Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-02-17
20:14:18 mriedem and the context doesn't have read_deleted='yes' or whatever
20:14:19 mriedem ok
20:14:20 mriedem yup
20:14:55 mriedem ok so in mitaka we have soft deleted compute_nodes, online data migration doesn't get them b/c of model_query, but 330 blocked migration pulls them in b/c it's not using the orm and it explodes
20:15:01 mriedem because we don't filter on deleted==0
20:15:07 melwitt the only gotcha would be if it were valid for any of those objects: compute nodes, aggregates, pci devices to be read in a deleted state
20:15:14 mriedem which is your fix
20:15:19 melwitt right
20:15:46 mriedem geez melwitt your commit message could have been several more paragraphs of explanation :)
20:16:11 melwitt because this fix would leave the soft-deleted unmigrated records around. the alternate fix would migrate potentially a lot of soft-deleted things for no reason
20:16:40 mriedem we don't read deleted on any of those records as far as i know
20:16:48 melwitt mriedem: yeah, sorry about that. I can add more info to it
20:16:56 mriedem melwitt: i'm only half joking
20:16:57 mriedem :)
20:16:58 melwitt yeah, that's what I thought too
20:17:08 mriedem superdan: do you know if we read deleted any of those?
20:17:21 superdan on compute nodes?
20:17:32 melwitt aggregates and pci devices also
20:17:38 mriedem i have like 1% knowledge of the pci_devices code
20:17:49 superdan definitely not pci_devices
20:18:32 superdan melwitt: if we did read aggregates, we'd generate on load: https://github.com/openstack/nova/blob/stable/newton/nova/objects/aggregate.py#L288
20:18:34 superdan so we're okay
20:18:57 melwitt ah, nice
20:20:38 mriedem we mostly read deleted instances
20:20:43 mriedem so yeah i think we're ok there
20:21:18 mriedem and flavors
20:24:23 mriedem just testing the test quick
20:24:28 mriedem do we want/need a release note for this?
20:24:32 mriedem since it's going to be backported
20:24:54 mriedem like even a "fixes" note or "upgrade" note? as i said earlier, if you hit this you can't get past 330 anyway
20:25:02 mriedem so this is more about fixing it before people hit it
20:25:23 mriedem last time we backported a db schema migration to stable and released it i got some flak from red hat
20:27:56 superdan this isn't backporting a new one,
20:28:00 superdan it's just tweaking an old one,
20:28:02 mriedem i know
20:28:03 superdan which we did in ocata too
20:28:06 mriedem melwitt: see my comment in the test
20:28:16 superdan but release notes are cool
20:29:03 mriedem so the test doesn't fail on the aggregate b/c you're using an object, which creates the aggregate in the api db
20:29:11 mriedem i don't know why it doesn't fail on the compute node when i remove your fi
20:29:13 mriedem *fix
20:29:15 melwitt oh, hm
20:29:21 superdan mriedem: it's because it has a uuid
20:29:24 superdan mriedem: if you reate it
20:29:27 mriedem ah
20:29:27 superdan I'm thinking
20:29:29 mriedem yup
20:29:35 superdan she'd have to use the db api
20:29:37 melwitt okay, let me fix this to put things in the old db
20:30:17 mriedem yup
20:30:24 mriedem pci device object one is fine
20:30:31 mriedem but use the db api directly for the compute notes and aggregates
20:30:44 melwitt right
20:30:44 mriedem fwiw i've put a hold on https://review.openstack.org/#/c/434928/ until we get this backported and merged
20:33:46 diana_clarke melwitt: Here's an easy review that fixes up the mismatch in metavars you noticed a couple of weeks back: https://review.openstack.org/#/c/431714/
20:34:06 diana_clarke Have a blast at the PTG, folks!
20:42:25 openstackgerrit Eric Fried proposed openstack/nova master: PowerVM Driver: console https://review.openstack.org/409402
20:42:50 diana_clarke melwitt: And here are some manual testing notes that should make that easier to review: http://paste.openstack.org/show/598292/
20:43:15 diana_clarke (anyhoo, no pressure, just cleaning up my TODO list)
20:46:18 melwitt diana_clarke: okay, thanks
20:46:30 diana_clarke lol, I just read "python-wsmanclient" as "python-womanclient"...
20:57:15 mriedem walmartclient
21:11:28 openstackgerrit Anusha Unnam proposed openstack/nova master: Stop using mox from tests/unit/test_service.py https://review.openstack.org/334395
21:24:56 openstackgerrit Dan Smith proposed openstack/nova master: DNM: Find InstanceUnknownCell instances https://review.openstack.org/435600
21:35:41 openstackgerrit Dan Smith proposed openstack/nova master: DNM: Find InstanceUnknownCell instances https://review.openstack.org/435600
21:38:04 openstackgerrit Sujitha proposed openstack/nova master: Remove mox from nova/tests/unit/api/openstack/compute/test_create_backup.py https://review.openstack.org/435610
21:38:59 openstackgerrit Sean Dague proposed openstack/nova master: Verify project id for flavor access calls https://review.openstack.org/435432
21:38:59 openstackgerrit Sean Dague proposed openstack/nova master: Verify project_id when quotas are checked https://review.openstack.org/435010
21:42:25 openstackgerrit Anusha Unnam proposed openstack/nova master: Remove mox from nova/tests/unit/objects/test_service.py https://review.openstack.org/435611
22:09:36 openstackgerrit Sean Dague proposed openstack/nova master: Verify project id for flavor access calls https://review.openstack.org/435432
22:09:36 openstackgerrit Sean Dague proposed openstack/nova master: Verify project_id when quotas are checked https://review.openstack.org/435010
22:09:45 openstackgerrit melanie witt proposed openstack/nova master: Skip soft-deleted records in 330_enforce_mitaka_online_migrations https://review.openstack.org/435546
22:24:20 melwitt mriedem, superdan: fixed the unit test ^
22:24:28 mriedem yup just pulling it down
22:35:58 jamielennox mriedem: you want auth_type, auth_section allows you to move auth options to a new ini section so you can reuse them
22:37:11 mriedem jamielennox: yup, thanks. turns out i don't need that check anyway
22:37:15 mriedem but good to know
22:39:47 mriedem melwitt: superdan: +2 on that fix now
22:39:56 mriedem melwitt: want to start the backport shenanigans?
22:40:10 melwitt mriedem: thanks. yeah, sure
23:04:25 superdan guh
23:04:31 superdan I see why the cells job is failing
23:06:30 mriedem what it be?
23:07:10 superdan that change makes us skirt around the cellsv1 replication
23:07:16 superdan so we can only target if we're non-cellsv1
23:07:21 openstackgerrit Dan Smith proposed openstack/nova master: Make servers API use cell-targeted context https://review.openstack.org/433362
23:07:22 openstackgerrit Dan Smith proposed openstack/nova master: DNM: Find InstanceUnknownCell instances https://review.openstack.org/435600
23:07:34 superdan which we did in the get and list case, it was just a while ago and I kinda forgot about why that matters
23:08:05 melwitt ah, yeah
23:25:43 mriedem superdan: +2 on those backports https://review.openstack.org/#/q/I2285005098b7dab7753366f53667ff8d4532d668,n,z
23:25:54 mriedem and now i'm getting the f out of here
23:43:38 openstackgerrit Dan Smith proposed openstack/nova master: Make servers API use cell-targeted context https://review.openstack.org/433362
23:48:35 openstackgerrit Anusha Unnam proposed openstack/nova master: Remove mox from nova/tests/unit/objects/test_service.py https://review.openstack.org/435611
23:56:54 openstackgerrit Sujitha proposed openstack/nova master: Remove mox from nova/tests/unit/api/openstack/compute/test_create_backup.py https://review.openstack.org/435610
#openstack-nova - 2017-02-18
00:16:00 openstackgerrit Pushkar Umaranikar proposed openstack/nova master: Add debug logs to output host weights for weighers https://review.openstack.org/385636
12:16:19 openstackgerrit Tomi Juvonen proposed openstack/nova-specs master: Planned host maintenance https://review.openstack.org/310510
13:48:03 narasimha_SV https://docs.openstack.org/kilo/config-reference/content/list-of-compute-config-options.html
13:48:42 narasimha_SV apart fromsetting compute monitor to compute_monitors =nova.compute.monitors.cpu.virt_driver
13:48:51 narasimha_SV which gives metrics of CPU
13:49:20 narasimha_SV is there any other way to get host metrics

Earlier   Later