| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-07-26 | |||
| 18:50:19 | efried | However, it's possible that powervm is blocked on "client side nrp stuff we haven't exercised yet". | |
| 18:50:20 | mriedem | i think we just, generally, speak in libvirt terms, given the adoption | |
| 18:50:26 | mriedem | and how libvirt is the first to have most any new feature | |
| 18:50:51 | dansmith | yes, I understand that boutique virt drivers are not blocked in the strictest sense | |
| 18:50:56 | dansmith | from a nova project perspective, | |
| 18:51:08 | dansmith | libvirt being blocked means it's blocked, IMHO | |
| 18:51:45 | mriedem | now, cyborg or whatever could maybe totally use whatever is available today, idk | |
| 18:51:47 | efried | We should consider nrp to be the framework, and things like NUMA & VGPU support to be the features we're talking about having "nova" support. | |
| 18:52:29 | efried | So from that point of view, NUMA & VGPU are blocked "in nova" (i.e. libvirt). But the nrp framework is (probably, mostly) ready, pending working out the kinks on the client side in the scheduler. | |
| 18:52:37 | mriedem | melwitt: let me guess, you're writing up the nova release highlights thing in the releases repo? | |
| 18:53:08 | mriedem | efried: that's a good summary i think | |
| 18:53:20 | dansmith | efried: IMHO, nova's notion of NRP being done would be that existing resources we're currently lying about are converted to the new thing | |
| 18:53:27 | mriedem | melwitt: if/when you do, you should definitely mention the is_bfv disk_gb tracking stuff finally being fixed | |
| 18:53:35 | dansmith | efried: otherwise we've just created a data structure and protocol to speak about it, | |
| 18:53:40 | dansmith | but no user will perceive anything as being different | |
| 18:53:46 | efried | mriedem: Example of client-side allocation stuff that's broke: put_allocations in the report client only works for the cn rp and ignores the rest. | |
| 18:53:50 | dansmith | therefore.. we've basically accomplished nothing visible | |
| 18:54:18 | mriedem | right so lots of *nova* isn't nrp-aware | |
| 18:54:27 | mriedem | like lots of nova isn't shared storage provider aware | |
| 18:54:31 | mriedem | you can model that in placement externally, | |
| 18:54:33 | mriedem | but nova won't honor it | |
| 18:54:44 | efried | right, swhat I'm sayin. | |
| 18:55:11 | efried | So nrp in placement is done, but nrp (i.e. the framework that the actual features need) in nova is not. | |
| 18:55:20 | dansmith | yeah, so call some specific blueprint done if you want, | |
| 18:55:39 | dansmith | but when I get asked "can we do any of the things that NRP is supposed to enable for us?" I still have to answer no | |
| 18:56:02 | melwitt | mriedem: just trying to get an idea of what's currently going on there ... and taking notes | |
| 18:56:12 | efried | dansmith: Yes, I will agree with that, stipulating s/we/nova/ (as opposed to some other placement consumer). | |
| 18:56:32 | dansmith | efried: yep, been trying to tag my comments with nova when I say stuff like that | |
| 18:56:44 | dansmith | that's my intent at least | |
| 18:57:14 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Fix accumulated nits from port binding for live migration series https://review.openstack.org/583994 | |
| 18:57:16 | mriedem | efried: ^ | |
| 18:57:42 | efried | ack | |
| 18:57:54 | mriedem | dansmith: i got our lazy load thing right here too https://review.openstack.org/#/c/583994/2/nova/compute/manager.py | |
| 18:59:00 | dansmith | mriedem: ack | |
| 18:59:15 | mriedem | i intentionally said that jersey-style | |
| 19:01:45 | mriedem | unrelated, but it seems that stephen's bottom change 564440 is perpetually stuck in the check queue or something weird | |
| 19:01:51 | mriedem | it's in both queues at the same time at least twice today | |
| 19:03:42 | mriedem | and it's already failed in the gate again, | |
| 19:03:47 | mriedem | and queued in check | |
| 19:05:33 | dansmith | cripes | |
| 19:05:48 | dansmith | and some guys keep naking the later patches too | |
| 19:05:49 | dansmith | GAWD | |
| 19:06:26 | mriedem | i nak while nak'ed | |
| 19:06:35 | dansmith | naked nak? | |
| 19:06:42 | mriedem | too far | |
| 19:10:13 | melwitt | mriedem: so the shared storage counting local usage for volume-backed instances saga is finally over? | |
| 19:11:28 | mriedem | not shared storage | |
| 19:11:31 | mriedem | volume-backed root_gb | |
| 19:12:05 | mriedem | https://review.openstack.org/#/q/(status:open+OR+status:merged)+project:openstack/nova+branch:master+topic:bug/1469179 | |
| 19:12:06 | melwitt | oh, I think I got that confused with shared storage reporting. that's still not done | |
| 19:12:26 | melwitt | ok | |
| 19:12:38 | mriedem | we will no longer claim root_gb for volume-backed instances against DISK_GB inventory in placement for new instances | |
| 19:12:41 | mriedem | and heal on moves | |
| 19:13:33 | dansmith | shared ephemeral reporting is also fixed in rocky I think | |
| 19:13:42 | mriedem | via placement? | |
| 19:13:48 | dansmith | we no longer report DISK_GB in inventory if MISC_SHARES_VIA_AGGREGATE is set | |
| 19:14:16 | mriedem | for the libvirt driver* | |
| 19:14:46 | mriedem | i wouldn't talk much about that in rocky though, because we have several places in the code that don't handle that | |
| 19:14:50 | dansmith | not sure, I would have expected that to be applied to each as they converted to update_provider_tree(), unless the others are still using get_inventory() | |
| 19:15:08 | openstackgerrit | Dan Smith proposed openstack/nova master: Assorted cleanups from numa-aware-vswitches series https://review.openstack.org/582651 | |
| 19:15:09 | openstackgerrit | Dan Smith proposed openstack/nova master: Add additional functional tests for NUMA networks https://review.openstack.org/585385 | |
| 19:15:13 | mriedem | yeah that's only libvirt and ironic that implement upt | |
| 19:15:26 | mriedem | and it doesn't apply to ironic | |
| 19:15:58 | dansmith | I didn't follow that change in, | |
| 19:15:58 | mriedem | but like, we don't have any docs on it, we don't have any CI on it, and we have lots of places that aren't aware of it (like move operations) | |
| 19:16:11 | dansmith | but I was hoping that if we're removing that inventory that we handle it the other places we need, | |
| 19:16:24 | dansmith | otherwise we're not going to be able to boot anything on computes that are configured that way :) | |
| 19:16:40 | mriedem | https://review.openstack.org/#/c/560459/ | |
| 19:17:11 | mriedem | well, we should, | |
| 19:17:20 | mriedem | the allocations could go against the provider that has the DISK_GB inventory | |
| 19:17:22 | dansmith | I'm not sure that move ops need to change? | |
| 19:17:31 | mriedem | which is in the MISC_SHARES_VIA_AGGREGATE relationship | |
| 19:17:34 | dansmith | right | |
| 19:17:48 | dansmith | so I'm not sure what you're saying... | |
| 19:17:49 | mriedem | there was something i look up every time this comes up that i know is broken | |
| 19:18:01 | mriedem | i've beaten cdent over the head with a few times already :) | |
| 19:18:32 | dansmith | have we updated the ceph job to do the right thing? | |
| 19:18:41 | dansmith | that'd be a good way to either show it's broken or show that it works | |
| 19:18:53 | dansmith | and you should be able to osc-placement one-shot yourself a sharing provider during setup | |
| 19:19:00 | mriedem | nope, and that's what i suggested in one of the placement update ML threads | |
| 19:19:05 | dansmith | okay | |
| 19:19:17 | dansmith | not sure why we merged that patch without that testing then | |
| 19:19:34 | dansmith | ^ that was rhetorical :D | |
| 19:20:51 | dansmith | mriedem: you're still trying to remember what the missing thing is right? | |
| 19:21:21 | mriedem | well, this is one https://github.com/openstack/nova/blob/master/nova/conductor/tasks/migrate.py#L57 | |
| 19:21:48 | dansmith | ooh right | |
| 19:22:03 | mriedem | and _revert_allocation in compute | |
| 19:22:33 | mriedem | resize to same host probably doesn't handle it either, not sure | |
| 19:22:34 | dansmith | yeah, I remember now.. I was thinking those were only problems we had with the doubling approach, but I remember now | |
| 19:22:39 | mriedem | anywho, | |
| 19:22:52 | mriedem | CI testing with ceph + modeling shared is a todo for the ptg etherpad most likely | |
| 19:22:56 | mriedem | i'll add it | |
| 19:23:10 | dansmith | yeah, that probably needs to be a priority for someone, | |
| 19:23:30 | dansmith | because right now if you associate a sharing provider, things are going to break in unhelpful ways | |
| 19:23:33 | mriedem | i'll assign it to my cat | |
| 19:23:55 | dansmith | melwitt: ^ | |
| 19:24:08 | mriedem | yeah - that's why i've tried to temper enthusiasm / communication that "now it's fixed" | |
| 19:24:17 | mriedem | also because i'm a debby downer | |
| 19:24:58 | dansmith | fair enough, I just saw it had merged when I was triaging a bug related to that stuff | |
| 19:25:05 | dansmith | and figured it had actually been finished | |