Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-01-17
15:42:26 gibi ildikov: thanks I think I see now. compute.api._attach_volume() calls _create_volume_bdm that will call the rpc
15:42:40 bauzas tetsuro: you pointed me sean-k-mooney's patch, was that intentional ?
15:42:43 gibi ildikov: in case of attaching a volume to an existing instance
15:43:07 ildikov gibi: yep, that flow sounds correct
15:43:28 gibi ildikov: so it fails before _check_attach_and_reserve_volume is called in _attach_volume
15:43:29 tetsuro bauzas: sorry, just a mistake
15:43:44 gibi ildikov: this is why we don't need an extra check in _check_attach_and_reserve_volume for the attach volume case
15:44:21 tetsuro bauzas: could you have a look at https://review.openstack.org/#/c/531347/ ?
15:44:52 bauzas tetsuro: short patch, will look :)
15:45:42 ildikov gibi: +1
15:45:47 tetsuro bauzas: thanks in advance
15:46:50 smcginnis bauzas: Hey, are you good with the new version on https://review.openstack.org/#/c/534807/ ?
15:47:07 smcginnis bauzas: I assume so, but would be good to have your +1 on there for historical reference.
15:47:22 bauzas smcginnis: I'm on it
15:47:23 bauzas +1d
15:47:28 gibi ildikov: then +2 from me :)
15:47:35 smcginnis bauzas: Thanks!
15:47:39 bauzas smcginnis: thanks for the catch on a minor release
15:48:08 smcginnis bauzas: Wouldn't have been the end of the world, but I think this version is more appropriate. ;)
15:48:28 ildikov gibi: much appreciated
15:48:32 bauzas smcginnis: any release documentation about semver we have could help me
15:48:45 ildikov gibi: will update the tests if I have to upload a new version
15:49:07 bauzas in general, I struggle with identifying .y releases, and just accept .zs
15:49:25 smcginnis bauzas: I only know of the site: https://semver.org/
15:49:26 sean-k-mooney smcginnis: a z stream release would techically have implied abi compatiblity aswell as api compatiblity. we are breaking the abi form 1.8.0 so you were right it should have been 1.9.0
15:50:16 gibi ildikov: cool. that test nit is not super important
15:50:21 smcginnis It's always a tricky and somewhat subjective thing to look at commit messages and determine. ;)
15:57:09 ildikov gibi: noted, tnx :)
16:11:59 openstackgerrit Tetsuro Nakamura proposed openstack/nova master: Test alloc_cands with indirectly sharing RPs https://review.openstack.org/519601
16:12:00 openstackgerrit Tetsuro Nakamura proposed openstack/nova master: Support relay RP for allocation candidates https://review.openstack.org/533437
16:27:13 claudiub stephenfin: hi. can you take a look at the autospec patch? https://review.openstack.org/#/c/447505/ Jenkins says it's ok. thanks!
16:27:35 stephenfin claudiub: Won't get it today, but it's now on the (rather big) list for tomorrow :)
16:27:44 claudiub yeay. :D
16:51:44 openstackgerrit Matthew Booth proposed openstack/nova master: Fix libvirt volume tests passing invalid disk_info https://review.openstack.org/529328
16:51:44 openstackgerrit Matthew Booth proposed openstack/nova master: Pass disk_info dict to libvirt_info https://review.openstack.org/529329
16:51:45 openstackgerrit Matthew Booth proposed openstack/nova master: Expose volume host type and path independent of libvirt config https://review.openstack.org/530786
16:51:45 openstackgerrit Matthew Booth proposed openstack/nova master: Remove redundant swap_volume tests https://review.openstack.org/531179
16:51:46 openstackgerrit Matthew Booth proposed openstack/nova master: Don't generate fake disk_info in swap_volume https://review.openstack.org/530787
16:51:46 openstackgerrit Matthew Booth proposed openstack/nova master: Local disk serial numbers for the libvirt driver https://review.openstack.org/529380
16:57:08 openstackgerrit Merged openstack/nova master: Address nits in I46d483f9de6776db1b025f925890624e5e682ada https://review.openstack.org/530462
16:57:28 openstackgerrit Merged openstack/nova master: Remove unnecessary execute permissions in files https://review.openstack.org/533686
17:10:22 sean-k-mooney bauzas: jaypipes stephenfin just fyi os-vif u-c/g-r patch https://review.openstack.org/#/c/534839/. ye dont need to do anything here ill follow up with infra once zuul comments back
17:10:38 jaypipes sean-k-mooney: ack, thank you sir
17:11:08 bauzas sean-k-mooney: +1d for the sake of consistency with myself
17:26:55 cdent jaypipes, efried : is there code in progress somewhere that includes parent references in the provider summaries of an /allocation_candidates response? (so much code at the moment, hard to find stuff). Or maybe it already done?
17:27:17 jaypipes cdent: I'm currently working on that area
17:27:48 cdent is there anything up in gerrit yet, or is that yet to come?
17:29:01 cdent jaypipes: the reason I ask is because I'm trying to confirm some of my understandings about how this ought to work.
17:29:15 cdent by cooking some gabbi scenarios
17:29:31 cdent but it looks like what I'm trying to do is not quite possible yet
17:29:41 jaypipes cdent: I'm currently working on the alloc candidates stuff with n-r-p in the https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/nested-resource-providers series
17:30:11 cdent yeah, that's where I was looking
17:31:18 cdent jaypipes: is the expectation that in provider summaries, anything with a parent will have parent_provider_uuid field?
17:33:01 jaypipes cdent: yup, zactly.
17:33:17 cdent roger, con aye, ten degrees down bubble
17:37:14 openstackgerrit Jackie Truong proposed openstack/nova master: Account for deprecation of personality files https://review.openstack.org/534849
17:51:33 cdent jaypipes: another, more theorectical, question. sean-k-mooney has a various times talked about moving vcpu and memory to the nested numa node, in which case the allocation records in allocation_requests will be against that. when that happens are we assuming that the nova-scheduler will climb the provider tree hierarchy...
17:52:01 cdent ...recreated from provider summaries, to find the compute node where the actual spawn will happen?
17:53:33 jaypipes cdent: yes
17:53:42 cdent cool
18:04:16 sean-k-mooney cdent: well we schedule resouce based on the leaf nodes to select the set of roots nodes that can fufil the request so yes.
18:04:20 openstackgerrit Jackie Truong proposed openstack/nova master: Add trusted_image_certificates to REST API https://review.openstack.org/486204
18:06:37 openstackgerrit Aditya Vaja proposed openstack/nova master: remove IVS plug/unplug as they're moved to separate plugin https://review.openstack.org/534371
18:08:07 openstackgerrit Jay Pipes proposed openstack/nova master: fix _get_trees_matching_all() when sum children https://review.openstack.org/534866
18:08:17 sean-k-mooney cdent: before we can move the vcpu or other resouce under a numa node i belive there was a distance/ajacency concept that was required also to allow affinity/antiaffinity to be expressed. that is out of scope for the current nested prodiers in queens right jaypipes?
18:08:22 jaypipes efried, gibi: eyeballs needed on the patch ^^
18:08:44 jaypipes sean-k-mooney: totally out of scope, yes
18:09:22 sean-k-mooney jaypipes: and also a prerequec for numa in placement or do you think they can be independent
18:09:41 jaypipes sean-k-mooney: pre-req
18:09:48 jaypipes sean-k-mooney: well....
18:10:05 jaypipes sean-k-mooney: pre-req if we want placement to do certain numa placement calculations...
18:10:31 jaypipes sean-k-mooney: I'm still not entirely convinced that placement should do that for NUMA nodes (as opposed to a weigher in scheduler doing it)
18:10:35 sean-k-mooney hehe pre-req or ptg topic :), i guess you could delegate to the numa topology filter
18:11:17 jaypipes sean-k-mooney: if you can only imagine the SQL in https://review.openstack.org/#/c/534866/1/nova/objects/resource_provider.py getting 10X more complicated ...
18:12:59 sean-k-mooney jaypipes: ya i have not looked at how nrp has evovled both the sql and usage of placement to have developed an intuition regarding if the logic should be kept in the schduler and just track numa info in placement
18:13:41 jaypipes ++
18:15:28 openstackgerrit Merged openstack/nova master: PowerVM driver: ovs vif https://review.openstack.org/422512
18:15:38 openstackgerrit Merged openstack/nova master: make unit tests compatible with os-vif 1.8.0 https://review.openstack.org/533766
18:17:01 sean-k-mooney oh ^ merged cool. now just waiting for https://review.openstack.org/#/c/534839/ and we are set for os-vif
18:18:07 cdent jaypipes: is your work also dealing with adding support for 'required' or is that rocky?
18:18:21 jaypipes cdent: for n-r-p?
18:18:44 cdent traits in n-r-p, yeah
18:19:17 sean-k-mooney cdent: im not sure you need special logic unless you want traits to inherit down the tree
18:19:48 cdent as far as I can tell 'required' is not merged in any fashion, yet, n-r-p or not
18:19:59 cdent (at the HTTP level)
18:21:11 sean-k-mooney cdent: i think alex xu was mainly working on required/traits
18:21:27 cdent yeah, just pulling up his pending stuff
18:21:51 sean-k-mooney cdent: :) you mean https://review.openstack.org/#/q/owner:hejie.xu%2540intel.com+status:open
18:22:05 cdent yes
18:22:57 cdent Basically I'm trying to figure out how much is not done
18:23:33 cdent and it looks like maybe it could just be a matter of it not being turned on
18:24:48 sean-k-mooney cdent: i think this is one of the main blockers for traits https://review.openstack.org/#/c/479776/
18:27:13 openstackgerrit Jackie Truong proposed openstack/nova master: Add trusted_image_certificates to REST API https://review.openstack.org/486204
18:27:49 sean-k-mooney cdent: im not sure how much of that has been pulled into https://review.openstack.org/#/q/topic:bp/granular-resource-requests+(status:open+OR+status:merged) however
18:31:49 sean-k-mooney efried: does https://review.openstack.org/#/c/517757/1 superceed https://review.openstack.org/#/c/479776 i belive the numbered RequireN sysntax does
18:36:51 mgoddard efried: whilst trying to get myself out of a hole caused by something a bit like https://bugs.launchpad.net/nova/+bug/1714248, I have hit this odd issue: http://paste.openstack.org/show/646537/hl
18:36:53 openstack Launchpad bug 1714248 in OpenStack Compute (nova) pike "Compute node HA for ironic doesn't work due to the name duplication of Resource Provider " [High,In progress] - Assigned to Matt Riedemann (mriedem)
18:37:30 mgoddard efried: can't delete a resource provider because it has a foreign key that points back to itself
18:38:03 mgoddard efried: I did try this via placement API first, with the same result
18:42:04 cdent mgoddard: that's new and different and fixed by...

Earlier   Later