| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-12-01 | |||
| 14:28:34 | mriedem | https://review.openstack.org/#/c/524425/ | |
| 14:28:53 | mriedem | turns out that now that people are starting to roll up to using placement, they actually care about configurable policy for the service :) | |
| 14:30:33 | cdent | mriedem: I saw it briefly this morning before I rolled out but didn’t have a chance to look at it for real yet. I had thought that we were planning to put it off until we actually had some reason to have policy other than “you must be this high to ride any of this ride” to align with the “minimal config” thing? | |
| 14:31:16 | leakypipes | mriedem, superdan, giblet: melwitt did some good sleuthing on the bug report associated with https://review.openstack.org/#/c/518022. I concur with her findings and think the patch is good to go. | |
| 14:31:17 | cdent | superdan: cdenture could be workable, but it’s not quite right enough for me to be all in | |
| 14:31:49 | mriedem | cdent: given it's all admin-only apis i think that's also why we put off configurable policy, | |
| 14:31:55 | mriedem | the issue that someone brought up last night was the admin role, | |
| 14:31:59 | mriedem | they don't have an "admin" role i guess | |
| 14:32:09 | mriedem | so they are hacking around it with a config option | |
| 14:32:11 | mriedem | mgagne: ^ | |
| 14:32:22 | mriedem | this https://gist.github.com/mgagne/b43c1e085c1f1d50bebc054a7d387688 | |
| 14:32:51 | mriedem | anyway, not sure it's high priority, just something to have on the radar | |
| 14:32:55 | cdent | ah | |
| 14:32:56 | cdent | hmmm | |
| 14:32:57 | mriedem | since i totally forgot about this | |
| 14:32:59 | superdan | cdent: hah, okay | |
| 14:35:12 | fried_rice | cdent Or "indenture" :) | |
| 14:35:27 | cdent | progress | |
| 14:36:04 | superdan | oof | |
| 14:36:06 | cdent | of greater concern than picking a good nick is the zomg, effort, of having to remember to change it on fridays. _So_ exhausting. | |
| 14:36:09 | superdan | friday nicks are supposed to be fun :) | |
| 14:36:15 | fried_rice | "decdent" would be a little morbid | |
| 14:39:16 | mriedem | superdan: melwitt: tssurya: i'll start throwing cellsv2 bugs in the meeting agenda for lack of a better place https://wiki.openstack.org/wiki/Meetings/NovaCellsv2#Agenda | |
| 14:39:56 | figleaf | cdent -> decent | |
| 14:40:32 | fried_rice | Mm, that's a good one. | |
| 14:40:39 | fried_rice | or dcent | |
| 14:40:40 | figleaf | or, more appropriately, descent | |
| 14:40:44 | fried_rice | heh | |
| 14:41:56 | cdent | this is reminding me of my server naming scheme from late 90s: depressing words that start with ‘d’ | |
| 14:42:18 | superdan | mriedem: okay | |
| 14:42:36 | cdent | Has anything placement-related happened in the last 12 hours that I ought to be aware of for the update? | |
| 14:43:29 | mriedem | just the policy thing | |
| 14:43:40 | cdent | thanks | |
| 14:45:37 | mriedem | mikal: https://bugs.launchpad.net/nova/+bug/1735663 | |
| 14:45:39 | openstack | Launchpad bug 1735663 in OpenStack Compute (nova) "privsep.daemon takes 9 seconds during taking volume attaching operation" [Undecided,New] | |
| 14:46:09 | fried_rice | cdent You got the _ensure bug on your list? | |
| 14:46:18 | cdent | yessir | |
| 14:46:27 | fried_rice | coo | |
| 14:46:52 | fried_rice | cdent And this series: https://review.openstack.org/#/c/523932/ | |
| 14:47:11 | openstackgerrit | Jay Pipes proposed openstack/nova master: handle traits with sharing providers https://review.openstack.org/517119 | |
| 14:47:12 | openstackgerrit | Jay Pipes proposed openstack/nova master: Test alloc_cands with one RP shared between two RPs https://review.openstack.org/519617 | |
| 14:47:12 | openstackgerrit | Jay Pipes proposed openstack/nova master: Test alloc_cands with non overlapping sharing RPs https://review.openstack.org/519380 | |
| 14:47:13 | openstackgerrit | Jay Pipes proposed openstack/nova master: Test helper: validate provider summaries https://review.openstack.org/518982 | |
| 14:47:13 | openstackgerrit | Jay Pipes proposed openstack/nova master: Test allocation candidates: multiple aggregates https://review.openstack.org/518633 | |
| 14:47:14 | openstackgerrit | Jay Pipes proposed openstack/nova master: Add aggregates check in allocation canditates https://review.openstack.org/522407 | |
| 14:47:14 | openstackgerrit | Jay Pipes proposed openstack/nova master: Fix accumulated nits in refactor series https://review.openstack.org/521189 | |
| 14:47:15 | openstackgerrit | Jay Pipes proposed openstack/nova master: Fix missing rps in allocation candidates https://review.openstack.org/522409 | |
| 14:47:15 | openstackgerrit | Jay Pipes proposed openstack/nova master: Add functional test for get_all_with_shared() https://review.openstack.org/522408 | |
| 14:47:16 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: func tests for multiple shared RPs https://review.openstack.org/498737 | |
| 14:47:54 | cdent | fried_rice: that’ll get covered with the n-r-p topic | |
| 14:48:00 | fried_rice | ight | |
| 14:48:18 | leakypipes | fried_rice, giblet, superdan, mriedem: ok, series starting at https://review.openstack.org/#/c/517119/ is rebased and ready for reviews (mostly new func tests for alloc candidate permutations) | |
| 14:48:28 | fried_rice | leakypipes ack | |
| 14:49:05 | mriedem | i have to spend quality time with https://review.openstack.org/#/c/330285/ this morning | |
| 14:50:53 | leakypipes | hehe | |
| 14:51:19 | leakypipes | coffee_cat: why? it's only on revision 165... that's not even a middle-aged patch. | |
| 14:51:41 | coffee_cat | leakypipes: if you happen to have time while Zuul is chewing on all of those above you could chime in too :) | |
| 14:51:50 | mriedem | guh, MOS bugs from mitaka https://bugs.launchpad.net/nova/+bug/1735407 | |
| 14:51:51 | openstack | Launchpad bug 1735407 in Mirantis OpenStack 9.x "[Nova] Evacuation doesn't respect anti-affinity rules" [High,New] - Assigned to Oleksiy Molchanov (omolchanov) | |
| 14:51:54 | leakypipes | coffee_cat: yeah, I have it open :) | |
| 14:52:02 | coffee_cat | leakypipes: yeah, I appreciate all the time I can spend rebasing it :) | |
| 14:52:14 | mriedem | i think at this point, any MOS bugs from mitaka can just be marked invalid with "can you recreate on master?" | |
| 14:52:32 | coffee_cat | leakypipes: sweet, I can see I will not be bored over the weekend :) | |
| 14:52:43 | leakypipes | mriedem: "can you recreated with Triple-O please?" | |
| 14:53:05 | mriedem | actually this is a latent bug | |
| 14:53:22 | mriedem | since you can't evacuate more than one instance in the same request, | |
| 14:53:27 | mriedem | and they are using the anti-affinity policy, | |
| 14:53:32 | mriedem | which relies on the late affinity check on the compute, | |
| 14:53:37 | mriedem | and even if that fails, we don't reschedule an evacuate | |
| 14:53:40 | mriedem | so :( | |
| 14:54:19 | leakypipes | fun | |
| 14:54:21 | giblet | mriedem: I think there is no late check for rebuild just for boot | |
| 14:54:30 | mriedem | giblet: ah, ok, | |
| 14:54:32 | mriedem | well in this case, | |
| 14:54:49 | mriedem | if they spread out the evacuates, it works - because they give time for the computes to send instance info updates back to the scheduler | |
| 14:55:00 | mriedem | which reminds me, | |
| 14:55:15 | mriedem | there is a CLI in novaclient which just gets all instances from a given host and evacuates them in a for loop right superdan? | |
| 14:55:25 | superdan | yeah | |
| 14:55:35 | mriedem | if those instances were in a server group policy, we are likely destroying those | |
| 14:57:32 | giblet | without the late check we break the group policy, with the late check even nova doesn't reschedule we at least fail the evac instead of breaking the policy | |
| 14:57:43 | giblet | that is already better | |
| 14:58:05 | giblet | of course it would be nice to reschedule | |
| 14:59:32 | mriedem | https://bugs.launchpad.net/mos/+bug/1735407/comments/3 | |
| 14:59:34 | openstack | Launchpad bug 1735407 in Mirantis OpenStack 9.x "[Nova] Evacuation doesn't respect anti-affinity rules" [High,New] - Assigned to Oleksiy Molchanov (omolchanov) | |
| 14:59:37 | mriedem | my analysis | |
| 14:59:50 | mriedem | i don't really want to add a reschedule loop for evacuate | |
| 15:00:20 | mriedem | long-term we model affinity in placement and use that, but i'm not entirely sure that will fix this case either | |
| 15:00:31 | leakypipes | mriedem: ++ | |
| 15:00:48 | giblet | mriedem: +1 for your analysis | |
| 15:01:01 | giblet | mriedem: can we just add the post check without reschedule? | |
| 15:02:01 | mriedem | well i guess the affinity stuff in placement would mean we somehow 'claim' that the instance we're scheduling / evacuating is going to be on some target host, | |
| 15:02:10 | mriedem | so a concurrent request gets that information when asking placement where it should go | |
| 15:02:18 | mriedem | not sure how that is all going to work out | |
| 15:02:23 | mriedem | since we don't want to model server group policies in placement | |
| 15:05:30 | mriedem | giblet: https://bugs.launchpad.net/mos/+bug/1735407/comments/4 | |
| 15:05:31 | openstack | Launchpad bug 1735407 in Mirantis OpenStack 9.x "[Nova] Evacuation doesn't respect anti-affinity rules" [High,New] - Assigned to Oleksiy Molchanov (omolchanov) | |
| 15:05:49 | mriedem | giblet: doing the late check might be reasonable | |
| 15:05:56 | mriedem | fail the evacuate, so the user can retry | |
| 15:06:55 | giblet | I vote for option 1) now and maybe consider prioritizing affinity impl with Placement in the future | |
| 15:07:52 | giblet | mriedem: I have to run soon but I have already looked into adding such late check before so I can take this bug if it can wait until Monday | |