| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-24 | |||
| 03:22:50 | mriedem | alex_xu: hmm, yeah probably not, which sounds similar to https://bugs.launchpad.net/nova/+bug/1712411 | |
| 03:22:51 | openstack | Launchpad bug 1712411 in OpenStack Compute (nova) pike "Allocations may not be removed from dest node during failed migrations" [High,Triaged] | |
| 03:23:11 | mriedem | remove allocations during reschedule is now going to be like deallocating ports from a host | |
| 03:23:39 | mriedem | alex_xu: do you want to open a new bug for the reschedule issue? | |
| 03:29:51 | openstackgerrit | Merged openstack/nova master: Delete instance allocations when the instance is deleted https://review.openstack.org/496942 | |
| 04:02:07 | alex_xu | mriedem: yea, I already opened one https://bugs.launchpad.net/nova/+bug/1712718 | |
| 04:02:08 | openstack | Launchpad bug 1712718 in OpenStack Compute (nova) "Allocations records aren't removed after rescheduling" [Undecided,New] | |
| 04:04:23 | alex_xu | mrhillsman: yea, very similar with https://bugs.launchpad.net/nova/+bug/1712411 | |
| 04:04:24 | openstack | Launchpad bug 1712411 in OpenStack Compute (nova) pike "Allocations may not be removed from dest node during failed migrations" [High,Triaged] | |
| 04:04:43 | alex_xu | mriedem: yea, very similar with https://bugs.launchpad.net/nova/+bug/1712411 | |
| 04:05:11 | alex_xu | just for the case instance boot rescheduling | |
| 06:14:16 | openstackgerrit | Alex Xu proposed openstack/nova master: Add functional test for rescheduling https://review.openstack.org/496995 | |
| 06:54:41 | openstackgerrit | Alex Xu proposed openstack/nova master: Remove allocation when booting instance rescheduled or aborted https://review.openstack.org/496995 | |
| 07:14:56 | openstackgerrit | liusheng proposed openstack/nova master: Trivail: remove unused instance_group_members_add method https://review.openstack.org/497115 | |
| 07:17:05 | openstackgerrit | liusheng proposed openstack/nova master: Trivial: remove unused instance_group_member_delete method https://review.openstack.org/497115 | |
| 08:30:33 | openstackgerrit | jichenjc proposed openstack/nova master: Do not attempt volume swap when guest is stopped/suspended https://review.openstack.org/389798 | |
| 08:49:31 | openstackgerrit | jichenjc proposed openstack/nova master: WIP: use conf for keystone session creation https://review.openstack.org/485121 | |
| 09:02:17 | openstackgerrit | jichenjc proposed openstack/nova master: fix race condition of instance host https://review.openstack.org/494458 | |
| 09:10:16 | ratailor | Can anybody tell me how to debug functional test ? | |
| 09:28:33 | openstackgerrit | jichenjc proposed openstack/nova master: propagate OSError to MigrationPreCheckError https://review.openstack.org/469779 | |
| 09:30:07 | openstackgerrit | jichenjc proposed openstack/nova master: WIP: fix warning: enable-extensions https://review.openstack.org/497170 | |
| 09:30:53 | openstackgerrit | Stephen Finucane proposed openstack/nova-specs master: Websockify security proxy framework https://review.openstack.org/496160 | |
| 09:48:51 | ratailor | Dinesh_Bhor, do you know how to debug functional test ? | |
| 09:56:33 | owalsh | moshele: do you have a minute to confirm some details about pci alias config? | |
| 10:08:25 | moshele | owalsh: yes | |
| 10:12:13 | owalsh | moshele: it appears to be required for nova-compute (comment #2 in https://bugs.launchpad.net/puppet-nova/+bug/1702730) | |
| 10:12:14 | openstack | Launchpad bug 1702730 in puppet-nova "[pci]/alias option is not present in the nova configuration on the computes" [Medium,New] - Assigned to Sai Sindhur Malleni (smalleni) | |
| 10:13:22 | moshele | owalsh: yes, it was added when I fixed the cold migration and resize for pci devices | |
| 10:15:00 | owalsh | moshele: ack, thanks | |
| 10:15:28 | moshele | owalsh: np | |
| 10:40:20 | sean-k-mooney | stephenfin: o/ congrats on the 2.0 patchwork release. did you ever get around to prototyping zuul intergration with the new rest api? | |
| 10:40:53 | stephenfin | sean-k-mooney: Not yet, but it's coming | |
| 10:40:59 | stephenfin | soon as I finish one or two other side projects | |
| 10:41:16 | stephenfin | Also, thanks :) | |
| 10:41:17 | sean-k-mooney | nice will cool to see. | |
| 10:43:30 | sean-k-mooney | stephenfin: im still subscribed to patchwork@lists.ozlabs.org from when i was testing the proxy support so i recived your annouce email which is always nice to see. any plans in place to get the oficall host instance over to 2.0 in the near future | |
| 10:43:53 | stephenfin | sean-k-mooney: https://lists.ozlabs.org/pipermail/patchwork/2017-August/004550.html | |
| 10:44:05 | stephenfin | (yup, this weekend :)) | |
| 10:45:23 | sean-k-mooney | awsome. i would love to set up an experimental ci to test new patchs to ovs to ensure i can still deploy ovs-dpdk with kolla and devstack | |
| 10:45:57 | sean-k-mooney | need to figure out how to do that upstream though or wait till i deploy our new dev cloud | |
| 11:09:42 | openstackgerrit | Jianghua Wang proposed openstack/nova master: XenAPI: image handler option improvement https://review.openstack.org/497201 | |
| 12:00:52 | openstackgerrit | Alex Xu proposed openstack/nova master: Remove allocation when booting instance rescheduled or aborted https://review.openstack.org/496995 | |
| 12:04:28 | openstackgerrit | liusheng proposed openstack/nova master: Trivial: remove unused instance_group_member_delete method https://review.openstack.org/497115 | |
| 12:15:00 | openstackgerrit | Viktor Varga proposed openstack/nova master: Remove mox from unit/conductor/test_conductor.py (1) https://review.openstack.org/496691 | |
| 12:22:12 | pooja_jadhav | mriedem :Hi | |
| 12:22:22 | openstackgerrit | Ed Leafe proposed openstack/nova master: docs: Document the scheduler workflow https://review.openstack.org/475810 | |
| 12:25:39 | pooja_jadhav | mriedem: Should we consider the urn format uuid like "urn:11111111-2222-4f44-5555-66666666666" or "urn:uuid:11111111-2222-4f44-5555-66666666666" for the nova APIs or not? | |
| 12:59:54 | mriedem | stephenfin: want to nudge this through? https://review.openstack.org/#/c/496930/ | |
| 13:00:31 | stephenfin | mriedem: Looks good to me. +Wd | |
| 13:01:51 | mriedem | thanks | |
| 13:02:18 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform instance.live_migration_force_complete notification https://review.openstack.org/480561 | |
| 13:02:25 | openstackgerrit | Gábor Antal proposed openstack/nova master: Transform instance.live_migration_rollback_dest notification https://review.openstack.org/480214 | |
| 13:07:46 | VAhl | Is it possible to change port on the Placement-api? If yes, which parameter should be set in the [placement] section in nova.conf | |
| 13:10:29 | VAhl | found it. It is sites-available/nova-placement-api.conf | |
| 13:11:48 | stephenfin | mriedem: I cherry-picked that cell deprecation to stable/pike, if we want to bring it in https://review.openstack.org/#/c/497166 | |
| 13:11:59 | stephenfin | (mostly so we can stay compliant with the deprecation policy) | |
| 13:12:58 | mriedem | :/ | |
| 13:22:56 | openstackgerrit | Eric Fried proposed openstack/nova master: Monkey patch the blockdiag extension https://review.openstack.org/476159 | |
| 13:23:04 | mriedem | alex_xu: thanks for starting this https://review.openstack.org/#/c/496995/ - i've got some comments inline | |
| 13:23:09 | mriedem | cdent: dansmith: ^ | |
| 13:30:22 | rabel | hi there. we have a patch for the vmware driver: https://review.openstack.org/#/c/494169/ ( cdent already reviewed it ). i posted a comment with vmware-recheck-patch about a week ago, but nothing happens. can you help me on how to trigger vmware ci tests? | |
| 13:32:24 | cdent | rabel: you may have a hit time when the vmware ci wasn’t happy. I’ve tried to trigger another | |
| 13:32:34 | rabel | cdent: thank you! | |
| 13:32:43 | mriedem | it hasn't been happy for about 6 months | |
| 13:33:04 | cdent | mriedem: I’m happy to report that that is almost but not quite on the agenda to get some attention soon | |
| 13:33:04 | mriedem | but, there haven't been a ton of changes proposed to make it a priority | |
| 13:33:20 | mriedem | heh | |
| 13:34:06 | mriedem | alex_xu: i'm not sure if you want to handle the reschedule for resize in the same patch, i know it's getting late for you and we don't have a functional regression test for the resize + reschedule case | |
| 13:34:11 | cdent | there’s been the usual such and such is in the way | |
| 13:38:59 | alex_xu | mriedem: yea, I also thought the resize case is another patch | |
| 13:39:26 | alex_xu | mriedem: cdent just replied this comment https://review.openstack.org/#/c/496995/5/nova/compute/manager.py@1733 | |
| 13:41:13 | lajoskatona | cdent: Gibi asked me to think about adding an extra testclass to test_servers to test the server moving stuff with custom resources | |
| 13:41:56 | lajoskatona | cdent: now I have a new class, and the setup works with custom resources and resorce providers, whatever.... | |
| 13:42:05 | cdent | lajoskatona: that’s a good idea | |
| 13:43:24 | lajoskatona | cdent: BUT: all tests are failing, due to exception from scheduler: 2017-08-24 14:49:41,491 DEBUG [nova.scheduler.manager] Got no allocation candidates from the Placement API. This may be a temporary occurrence as compute nodes start up and begin reporting inventory to the Placement service. | |
| 13:44:03 | lajoskatona | cdent: shall I upload as WIP, and somebody should check it what am I missing from the workflow? | |
| 13:44:03 | cdent | can you push up your code as a wip so it is easier to look at? | |
| 13:44:06 | cdent | :) | |
| 13:44:11 | cdent | yes! | |
| 13:44:16 | lajoskatona | cdent, ok, I upload :-) | |
| 13:44:23 | alex_xu | mriedem: cdent in the original RT behaviour, the instance claim will release the resource, we also nill out instance's host and name, that also means the resource will be released by RT | |
| 13:45:32 | mriedem | i'm replying | |
| 13:47:51 | mriedem | guh | |
| 13:47:54 | mriedem | ok replied | |
| 13:48:07 | mriedem | alex_xu: so you've got a point, as usual :) | |
| 13:48:28 | mriedem | if the instance.host is None, then the API isn't going to cast to the compute to complete the deletion which would remove the allocation | |
| 13:48:39 | mriedem | and we don't currently remove the allocations in the API during a 'local' delete | |
| 13:48:45 | mriedem | so we're kind of stuck | |
| 13:49:02 | cdent | is there a reason we don’t do it from the API? | |
| 13:49:09 | cdent | s/don’t/don’t want to/ | |
| 13:49:09 | mriedem | hasn't been fixed yet | |
| 13:49:17 | openstack | Launchpad bug 1679750 in OpenStack Compute (nova) "Allocations are not cleaned up in placement for instance 'local delete' case" [Medium,Confirmed] | |
| 13:49:17 | mriedem | no, it's just an open bug https://bugs.launchpad.net/nova/+bug/1679750 | |
| 13:49:37 | mriedem | cdent: it was less of a problem at the time it was reported because the periodic task would heal and remove the allocations | |
| 13:50:02 | mriedem | well, actually let me see about that | |
| 13:50:41 | mriedem | yeah, actually the periodic task should still clean these up, in _remove_deleted_instances_allocations | |
| 13:50:46 | mriedem | that gets all allocations for the node | |
| 13:51:09 | mriedem | and if we get InstanceNotFound looking up the instance, it deletes the allocations for that instance | |
| 13:55:06 | mriedem | left some more comments | |
| 13:56:05 | mriedem | probably need to get dansmith's opinion when he's around. i think we could go with this patch even though the periodic would cleanup the allocations for the node once the instance is deleted, and if we removed the allocations from the api during local delete that would also take care of it - but we don't have that fix yet | |