| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-02-05 | |||
| 14:22:41 | openstackgerrit | Merged openstack/nova stable/pike: Fix the ocata config-reference URLs https://review.openstack.org/515410 | |
| 14:22:48 | openstackgerrit | Merged openstack/nova master: doc: update the GPU passthrough HPC feature entry https://review.openstack.org/539638 | |
| 14:23:03 | openstackgerrit | Merged openstack/nova master: Provide support matrix and doc for VGPU https://review.openstack.org/539266 | |
| 14:27:04 | ameeda | jaypipes_: could you please check this https://review.openstack.org/#/c/526900/ ? | |
| 14:27:33 | openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient master: Remove 2 redundant methods https://review.openstack.org/528325 | |
| 14:36:32 | gibi | bauzas: If you have some free time, then there is a bugfix that needs some review https://review.openstack.org/#/c/525242/ :) | |
| 14:37:14 | openstackgerrit | Lee Yarwood proposed openstack/nova stable/pike: add "--until-complete" option for nova-manage db archive_deleted_rows. https://review.openstack.org/540871 | |
| 14:37:19 | bauzas | gibi: free time is not exactly what I call as my current workload, but I'll ;) | |
| 14:38:37 | gibi | bauzas: Ohh then I was misled by your feeling the breeze comment above :) | |
| 14:39:07 | bauzas | gibi: oh, your patch is crucial https://review.openstack.org/#/c/525242/ | |
| 14:39:54 | mriedem | lyarwood: i think it would be ok to just leave that as a separate refactor/test coverage change on top | |
| 14:40:14 | mriedem | don't really want to hold up https://review.openstack.org/#/c/539739/ since it's breaking cinder 3rd party CI | |
| 14:40:40 | lyarwood | mriedem: yup np | |
| 14:43:18 | gibi | mriedem: hi, I think we have some indirect disagreement in https://review.openstack.org/#/c/528385 I think the duplicate detach will cause confusions in the logs and in the notification interface | |
| 14:46:26 | efried | cdent: This sentence does not parse for me: "That new structure allowed for the creation a POST to /allocations that allows an atomic action to manage allocations for multiple consumers in one request, supporting race-free migration allocations." | |
| 14:51:25 | bauzas | gibi: mriedem: I'm a bit concerned by a bugfix having RPC versions bumped but it's understandable given the upcall issue | |
| 14:51:52 | bauzas | context being https://review.openstack.org/#/c/525242/14 | |
| 14:51:56 | gibi | bauzas: First I tried without the rpc bump but that solution would create yet another cell upcall | |
| 14:52:14 | bauzas | yeah I know | |
| 14:52:39 | bauzas | FWIW, I have a long standing blueprint for passing request specs to computes | |
| 14:52:50 | bauzas | it's purely procedural tho | |
| 14:53:03 | bauzas | I'm fine with approving that change if everything else is ok | |
| 14:53:14 | bauzas | just thinking out loud | |
| 14:54:15 | gibi | yes, we need to agree on such a bump together. Especially as dansmith might plan to do a major rpc version bump before the release | |
| 14:54:46 | dansmith | yeah, unfortunately, that reqspec cleanup would be really nice to get into place | |
| 14:56:30 | mriedem | i got 2 patches in against it for queens! | |
| 14:56:31 | gibi | bauzas: if that helps with the current bugfix, I can offer some help with the request spec blueprint in Rocky | |
| 14:56:33 | mriedem | with great success | |
| 14:57:29 | efried | bauzas: What's a request spec, as distinct from allocation request data? | |
| 14:58:01 | efried | That is, what does a request spec has that an allocation request doesn't? | |
| 14:58:19 | mriedem | gibi: replied in https://review.openstack.org/#/c/528385/ | |
| 14:58:32 | mriedem | efried: all sorts of stuff, | |
| 14:58:38 | mriedem | instance groups, scheduler hints, | |
| 14:58:41 | mriedem | image id | |
| 14:58:46 | mriedem | it's basically the instance | |
| 14:58:53 | mriedem | minus host stuff (and BDMs for some reason) | |
| 14:59:03 | efried | okay, thanks. | |
| 15:00:20 | efried | cause jfyi, we are passing the allocation to the compute (https://review.openstack.org/#/c/511879/) | |
| 15:00:32 | bauzas | efried: originally the request spec is what was passed to the scheduler for filtering purposes | |
| 15:00:44 | mriedem | efried: in this case we need the scheduling hints | |
| 15:00:52 | efried | coo | |
| 15:00:58 | bauzas | efried: it became used for more than just the scheduler | |
| 15:01:20 | bauzas | because we have the original intent of the user + some quirks | |
| 15:01:57 | bauzas | the flaw in that model (and I apologize for not having considered it at the design state) is that we don't know which fields are immutable and which not | |
| 15:02:25 | gibi | mriedem: thanks the reply in https://review.openstack.org/#/c/528385/, I'm OK with the extra kwargs | |
| 15:02:36 | bauzas | so, when you review a change using that RequestSpec object record, try to consider all fields as potentially mutable | |
| 15:03:10 | gibi | ameeda: I think you have a way forward now in https://review.openstack.org/#/c/528385/ | |
| 15:04:08 | openstackgerrit | Peter Hamilton proposed openstack/nova-specs master: Add support for certificate validation https://review.openstack.org/540879 | |
| 15:16:54 | lyarwood | mriedem: https://review.openstack.org/#/c/531422/ - happy to +W this now that https://review.openstack.org/#/c/540680/1 is posted? | |
| 15:25:09 | mriedem | gibi: comments in https://review.openstack.org/#/c/525242/ | |
| 15:26:51 | gibi | mriedem: looking | |
| 15:28:10 | cdent | efried: sorry, was walking. On the POST, do the nearby links clarify things at all? On the allocations have a resource provider with a generation that's here: https://github.com/openstack/nova/blob/master/nova/api/openstack/placement/handlers/allocation.py#L210-L220 | |
| 15:29:01 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Start using openstackdoctheme's extlink extension https://review.openstack.org/540888 | |
| 15:29:01 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Don't use single backticks in man pages https://review.openstack.org/540887 | |
| 15:33:06 | mriedem | gibi: and the functional test below | |
| 15:33:44 | mriedem | dansmith: melwitt: easy quick cells v2 layout caveats update for stuff that is now fixed in queens https://review.openstack.org/#/c/540667/ | |
| 15:34:01 | gibi | mriedem: thanks | |
| 15:34:26 | mriedem | bauzas: this looks nice https://docs.openstack.org/nova/latest/admin/virtual-gpu.html thanks for doing that | |
| 15:40:10 | efried | cdent: I knew what you meant; was just commenting that the sentence didn't grammar for me. | |
| 15:40:38 | cdent | efried: because of the missing "of" or more than that? | |
| 15:41:20 | efried | cdent: TBH, I couldn't even interpolate what was missing/wrong. Let me go read it again... | |
| 15:41:28 | cdent | i've just put the "of" in | |
| 15:41:38 | mriedem | lyarwood: done | |
| 15:42:00 | lyarwood | mriedem: thanks | |
| 15:42:20 | efried | cdent: Oh, yeah, that (or the addition of caffeine since first reading) makes it a lot clearer. | |
| 15:42:37 | efried | cdent: And when you get a chance, you said you could point me to where allocations were managing generations. | |
| 15:42:41 | mriedem | lyarwood: can you hit the other 2 pike doc backports here? https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:stable/pike+topic:bug/1714017 | |
| 15:43:03 | lyarwood | mriedem: sure, looking now | |
| 15:43:09 | purplerbot | ||
| 15:43:09 | cdent | efried: second part of [t vPW] | |
| 15:43:51 | efried | cdent: That's on the return from POST /allocations? | |
| 15:44:02 | mriedem | lyarwood: i'm not sure about this https://review.openstack.org/#/c/540846/ | |
| 15:44:19 | mriedem | it's going to be super annoying if we have to update all doc backports to point at the branch they are on | |
| 15:44:19 | cdent | that gives the Allocation object and rp (with a genaration), which is then checked at https://github.com/openstack/nova/blob/master/nova/objects/resource_provider.py#L2079-L2085 | |
| 15:44:25 | mriedem | we don't do that when we branch | |
| 15:44:29 | cdent | efried: that's on the _start_ of PUT or POST | |
| 15:44:35 | cdent | on the way in, not the way out | |
| 15:44:41 | openstackgerrit | Merged openstack/nova master: Fix invalid UUIDs in test_compute.py https://review.openstack.org/531479 | |
| 15:44:49 | cdent | efried: so the generation is checked, but the client doesn't provide | |
| 15:44:50 | openstackgerrit | Merged openstack/nova stable/pike: Don't update RT in _allocate_network https://review.openstack.org/519942 | |
| 15:44:55 | openstackgerrit | Merged openstack/python-novaclient master: Zuul: Remove project name https://review.openstack.org/540304 | |
| 15:44:56 | cdent | which seems ... useless? | |
| 15:45:04 | efried | cdent: Right. What are we checking the generation *against*? | |
| 15:45:30 | efried | cdent: (BTW, in that code block, I still don't see where the generation is being checked, even internally) | |
| 15:45:43 | cdent | efried: anything that changes from the start of the PUT to the time when _increment_provider_generation is called | |
| 15:46:13 | bauzas | mriedem: thanks for pushing me the idea :-) | |
| 15:46:22 | efried | cdent: Okay, well, I should hope that *all* the APIs are doing *that*, to guarantee atomicity of a single call. But yeah, totally not the point I'm actually making in the bug report. | |
| 15:46:43 | cdent | efried: yes, your bug report is a different concern | |
| 15:46:54 | cdent | and no, we shouldn't be doing that in all single calls | |
| 15:46:58 | cdent | it is pointless in allocations | |
| 15:47:00 | mriedem | lyarwood: are you planning on backporting this entire series to ocata also https://review.openstack.org/#/c/540861/ ? | |
| 15:47:03 | cdent | if the allocation fits, write it | |
| 15:47:04 | mriedem | that's a lot of refactor | |
| 15:47:11 | cdent | efried: but we don't currently do that | |
| 15:47:34 | lyarwood | mriedem: still working out if that's even possible | |
| 15:47:46 | efried | cdent: So that's the point we need to consider. I'm not sure I agree that we should always write the allocation if it still fits. | |
| 15:47:51 | lyarwood | mriedem: yeah, the refactor also fixes a load of connect/disonnect issues | |
| 15:48:22 | lyarwood | mriedem: really I should've landed a smaller bugfix ahead of this on master to backport | |
| 15:49:03 | mriedem | hmmm, | |
| 15:49:04 | mriedem | like what? | |