/nova/+bug/1868033 | |
| 14:47:37 | openstack | Launchpad bug 1868033 in OpenStack Compute (nova) "Booting instance with pci_device fails during rocky->stein live upgrade" [High,Confirmed] - Assigned to Stephen Finucane (stephenfinucane) | |
| 14:47:38 | stephenfin | Because of https://github.com/openstack/oslo.versionedobjects/blob/master/oslo_versionedobjects/base.py#L555-L564 | |
| 14:50:03 | dansmith | stephenfin: well, he's saying that it _is_ trying to backport, but it's not backporting far enough, which makes it sound like the manifest being sent it wrong or something | |
| 14:50:08 | dansmith | which really shouldn't happen | |
| 14:50:31 | dansmith | he's wrong to equate the upgrade_level to anything with objects though -- they are unrelated | |
| 14:50:55 | stephenfin | yeah, that patch isn't a suitable fix | |
| 14:53:21 | dansmith | stephenfin: I wonder if it's because that object isn't calling super() | |
| 14:53:56 | dansmith | and its obj_make_compat() is only handling the sub-object backporting, which shouldn't be needed anymore anyway, so maybe this one got skipped when those were culled? | |
| 14:55:23 | stephenfin | Ah, good spot. That's likely it | |
| 14:55:53 | stephenfin | the superclass handles ListOfObjectField and ObjectField fields explicitly | |
| 14:56:04 | dansmith | right, I'm having to reboot my dev machine, | |
| 14:56:18 | dansmith | but I think you'll find most list objects don't have an explicit backport routine anymore | |
| 14:57:33 | dansmith | yeah, confirmed, but let me look up the culling patch | |
| 14:59:29 | dansmith | ah, I think I see | |
| 15:00:12 | dansmith | we used to use that obj_relationships mapping, and obj_make_compatible was still centralized based on that map for lists | |
| 15:00:35 | dansmith | but this one didn't use that, so it was overriding from the beginning and thus didn't get caught in the grep | |