Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-10-16
19:44:29 efried cdent Aha, perfect, so yeah, this all makes total sense when traits are NOT in the picture at all.
19:44:57 efried And breaks down pretty handily once traits are introduced.
19:45:39 cdent so yeah, if a compute node is satisfy VCPU, and has a trait that is “about disk”
19:45:42 cdent things do get very weird
19:45:50 efried btw, I'm asserting that the scenario we walked through above is representative, not necessarily realistic. That is, even if we say "that'll never happen", something with the same characteristics is sure to pop up.
19:46:17 cdent yeah, I get that
19:46:27 cdent it’s better to be robust in the face of possibilities
19:46:52 efried One could imagine similar scenarios where network resources are provided by both the compute node and some shared RP.
19:46:53 cdent for a while the stake you mention above did exist, but then it went away, one option is for it to come back
19:48:11 efried Okay, what about when I have my ephemeral boot disk via local storage and my volumes shared (as they're likely to be)
19:48:33 efried Could solve it by modeling those storage types with different resource classes. But ew.
19:49:13 efried And in any case, there we get into a situation where we need to request separate chunks of storage inventory from separate RPs.
19:49:34 cdent yeah, this is not the first time that question comes up. I don’t know the answer right now.
19:50:27 efried But it sounds like the change set you pointed out means we do actually need an answer for this right now.
19:51:05 cdent efried: except that I’m not sure that code is doing anything
19:51:53 cdent it may very well be
19:52:08 efried cdent Well, it looks to me like it's letting you use aggregates.
19:52:45 cdent yes, but aggregates aren’t yet used in any particular way
19:53:23 efried Meaning there's no code that's creating aggregations yet?
19:53:29 cdent unless there are lots of clients of placement out there that we are not away of (which there very well may be, vmware has an entire feature set based around placement that has nothing to do with nova)
19:53:38 efried whee
19:53:49 cdent upstream, as far as I know, don’t think so, but I don’t know
19:55:49 mriedem wtf, why would hasattr return True on something and then immediately after that if i try to use it, i get an AttributeError?
19:56:16 efried mriedem Overrides?
19:56:21 cdent is it an ovo with a coerce?
19:56:27 mriedem no, fakelibvirtr
20:01:11 mriedem it's not even a Mock so getattr shouldn't be getting in the way
20:06:00 cfriesen are there any gotchas around running two instances of nova-compute one one system, one for local kvm instances and one for managing separate baremetal instances (for use with ironic)?
20:13:11 cfriesen I guess it might cause problems if the binary and hostname are the same for the two of them
20:22:29 mriedem my god,
20:22:40 mriedem when you find out you've been bashing your brains for an hour b/c of a copy paste error
20:22:43 cdent it’s full of stars
20:27:14 openstackgerrit Matt Riedemann proposed openstack/nova stable/ocata: libvirt: add check for VIR_DOMAIN_BLOCK_REBASE_COPY_DEV https://review.openstack.org/512406
20:27:15 mriedem artom: ^
20:54:35 openstackgerrit Chris Dent proposed openstack/nova master: [placement] Enable PUT dict to /allocations/{consumer_uuid} https://review.openstack.org/510625
20:54:36 openstackgerrit Chris Dent proposed openstack/nova master: [placement] POST /allocations to set allocations for >1 consumers https://review.openstack.org/500073
20:54:36 openstackgerrit Chris Dent proposed openstack/nova master: [placement] Document the changes to PUT /allocations/{consumer_uuid} https://review.openstack.org/510626
20:54:37 openstackgerrit Chris Dent proposed openstack/nova master: [placement] Include project_id and user_id in GET /allocations https://review.openstack.org/512421
20:54:37 openstackgerrit Chris Dent proposed openstack/nova master: Include project_id and user_id in AllocationList.get_all_by_consumer_id https://review.openstack.org/512420
20:55:08 mriedem dansmith: melwitt: need an opinion on a backport scenario for https://review.openstack.org/#/c/505771/
20:55:26 mriedem i need to do the same fix as in ^ but generically since we also need that fix for resize, unshelve and evacuate
20:55:36 mriedem i don't want to rename the method and then backport
20:56:02 mriedem so i can either copy the same fix outside the method in the 4 different places and backport it, or move the fix into the method with a todo to rename later
20:56:11 mriedem the rename would be master only
20:57:16 mriedem i'm leaning toward the latter - fix in one place, rename later
21:05:30 mgagne mriedem: what's the proper channel to get a blueprint approved/reviewed? https://blueprints.launchpad.net/nova/+spec/rebuild-ironic-config-drive
21:05:58 mriedem mgagne: generally during the weekly nova meeting if it's a specless blueprint
21:07:17 mriedem i think this would be ok as a specless feature parity bp
21:08:30 mriedem mgagne: so the issue is you create an ironic node instance with a config drive, and then when you rebuild the instance in nova, the new instance doesn't have a config drive, correct?
21:08:32 mgagne will add to open discussion for next meeting
21:08:39 mgagne mriedem: yes
21:08:48 mriedem ok, really a bug fix
21:09:14 mgagne mriedem: looks like someone from nova team had a different opinion in the bug report
21:09:19 mgagne and suggested a blueprint
21:10:09 mgagne https://bugs.launchpad.net/ironic/+bug/1575935/comments/10
21:10:10 openstack Launchpad bug 1575935 in Ironic "Rebuild should also accept a configdrive" [Wishlist,In progress] - Assigned to Mathieu Gagné (mgagne)
21:10:12 mriedem "reno: nova rebuild of ironic instances will no longer trash your config drive. you're welcome."
21:10:41 mriedem ok, well, i'm cool with fixing it either way
21:12:28 mgagne mriedem: I'm not sure what you were referring to? any link?
21:12:53 mriedem i was being facetious,
21:13:00 mriedem i.e. how would one write a release note for this as a 'feature'?
21:13:09 mriedem anyway, bp approved
21:13:46 mgagne mriedem: Configdrive is now regenerated when rebuilding an Ironic instance (matching behaviour of other virt drivers)
21:14:09 mgagne we can say it's done for feature parity
21:21:23 openstackgerrit Matt Riedemann proposed openstack/nova master: Don't ignore original retried hosts when moving an instance https://review.openstack.org/505771
21:21:25 mriedem tasker: updated and simplified that change to also fix the issue for resize, unshelve and evacuate ^
21:21:29 mriedem zioproto: ^
21:27:55 dansmith mriedem: not sure what you're asking: if we should unify all the places into one call and backport that change?
21:28:07 mriedem dansmith: yeah that's just what i did
21:33:54 mriedem totally unrelated, but when querying stuff out of the db, if a sort key/direction is not specified, is it just left up to the rdbms?
21:34:29 mriedem zzzeek: ^?
21:34:45 mriedem or is there oslo.db magic happening somewhere by default?
21:38:19 dansmith I think it's just undefined/insert order if you don't ask
21:38:26 dansmith asking for it to be sorted when you don't care would just be overhead
21:38:41 mriedem yeah, trying to reply to some questions on this spec for paging migrations
21:38:47 mriedem and sort orders since we don't impose one today
21:39:03 mriedem and if the changes-since filter should be on the created_at or updated_at fields
21:40:32 melwitt mriedem: yeah, I think what you did makes sense. do the change in one place and rename the method in a separate change that won't be backported
21:59:38 edleafe What is the ovo trick for testing if an object's field is populated, without getting the "Cannot load 'foo' in the base class" error?
22:01:05 melwitt edleafe: there's obj_attr_is_set() for checking if a field has been set
22:01:09 mriedem 'in'
22:01:17 mriedem if 'foo' in bar
22:01:37 mriedem uses obj_attr_is_set() under the covers
22:01:48 melwitt yeah, use that
22:01:48 melwitt oh. I keep forgetting you added that
22:02:51 openstackgerrit Mathieu Gagné proposed openstack/nova master: Regenerate and pass configdrive when rebuild Ironic nodes https://review.openstack.org/503088
22:03:42 edleafe mriedem: melwitt: thanks
22:04:06 mriedem edleafe: api wg question,
22:04:11 mriedem does changes-since come up in guidelines?
22:04:16 mriedem changes-since filtering i mean
22:04:50 mriedem and if so, are there guidelines on which column to filter it on? we use updated_at for filtering instances by changes-since
22:07:11 edleafe mriedem: not sure off the top of my head
22:08:50 edleafe mriedem: ok, a quick review doesn't turn up anything
22:09:00 edleafe that means it's ripe for defining :)
22:09:45 edleafe I can add it to the discussion for Thursday's meeting
22:09:52 mriedem not a big deal,
22:10:15 mriedem someone is just proposing to add changes-since filtering to the os-migrations api and i said they should filter on updated_at to be consistent with changes-since on filtering instances
22:10:40 edleafe well, updated_at sounds like the most logical choice
22:36:12 openstackgerrit Eric Fried proposed openstack/nova master: Send Allocations to spawn https://review.openstack.org/511879
22:36:13 openstackgerrit Jackie Truong proposed openstack/nova master: Implement certificate_utils https://review.openstack.org/479949

Earlier   Later