| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-30 | |||
| 16:52:09 | kashyap | melwitt: And your test (noted in the review) proves it, too. | |
| 16:52:31 | mdbooth | mriedem: I'm just trying to convince myself it solves the problem | |
| 16:53:22 | mriedem | cells v2 meeting in 7 minutes | |
| 16:53:31 | melwitt | kashyap: thanks for the sanity check. I'm thinking I'll add to the patch save of the original XML to write back in the exception case, if rebase/resize fails | |
| 16:53:32 | mriedem | in #openstack-meeting-3 | |
| 16:53:53 | mdbooth | melwitt: Do you happen to know what XMLDesc(0) returns for an inactive domain? | |
| 16:54:13 | kashyap | melwitt: Yep, saving the original XML to handle the failure path sounds correct. | |
| 16:54:45 | melwitt | mdbooth: you mean a stopped instance? no actually. I would also wonder what the rebase does in that case too | |
| 16:55:13 | mdbooth | I'm thinking in the context of detach_volume currently | |
| 16:56:00 | mdbooth | My feeling is that the use of the live flag is itself wrong | |
| 16:56:21 | melwitt | mdbooth: it solved the problem in a repro environment that I tested | |
| 16:57:05 | mdbooth | melwitt: Right, I just suspect it could be both simpler and more robust | |
| 16:57:18 | mdbooth | It looks like a haven for edge cases, and we closed one | |
| 16:58:37 | mdbooth | For eg, 'live' could change during the call | |
| 16:58:42 | mdbooth | Because the domain is stopped, for eg | |
| 16:58:45 | melwitt | mdbooth: sorry, that was in response to "trying to convince myself it solves the problem." I agree with you that there's probably a better way to refactor that function | |
| 16:59:07 | melwitt | it's bug prone for sure | |
| 16:59:48 | openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: config drive https://review.openstack.org/409404 | |
| 17:00:11 | mdbooth | melwitt: Yeah, I'm mid-stream here :) Not entirely sure where I'm going with this. | |
| 17:00:45 | dansmith | mriedem: cells meeting? | |
| 17:00:46 | melwitt | I think you're just surveying the horror | |
| 17:00:47 | mdbooth | It's just such an over-complex function, I spent way too much time thinking about it already | |
| 17:06:06 | kashyap | melwitt: I know you're multi-tasking between two reviews. For later, some related notes I did a brain-dump on debugging blockRebase() -> _swap_volume() -- http://lists.openstack.org/pipermail/openstack-dev/2016-October/105158.html | |
| 17:06:40 | kashyap | melwitt: Even after staring at these libvirt / QEMU block APIs, damned if I fully wrapped my head around. Just when I think I got a little hang of something, there it comes ... another corner case that I didn't think of | |
| 17:09:39 | andreaf | jamespage: I have a change in Tempest that if merged would break some nova-lxd integration tests from the in-tree tempest plugin | |
| 17:10:28 | andreaf | jamespage: but I don't see those tests running anywhere in nova-lxd gate, so I was wondering if I can just change tempest and propose a patch to fix nova-lxd afterward? | |
| 17:23:29 | jamespage | andreaf: +1 that's good with me | |
| 17:23:55 | jamespage | andreaf: sorry for no response yesterday - just back from hols so catchup was a bit extreme | |
| 17:24:09 | andreaf | jamespage: cool - so is that right that tests are not running in any CI right now? | |
| 17:24:32 | andreaf | jamespage: np thanks for your reply | |
| 17:24:39 | jamespage | andreaf: we have a tempest devstack gate, but I don't think the in-tree tests actually get executed | |
| 17:25:27 | andreaf | jamespage: yeah that was what I found as well | |
| 17:28:47 | openstackgerrit | Elod Illes proposed openstack/nova master: WIP: delete after failed evac https://review.openstack.org/499237 | |
| 17:29:50 | melwitt | kashyap: cool, thanks for the link | |
| 17:31:24 | openstackgerrit | Ed Leafe proposed openstack/nova master: WIP - Add ${Destination} and ${Destination}List objects https://review.openstack.org/499239 | |
| 17:31:24 | openstackgerrit | Ed Leafe proposed openstack/nova master: Add alternate hosts https://review.openstack.org/486215 | |
| 17:37:58 | mriedem | alaski: did you or someone else at some point have a doc with the known gaps in what we track for reporting instance action events? | |
| 17:38:02 | mriedem | i could have sworn you did | |
| 17:39:29 | alaski | I may have. You might be thinking of a spec opened by rosimata a while back looking to improve instance actions which listed some of the things it missed. | |
| 17:40:57 | mriedem | i also found https://blueprints.launchpad.net/nova/+spec/improve-instance-action-events-2 and something it depends on | |
| 17:41:17 | alaski | https://review.openstack.org/#/c/256743/ maybe | |
| 17:41:44 | alaski | https://review.openstack.org/#/q/project:openstack/nova-specs+owner:%22Brian+Rosmaita+%253Crosmaita.fossdev%2540gmail.com%253E%22 | |
| 17:42:33 | mriedem | yup https://review.openstack.org/#/c/256743/6/specs/newton/approved/expand-instance-actions-coverage.rst it is | |
| 17:42:38 | mriedem | i remember seeing that and commenting on it | |
| 17:42:39 | mriedem | thanks | |
| 17:43:14 | alaski | np, glad to help | |
| 17:47:10 | openstackgerrit | Merged openstack/osc-placement master: Fix the bug link in the readme https://review.openstack.org/499206 | |
| 17:50:15 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Hyper-V: Perform proper cleanup after cold migration https://review.openstack.org/486955 | |
| 18:24:55 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Cleanup allocations on invalid dest node during live migration https://review.openstack.org/498861 | |
| 18:24:56 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Refactor LiveMigrationTask._find_destination https://review.openstack.org/498874 | |
| 18:44:43 | mriedem | huh, hitting a 500 in the PUT /allocatoins API | |
| 18:44:48 | mriedem | */allocations | |
| 18:44:58 | mriedem | index error shenanigans | |
| 18:45:58 | mriedem | Sending updated allocation [{'resource_provider': {'uuid': '7ab9dab7-65c6-4961-9403-c8fc50dedb6b'}, 'resources': {}}] for instance dc8a686c-ad92-48f3-8594-d00c6e671a1c after removing resources for 7ab9dab7-65c6-4961-9403-c8fc50dedb6b | |
| 18:46:57 | mriedem | remove_provider_from_instance_allocation thinks we're doing a resize to same host because there is only one compute provider because of the bug where we force the evacuate host and don't have allocations on the dest node | |
| 18:47:54 | mriedem | funny the jsonschema validation in PUT /allocations doesn't fail on resources being an empty dict | |
| 18:48:54 | mriedem | oh i guess we dont have schema validation for that | |
| 18:49:36 | cdent | mriedem: huh, I thought we had either a bug or fix for that, it’s tweaking a memory | |
| 18:49:50 | cdent | there was at least some discussion about whether that should mean anything | |
| 18:51:51 | dansmith | mriedem: am I good to approve that assisted snapshot backport? just wanted to make sure we weren't in a quiet period for release things or something | |
| 18:52:12 | mriedem | dansmith: yes, pike is GAed | |
| 18:52:13 | dansmith | er, I guess I'm back to just +2ing now anyway, so .. nevermind | |
| 18:52:19 | mriedem | yeah | |
| 18:52:25 | mriedem | i appreciate your gusto | |
| 18:52:29 | dansmith | got so used to the pike body slams | |
| 18:53:17 | mriedem | cdent: https://bugs.launchpad.net/nova/+bug/1714072 | |
| 18:53:18 | openstack | Launchpad bug 1714072 in OpenStack Compute (nova) "PUT /allocations/{consumer_id} fails with a 500 if "resources: {}"" [Medium,Triaged] | |
| 18:53:46 | mriedem | guess we just have to detect that in code and return a 400? | |
| 18:53:56 | mriedem | i don't know if jsonschema allows min_items for an object type | |
| 18:54:05 | mriedem | *minItems | |
| 18:54:47 | cdent | mriedem: left a query | |
| 18:54:53 | mriedem | http://json-schema.org/latest/json-schema-validation.html#rfc.section.6.12 | |
| 18:55:17 | mriedem | oh minProperties | |
| 18:55:23 | cdent | the same question is coming up on the spec dansmith has prompted me to write (which is nearly ready) | |
| 18:57:48 | mriedem | cdent: replied | |
| 18:58:49 | mriedem | also means i can't really do the whole (1) provide recreate test patch then (2) provide fix | |
| 18:59:02 | mriedem | because (1) is essentially blocked | |
| 18:59:30 | mriedem | but that's not the end of the world | |
| 18:59:55 | openstackgerrit | Chris Dent proposed openstack/nova-specs master: Add a spec for POST /allocations in placement https://review.openstack.org/499259 | |
| 19:00:04 | cdent | dansmith: at your service ^ | |
| 19:00:19 | cdent | it’s got a fair few open questions on it, but I think they’ll resolve relatively quickly | |
| 19:00:24 | dansmith | wow, not even a WIP in front.. ballsy | |
| 19:01:18 | cdent | it should be fairly obivousl from the not just one but two entirely different body representations that it is a wip | |
| 19:01:36 | cdent | i’ll take that bug right now, mriedem | |
| 19:01:47 | cdent | unless you’ve already started? | |
| 19:02:22 | mriedem | it's yours | |
| 19:02:31 | mriedem | i have to get my hairs cut in <1 hour | |
| 19:03:24 | cdent | check | |
| 19:22:48 | cfriesen_ | mriedem: if you're still around, would the changes you're talking about doing for https://bugs.launchpad.net/nova/+bug/1427772 (with the claims in the scheduler) still handle the resources that aren't handled by Placement? (all the per-numa stuff, basically) | |
| 19:22:50 | openstack | Launchpad bug 1427772 in OpenStack Compute (nova) "Instance that uses force-host still needs to run some filters" [Low,Confirmed] | |
| 19:33:17 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Add test for empty resources in allocation https://review.openstack.org/499269 | |
| 19:33:18 | openstackgerrit | Chris Dent proposed openstack/nova master: [placememnt] Require at least one resource class in allocation https://review.openstack.org/499270 | |
| 19:33:26 | cdent | mriedem_away, dansmith ^^ | |
| 19:47:44 | openstackgerrit | Dan Smith proposed openstack/nova master: Revert allocations by migration uuid https://review.openstack.org/498949 | |
| 19:47:44 | openstackgerrit | Dan Smith proposed openstack/nova master: Make allocation cleanup honor new by-migration rules https://review.openstack.org/498948 | |
| 19:47:45 | openstackgerrit | Dan Smith proposed openstack/nova master: Move allocation manipulation out of drop_move_claim() https://review.openstack.org/498947 | |
| 19:47:45 | openstackgerrit | Dan Smith proposed openstack/nova master: Pre-create migration object https://review.openstack.org/498950 | |
| 20:27:25 | openstackgerrit | melanie witt proposed openstack/nova master: Save updated libvirt domain XML after swapping volume https://review.openstack.org/498983 | |
| 20:52:36 | openstackgerrit | Dan Smith proposed openstack/nova master: Pre-create migration object https://review.openstack.org/498950 | |
| 21:00:27 | cdent | dansmith: i responded to your comments on that spec, but haven’t pushed a new version yet figured may as well wait for ed, matt, jay | |