| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-07-26 | |||
| 07:51:38 | openstackgerrit | Alex Xu proposed openstack/nova master: Add DeviceManager to the libvirt virt driver https://review.opendev.org/671388 | |
| 07:51:38 | openstackgerrit | Alex Xu proposed openstack/nova master: Populates the existing mediated devices in the libvirt device manager https://review.opendev.org/670787 | |
| 07:51:39 | openstackgerrit | Alex Xu proposed openstack/nova master: Using the claim/unclaim_for_instance for mdevs https://review.opendev.org/671222 | |
| 07:51:39 | openstackgerrit | Alex Xu proposed openstack/nova master: Adds functional test for creating the instance with vgpus https://review.opendev.org/671398 | |
| 10:11:55 | aspiers | sean-k-mooney, efried, bbobrov: exciting progress! http://paste.openstack.org/show/754889/ | |
| 10:13:21 | aspiers | cdent: inventory stuff seems to be working nicely :) ^^^ | |
| 10:14:01 | cdent | excellent, glad to hear it | |
| 10:42:27 | sean-k-mooney | aspiers: yep that does look like its working well | |
| 10:42:59 | sean-k-mooney | aspiers: you said you finally booted an sev instance with openstack yesterday right or is that the goal for today | |
| 10:43:35 | sean-k-mooney | aspiers: cdent actully should th max unit on that inventory not be 1 | |
| 10:44:02 | sean-k-mooney | you can only have one mem_encryption_context per vm right | |
| 10:45:19 | openstackgerrit | sean mooney proposed openstack/os-vif master: only disable mac ageing for ovs hybrid plug https://review.opendev.org/672834 | |
| 10:45:33 | sean-k-mooney | this a relitivly minor nit | |
| 10:55:12 | aspiers | sean-k-mooney: booting is the goal for today | |
| 10:56:34 | aspiers | sean-k-mooney: I don't get why you think max_unit should be 1? there are 16 slots configured on that machine | |
| 10:56:57 | sean-k-mooney | yes but max_unit is the maxium amount you can request in a singe request | |
| 10:56:57 | aspiers | ohhh, are you saying max_unit represents the maximum per VM, not the maximum per host? | |
| 10:57:03 | sean-k-mooney | yes | |
| 10:57:04 | aspiers | ok I didn't know that | |
| 10:57:10 | aspiers | in that case yeah it should be 1 | |
| 10:57:12 | aspiers | good catch | |
| 10:57:14 | sean-k-mooney | total is the total capasity | |
| 10:57:21 | aspiers | right that never quite made sense before | |
| 10:57:23 | aspiers | easy to fix | |
| 10:57:39 | sean-k-mooney | ya as i said minor nit | |
| 10:57:46 | aspiers | minor but v important! | |
| 10:57:50 | sean-k-mooney | we will only ever ask for 1 anyway | |
| 10:57:56 | aspiers | true | |
| 10:57:59 | aspiers | well | |
| 10:58:14 | aspiers | yeah | |
| 10:58:32 | aspiers | unless some operator requested the resource directly via the flavor | |
| 10:58:37 | sean-k-mooney | still it does not hurt to be defensive | |
| 10:58:40 | sean-k-mooney | ya | |
| 10:59:48 | aspiers | sean-k-mooney: is min_unit=1 correct though? | |
| 11:00:11 | sean-k-mooney | can you ask for less then 1 allocation of something if you want a resouce | |
| 11:00:14 | sean-k-mooney | yes it is | |
| 11:00:45 | sean-k-mooney | the min unit only applie if you ask for a resouce of that type | |
| 11:01:22 | sean-k-mooney | and since we dont want people saying resouce:MEM_ENCRYPTION_CONTEXT=0 it is correct | |
| 11:01:28 | aspiers | got it | |
| 11:01:30 | aspiers | thanks | |
| 11:02:35 | sean-k-mooney | resouce:MEM_ENCRYPTION_CONTEXT=0 would actully for the vm to land on a host that supported sev but not consume sev resouces so it could be used incorrectly by opertors to stave those nodes or ram cpus and disk | |
| 11:03:02 | sean-k-mooney | /for/force | |
| 11:03:51 | aspiers | good point | |
| 11:28:25 | openstackgerrit | Luyao Zhong proposed openstack/nova master: libvirt: Enable driver configures PMEM namespace https://review.opendev.org/634549 | |
| 11:28:25 | openstackgerrit | Luyao Zhong proposed openstack/nova master: Report pmem namespaces resources by provider tree https://review.opendev.org/634551 | |
| 11:28:26 | openstackgerrit | Luyao Zhong proposed openstack/nova master: Indroduce new LibvirtConfigGuestPMEM https://review.opendev.org/672956 | |
| 11:28:26 | openstackgerrit | Luyao Zhong proposed openstack/nova master: Populates existing pmems in libvirt device manager https://review.opendev.org/672957 | |
| 11:28:27 | openstackgerrit | Luyao Zhong proposed openstack/nova master: Support instance creation and resize with pmems https://review.opendev.org/672958 | |
| 11:31:38 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: API microversion 2.75: Add 'power-update' external event https://review.opendev.org/645611 | |
| 11:39:58 | efried | mnaser: re openstacksdk: | |
| 11:39:58 | efried | spec: http://specs.openstack.org/openstack/nova-specs/specs/train/approved/openstacksdk-in-nova.html | |
| 11:39:58 | efried | code: https://review.opendev.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/openstacksdk-in-nova | |
| 11:40:58 | efried | mnaser: note that we're not using clouds.yaml. We're using newly introduced functionality to consume ksa.loading oslo.config opts to build the Connection obj. | |
| 11:42:27 | efried | aspiers, sean-k-mooney: You can't ask placement for an allocation of 0. | |
| 11:43:13 | sean-k-mooney | efried: because we explcitly check or because we set min_unit to 1 | |
| 11:43:15 | efried | If you want to land on a SEV host but not consume SEV resources (why??) you would have to require the SEV trait. Which of course wouldn't be generic for SEV/MKTME like the rc is. | |
| 11:43:30 | sean-k-mooney | efried: we dont | |
| 11:43:43 | efried | sean-k-mooney: I'm actually not sure. Because I'm pretty sure you can't set min_unit to 0 either. | |
| 11:43:59 | sean-k-mooney | efried: aspiers was just asking if min_unit=1 was correct | |
| 11:44:04 | sean-k-mooney | and i was saying yes it is | |
| 11:44:08 | efried | yeah, it certainly is. | |
| 11:44:17 | efried | and per the design, max_unit=1 is also correct. | |
| 11:44:21 | efried | as you said | |
| 11:48:06 | efried | alex_xu: yt? | |
| 11:50:04 | efried | sean-k-mooney, aspiers: Who among cores do y'all think would be appropriate reviewers for https://review.opendev.org/670189 (harden get_domain_capabilities)? | |
| 11:52:32 | sean-k-mooney | good question. am most of the redhat folk are familar with libvirt, matt knows all but is also always over subsribed, maybe alex_xu on the intel side | |
| 11:53:16 | sean-k-mooney | to be honest we added a tone of comments so most core should be able to review at this point | |
| 11:54:36 | sean-k-mooney | short list would proably be dansmith bauzas alex_xu or mriedem | |
| 12:00:19 | sean-k-mooney | stephenfin is on pto today and maybe monday hence why he is not on the list | |
| 12:04:19 | sean-k-mooney | johnthetubaguy: by the way regarding https://review.opendev.org/#/c/669190/ | |
| 12:05:29 | sean-k-mooney | one i assume you still want to merge ^ but second i assume you are not against either reverting the feature entrily or making it work properly | |
| 12:24:39 | openstackgerrit | Merged openstack/os-vif master: Blacklist sphinx 2.1.0 (autodoc bug) https://review.opendev.org/672856 | |
| 12:57:37 | aspiers | efried: re. core reviewers, I can't add anything to sean-k-mooney's answer | |
| 12:57:52 | aspiers | I just found an issue with some of the SEV tests | |
| 12:58:11 | aspiers | turns out you can't use decorators on classes after all :-( or at least in some circumstances. digging more now | |
| 13:06:25 | mriedem | if anyone has a stein env handy could you see if you can recreate this? https://bugs.launchpad.net/nova/+bug/1837995 | |
| 13:06:25 | openstack | Launchpad bug 1837995 in OpenStack Compute (nova) ""Unexpected API Error" when use "openstack usage show" command" [Undecided,New] | |
| 13:06:33 | mriedem | i don't see why it would happen but if it is it should be fairly obvious | |
| 13:53:38 | sean-k-mooney | mriedem: i can unstack and restack an env but i dont have one to hand | |
| 14:09:45 | aspiers | stupid question #734: what's a "remote security group" as returned by "openstack security group rule list" ? | |
| 14:10:29 | openstackgerrit | Eric Fried proposed openstack/nova master: DNM: Where have all the global_request_id gone? https://review.opendev.org/672986 | |
| 14:10:43 | aspiers | and what's the effect of rules where the IP protocol / range are both None? | |
| 14:11:02 | aspiers | I would assume it just allows everything, but I'm not seeing that behaviour | |
| 14:13:11 | efried | aspiers: This sounds like a question for #openstack-keystone | |
| 14:13:32 | aspiers | oh really? I thought secgroups were part of nova | |
| 14:13:45 | aspiers | I guess my knowledge in this area is 5 years out of date or something | |
| 14:15:32 | mriedem | you're both wrong | |
| 14:15:34 | mriedem | neutron | |
| 14:15:39 | aspiers | haha | |
| 14:15:45 | aspiers | that would have been my second guess | |
| 14:16:01 | aspiers | actually I don't know why it wasn't my first | |
| 14:16:16 | aspiers | since I know full well that neutron manages the network namespaces and the rules within them | |
| 14:18:12 | slaweq | aspiers: if both IP protocol and range are None than it's not inserted into iptables rule to match, so all will be matched | |
| 14:18:42 | aspiers | slaweq: isn't the default is to reject? | |
| 14:19:05 | aspiers | if there's no corresponding iptables rule, I don't see what could match | |
| 14:19:08 | slaweq | aspiers: with rules You are defining what should be accepted | |
| 14:19:15 | aspiers | exactly that's what I mean | |
| 14:19:21 | slaweq | everything else, what will not match rules will be rejected | |
| 14:19:37 | aspiers | right but before you said "all will be matched" | |
| 14:19:44 | slaweq | so e.g. if You have rule with only "--direction ingress" than everything with ingress direction will be accepted | |