Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-02-21
20:56:06 tssurya dansmith , melwitt : wow, thanks :D truly honoured
20:56:28 belmoreira dansmith thanks for all the work in the placement filtering. Do you think it will be easy to backport to Pike?
20:56:42 melwitt tssurya: and thank you for all the great work
20:57:09 dansmith belmoreira: it's not backportable upstream for sure.. I think pike lacks the placement aggregate stuff that will be needed to make it actually work
20:57:30 tssurya melwitt : couldn't have done it without the cells team ;)
20:58:23 belmoreira dansmith I was not asking for the upstream backport. But I would like/need to backport it for us
20:58:36 dansmith belmoreira: the actual filter mechanism should apply back that far I think, so if you'd be willing to run a newer placement with the newer support it has then it's probably not too bad yeah
20:59:57 belmoreira Currently I started collecting data about the scheduling time (I'm running local placement). I would like to see the impact when starting consolidating placement in only one instance
21:00:19 belmoreira I will let you the numbers
21:01:11 openstackgerrit Matt Riedemann proposed openstack/nova stable/queens: Update plugs Contrail methods to work with privsep https://review.openstack.org/546778
21:01:25 dansmith belmoreira: cool
21:01:32 melwitt fyi I've added the cells review links to the priorities etherpad https://etherpad.openstack.org/p/rocky-nova-priorities-tracking
21:02:17 tssurya melwitt : thank you
21:04:40 mriedem efried: do you see any issues with including these in RC3? https://review.openstack.org/#/q/project:openstack/nova+topic:bug/1734625+branch:stable/queens
21:32:04 efried mriedem: No. Go for it.
21:32:30 efried mriedem: But they'll have to be rebased.
21:32:51 efried The alternative is to rebase https://review.openstack.org/#/c/546740/ on top of them instead.
21:33:02 efried which will make the cherry-pick clean.
21:33:14 efried I think
21:41:36 melwitt smcginnis is already in the March spirit :)
21:47:32 mriedem efried: that's why i'm asking,
21:47:35 mriedem if you put yours on top
21:48:05 mriedem efried: i didn't include those request id backports since they were large and didn't think we needed them for an RC
21:48:14 mriedem now if it helps with conflicts both ways, we might as well
21:48:55 efried mriedem: The existing cherry-pick is already proposed, and it reads pretty clean. So it's more work, not less, to backport the global_id stuff.
21:49:05 efried mriedem: It may make subsequent backports easier.
21:49:18 efried mriedem: It will make this cherry-pick cleaner (possibly zero-conflict)
21:49:27 efried mriedem: But it'll be work to get that done.
21:49:35 efried mriedem: Which I'm happy to do if you wish it to be so.
21:50:40 melwitt which additional backports would the global_id patches help?
21:50:56 melwitt because I agree, the one that's already proposed reads clean to me too
21:51:12 mriedem if it's not clean on top of https://review.openstack.org/#/q/project:openstack/nova+topic:bug/1734625+branch:stable/queens then nvm
21:51:20 mriedem we can just go with what's there, that's fine
21:51:36 mriedem https://review.openstack.org/#/q/project:openstack/nova+topic:bug/1734625+branch:stable/queens is just more LOC than i'm comfortable with for an RC when it's not something that is required
21:51:49 melwitt but if not having global_id patches dominos into making other things difficult too, then we bring in global_id patches
21:57:58 mriedem dansmith: want to get this for RC3? https://review.openstack.org/#/c/546729/
21:58:15 mriedem and https://review.openstack.org/#/c/546778/
21:58:26 efried melwitt: Depends how many backport-worthy bugs we find in the report client.
21:58:40 efried Considering the amount of change it saw in Q....
21:59:29 melwitt yeah. I'd say since we don't know yet and you already dealt with the conflicts and it looks clean now, just go with what we have and we'll burn the other bridge when we get to it
22:00:19 openstackgerrit Eric Berglund proposed openstack/nova master: Add check for redundant import aliases https://review.openstack.org/543995
22:02:32 efried melwitt: Dig
22:03:27 openstackgerrit Merged openstack/nova master: Only pull associated *sharing* providers https://review.openstack.org/545494
22:05:35 openstackgerrit Merged openstack/nova master: Fix error handling in compute API for multiattach errors https://review.openstack.org/545478
22:05:54 openstackgerrit Merged openstack/nova master: Avoid exploding if guest refuses to detach a volume https://review.openstack.org/546423
22:08:40 mriedem dansmith: i'm +2 on the backport for efried's fix https://review.openstack.org/#/c/546740/ - do we want to hack in the (cherry picked from x) thing?
22:09:12 mriedem oh nvm it's higher up
22:09:36 efried yuh
22:09:54 efried git review complained about it not being at the bottom for some reason. Possibly because of the conflicts section? Dunno.
22:10:30 mriedem melwitt: in case you haven't seen https://review.openstack.org/#/c/546315/ and the one after it
22:10:41 mriedem mnaser found some problems after working on those local delete backports
22:14:43 mriedem i kind of sort of wonder if we should get those into rc3 as well
22:15:07 mriedem https://review.openstack.org/#/q/topic:bug/1404867+status:open+branch:stable/queens plus the 2 on master
22:15:14 mriedem maybe that's too much
22:17:35 mnaser mriedem: part of me says - this bug existed for a long time so it's not exactly affecting this release - the other part says it might manifest itself much more after an upgrade where all computes end up being a minimum of pike
22:18:04 mnaser if the computes were behind and are being updated *just enough* to upgrade api, then it'll happen
22:21:03 mriedem right that's my worry, any bfv with an existing volume where you have at least 1 pike compute, and scheduling fails, and you orphan the volume when deleting the instance
22:22:03 mriedem less of an issue in pike where you might have 1 ocata compute so we don't reserve the volume in the api,
22:22:13 mriedem more of an issue in pike and queens where the computes would be new enough to reserve the volume in the api
22:22:30 mriedem *more of an issue in queens
22:24:42 melwitt hm, sounds like we should include it then. the patches are small
22:26:56 openstackgerrit Eric Berglund proposed openstack/nova master: Add check for redundant import aliases https://review.openstack.org/543995
22:29:53 mriedem except your big one
22:30:01 openstackgerrit Takashi NATSUME proposed openstack/nova master: Transform live_migration_post_dest notification https://review.openstack.org/469784
22:30:20 melwitt me? no
22:31:32 mriedem ok i'll hit these queens ones and let dansmith make a call
22:33:21 openstackgerrit Takashi NATSUME proposed openstack/nova master: Adds view builders for keypairs controller https://review.openstack.org/347289
22:34:01 melwitt was just kidding on that. I had thought the others had already gone to stable/queens
22:35:56 mriedem done https://review.openstack.org/#/q/topic:bug/1404867+status:open+branch:stable/queens
22:36:43 mriedem https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/queens+label:Code-Review=2
22:37:15 openstackgerrit Takashi NATSUME proposed openstack/nova master: Fix 500 error while passing 4-byte unicode data https://review.openstack.org/407514
22:37:22 melwitt thanks. dansmith ^ if you think those are cool for rc3
22:38:51 openstackgerrit Merged openstack/nova master: Update plugs Contrail methods to work with privsep https://review.openstack.org/533212
22:39:27 mriedem will likely need a +2/+W from dan on the backport for ^ as well
22:40:10 dansmith sorry I had a thing
22:41:09 dansmith mriedem: so just the stable/queens list from above?
22:43:03 openstackgerrit Takashi NATSUME proposed openstack/nova master: List/show all server migration types (1/2) https://review.openstack.org/430608
22:44:26 dansmith mriedem: need someone else for that backport of mine I guess
22:44:31 mnaser mriedem: idk if we want to land those other last 2 on top for rc3
22:46:40 openstackgerrit Takashi NATSUME proposed openstack/nova master: List/show all server migration types (2/2) https://review.openstack.org/459483
22:48:09 mriedem mnaser: we do
22:48:30 openstackgerrit Takashi NATSUME proposed openstack/nova-specs master: Abort Cold Migration https://review.openstack.org/334732
22:48:51 mriedem dansmith: need this one also https://review.openstack.org/#/c/546778/
22:48:58 mriedem dansmith: you'll likely need to just approve that one,
22:49:04 mriedem consider my stable core backport as a +2
22:49:32 mriedem mnaser: mel approved https://review.openstack.org/#/c/546315/ and the one after it so can you backport those to queens also so we can get them into rc3?
22:51:13 dansmith mriedem: ack
22:51:51 wolverineav hey, mriedem stephenfin, just a quick heads-up on the IVS vif change https://review.openstack.org/#/c/534371/ i've addressed the comments. let me know if any further changes are required :)
22:55:02 openstackgerrit Takashi NATSUME proposed openstack/python-novaclient master: Microversion 2.61 - List/Show all server migration types https://review.openstack.org/430839
22:55:19 mnaser mriedem: I’ll do that right now
22:56:18 openstackgerrit Takashi NATSUME proposed openstack/nova master: Transform servergroup.addmember notification https://review.openstack.org/541101
23:07:53 openstackgerrit Mohammed Naser proposed openstack/nova stable/queens: Add functional test for deleting BFV server with old attach flow https://review.openstack.org/546803
23:07:54 openstackgerrit Mohammed Naser proposed openstack/nova stable/queens: Ensure attachment_id always exists for block device mapping https://review.openstack.org/546804
23:08:54 mnaser dansmith: mriedem ^
23:09:04 mnaser ill do the pike one but that's not a big priority i guess
23:09:46 dansmith moar?
23:10:06 dansmith I'm going off mriedem's judgment for these attachment things
23:10:51 mnaser dansmith: yeah it was a corner case we discovered upon backporting to pike that we fixed
23:25:46 mnaser mriedem: i think i will have to squash https://review.openstack.org/#/c/546398 + https://review.openstack.org/#/c/545132 in the stable/pike backport. if i put 546398 first, the integration test for the fix fails (because early-delete patch is not there). if i put 545132 before, the patch doesn't pass because 546398 is not there to make it do the right thing(tm)
23:25:49 mnaser is that ok?

Earlier   Later