Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-09-06
21:09:39 dansmith mnaser: yeah
21:10:24 mnaser or i'll add a note as well
21:11:11 mriedem TODO i'd say
21:11:19 mnaser damn
21:11:21 mnaser um
21:11:21 mriedem you'll also have to keep track of which cells each instance goes into
21:11:23 mnaser i dont know what cell
21:11:24 mnaser yeah
21:11:32 mnaser i'll have to figure out a clean way to do that
21:11:36 mriedem just save that into a dict in the first loop
21:11:44 mnaser alright
21:12:47 mriedem everything already goes into that instances list variable, and is sometimes None for things that failed, but for things that did get mapped, i think you could map the instance uuid to the cell in a dict and then use those 2 variables to do the mapping in the failure block
21:12:57 mriedem s/mapped/created/
21:13:57 mriedem hmm, i think i just realized another bug here
21:14:03 dansmith um
21:14:06 dansmith hah yesh
21:14:12 mriedem we're always mapping to the last cell
21:14:14 dansmith yeah
21:14:17 mriedem :)
21:14:27 dansmith oof
21:15:27 dansmith mnaser: so you might want to stack your change on top of the fix for that bug
21:15:35 dansmith which I can cook up
21:15:54 mnaser okay cool, ill work on the stuff to set mappings
21:17:45 dansmith mriedem: I'm guessing that this would have uncovered that if I had ever finished it: https://review.openstack.org/#/c/452006/
21:18:41 dansmith ah, no,
21:18:44 mriedem https://bugs.launchpad.net/nova/+bug/1715493
21:18:44 dansmith that's the migrate one
21:18:45 openstack Launchpad bug 1715493 in OpenStack Compute (nova) "Instances always get mapped into the last processed cell in conductor" [High,Triaged]
21:19:55 dansmith I guess this bug came from splitting that loop amirite?
21:19:58 mriedem yup
21:20:11 mriedem so when we combine we can remove both workarounds
21:20:17 dansmith yeah
21:20:24 openstackgerrit Chris Dent proposed openstack/nova master: WIP: [placement] POST /allocations to set allocations for >1 consumers https://review.openstack.org/500073
21:20:25 melwitt ugh :(
21:20:25 openstackgerrit Chris Dent proposed openstack/nova master: [placement] Allow _set_allocations to delete allocations https://review.openstack.org/501051
21:20:35 mriedem plus, if mnaser's patch creates a instance -> cell mapping dict to keep track, the 2nd loop could use that too
21:20:43 dansmith yeah
21:20:52 openstackgerrit Dan Smith proposed openstack/nova master: Track which cell each instance is created in and use it consistently https://review.openstack.org/501452
21:21:02 dansmith gonna work on tests, but pushed this up in case I have to run ^
21:25:16 dansmith hmm, we kinda have a test for this, I'm not sure why it's not failing
21:29:32 mnaser dansmith is it okay that when i try to do git review with your patch below mine, it mentions that it will submit two commits?
21:29:44 dansmith mnaser: yep
21:29:49 mnaser okay cool
21:29:53 openstackgerrit Mohammed Naser proposed openstack/nova master: Ensure instance mapping is updated in case of quota recheck fails https://review.openstack.org/501408
21:30:05 mnaser oh it didnt send both, nice. that's leveraging your cell_instance_cache
21:30:15 mnaser oh i should update the commit msg
21:30:22 melwitt mnaser: I usually double check to make sure the commit hash of the dependent change is the same as what shows on the review being rebased upon
21:30:41 melwitt that's how you can tell whether it will push more than just your change
21:30:51 mnaser melwitt oh, if it matches it wont submit it? i'm just terrified of the embarassement that happens sometimes when an irc channel gets spammed :p
21:31:10 mriedem mnaser: git review -R also avoids the rebase of the base changes
21:31:19 melwitt mnaser: yeah. if the hash hasn't changed it won't submit it. so you can know before you do it
21:31:28 mriedem melwitt: btw i think i figured out the git review / rebase author change thing,
21:31:36 mriedem it happens when rebasing on a series that involves a merge conflict,
21:31:46 mriedem so you rebase, merge conflict, fix, git add / commit, git rebase continue
21:31:50 mriedem and that has changed the author
21:32:05 openstackgerrit Mohammed Naser proposed openstack/nova master: Ensure instance mapping is updated in case of quota recheck fails https://review.openstack.org/501408
21:32:19 melwitt weird. I feel like I've rebased other ppl's changes with merge conflict and it didn't change the author before
21:32:32 mnaser there, that should have most of it covered
21:32:34 melwitt but, that's great data to go on. I'll pay attention next time
21:33:31 dansmith okay, the test I thought we had was build not schedule_and_build
21:33:42 dansmith I have to board soon, but I can work on tests for this on the plane unless someone else wants to?
21:34:28 mnaser i'd help but it sits outside my relm of knowledge :) have a safe flight though dansmith
21:35:14 dansmith mnaser: np, thanks
21:35:38 mnaser time to go figure out why horizon pike is logging users out now :> thanks for the help everyone, i'll keep an eye out on reviews for the patch
21:36:01 mriedem mnaser: one issue in your patch
21:36:09 mnaser sure whats up
21:36:15 mnaser oh /me looks
21:36:16 mriedem comment inline
21:36:57 melwitt dansmith: I can write the tests for your change if you want to not do it on the plane
21:38:55 mnaser mriedem took me a while to figure out what you meant till i realized that super hidden for loop in the top lol
21:39:31 mriedem :) everything else looks good
21:42:07 mriedem sdague: low hanging fruit review https://review.openstack.org/#/c/499001/
21:42:16 mriedem skip more cells v1 random failure tests
21:42:27 openstackgerrit Mohammed Naser proposed openstack/nova master: Ensure instance mapping is updated in case of quota recheck fails https://review.openstack.org/501408
21:44:07 openstackgerrit Eric Fried proposed openstack/nova master: nova.utils.get_ksa_adapter() https://review.openstack.org/488137
21:44:08 openstackgerrit Eric Fried proposed openstack/nova master: Get auth from context for glance endpoint https://review.openstack.org/490057
21:45:32 openstackgerrit Merged openstack/nova master: First attempt at adding a privsep user to nova itself. https://review.openstack.org/459166
21:48:42 openstackgerrit Eric Fried proposed openstack/nova master: nova.utils.get_ksa_adapter() https://review.openstack.org/488137
21:52:39 openstackgerrit Chris Dent proposed openstack/nova-specs master: Add a spec for POST /allocations in placement https://review.openstack.org/499259
21:55:52 openstackgerrit Matt Riedemann proposed openstack/nova stable/pike: Add release note for force live migration allocations https://review.openstack.org/501477
21:58:13 openstackgerrit Eric Fried proposed openstack/nova master: Get auth from context for glance endpoint https://review.openstack.org/490057
22:05:07 openstackgerrit Eric Fried proposed openstack/nova master: Use ksa adapter for placement conf & requests https://review.openstack.org/492247
22:15:09 mriedem melwitt: i'm going to step away to make dinner but if you get that test written ping me and i'll take a look tonight
22:16:27 melwitt mriedem: okay. I wasn't sure if dansmith got my message before boarding. so I wasn't sure if he's already working on it on the plane
22:46:53 mnaser do live migration jobs often fail
22:46:55 mnaser or did i break something :>
22:47:06 mnaser Details: (LiveAutoBlockMigrationV225Test:test_live_block_migration) Server 44884988-d3f2-4e32-a367-4962acda6346 failed to reach ACTIVE status and task state "None" within the required time (196 s). Current status: BUILD. Current task state: scheduling.
23:21:20 openstackgerrit Merged openstack/nova master: Amend uuid4 hacking rule https://review.openstack.org/488315
23:21:54 openstackgerrit Merged openstack/nova stable/pike: Fix broken link https://review.openstack.org/501403
23:40:07 mriedem mnaser: ah ./nova/conductor/manager.py:1017:20: F821 undefined name 'instance_mapping_cache'
23:40:19 mnaser oopsie
23:40:50 mriedem that's a bug in dan's change
23:40:52 mnaser oh thats not my path
23:40:53 mnaser yeah
23:41:02 mriedem i'll fix quick
23:44:17 openstackgerrit Matt Riedemann proposed openstack/nova master: Ensure instance mapping is updated in case of quota recheck fails https://review.openstack.org/501408
23:44:17 openstackgerrit Matt Riedemann proposed openstack/nova master: Track which cell each instance is created in and use it consistently https://review.openstack.org/501452
23:58:42 openstackgerrit Merged openstack/nova master: Add release note for force live migration allocations https://review.openstack.org/501314
#openstack-nova - 2017-09-07
00:00:40 openstackgerrit Michael Still proposed openstack/nova master: Move nbd commands to privsep. https://review.openstack.org/500351
00:00:41 openstackgerrit Michael Still proposed openstack/nova master: Move lvm handling to privsep. https://review.openstack.org/495516

Earlier   Later