| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-20 | |||
| 19:58:17 | sean-k-mooney | jaypipes: how did the dentist go. | |
| 19:58:20 | jaypipes | efried: suffice to say it involves me cleaning up a giant pile of dog shit in the back of my car in the rain. | |
| 19:58:29 | jaypipes | efried: ^ not related to the dentist | |
| 19:58:36 | mriedem | but at least your teeth are clean | |
| 19:58:37 | efried | jaypipes: Geez, you try to do something nice for the dog... | |
| 19:58:51 | jaypipes | indeed. | |
| 19:58:53 | mriedem | unless there was...splatter | |
| 19:59:21 | sean-k-mooney | mriedem: ew, let hop not for jaypipes sake | |
| 19:59:24 | efried | jaypipes: You want a quick feeling of satisfaction, https://review.openstack.org/#/c/545111/ ought to be eligible for quick-approve :) | |
| 19:59:29 | sean-k-mooney | *hope | |
| 19:59:55 | mriedem | efried: i'll take a look again at that after this | |
| 19:59:57 | jaypipes | my dogs have become a nested provider tree with inventory of DOG_SHIT_KG | |
| 20:00:07 | efried | rofl | |
| 20:00:08 | mriedem | CUSTOM_DOG_SHIT_KG? | |
| 20:00:11 | tssurya | lol | |
| 20:00:14 | jaypipes | yes, sorry. | |
| 20:00:19 | jaypipes | well corrected, mriedem | |
| 20:00:22 | mriedem | :) | |
| 20:00:31 | sean-k-mooney | mriedem: i dont know me might want to standarise it | |
| 20:01:03 | jaypipes | sean-k-mooney: os-dog-poop? | |
| 20:01:26 | jaypipes | though that's a big specific. os-poop would be more generic and future proof. | |
| 20:01:27 | efried | Well, we already have os-brick | |
| 20:01:35 | jaypipes | well played efried | |
| 20:01:39 | sean-k-mooney | jaypipes: you know im surprised you have not come up with an os-pug yet | |
| 20:03:16 | jaypipes | sean-k-mooney: Provider Usage Group. | |
| 20:03:40 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Don't log a warning for InstanceNotFound with deleted VIFs https://review.openstack.org/554591 | |
| 20:07:34 | dansmith | jaypipes: loaded question for you in here: https://review.openstack.org/#/c/544694/2/specs/rocky/approved/alloc-candidates-member-of.rst | |
| 20:07:40 | mriedem | efried: edmondsw: question inline about evacuate https://review.openstack.org/#/c/545111/ | |
| 20:10:03 | edmondsw | mriedem looking | |
| 20:10:36 | mriedem | also, | |
| 20:10:53 | mriedem | at some point, we should not have powervm specs for 'implement random feature parity stuff in our driver', | |
| 20:11:04 | mriedem | just do what other virt drivers do and have specless feature parity blueprints per feature in questoin | |
| 20:11:34 | sean-k-mooney | mriedem: for https://review.openstack.org/554591 are the network-vif-deleted event generated by neutron port being deleted as part of instance deleteion or are they neutron vif unplugged events form the ports being removed from ovs? | |
| 20:11:40 | efried | mriedem: That would be lovely. Are we at that point yet? | |
| 20:11:52 | mriedem | efried: i'm ok with that if melwitt is | |
| 20:12:17 | mriedem | sean-k-mooney: the former | |
| 20:12:29 | melwitt | mriedem, efried: sounds fine to me | |
| 20:12:52 | sean-k-mooney | mriedem: ok then ya that makes sense to me. i just was not sure where the event was comming from | |
| 20:12:52 | efried | Cool beans. edmondsw esberglu ^. | |
| 20:13:26 | efried | mriedem, melwitt: But for this release, since we're already here, use the bp/spec that's proposed, yah? | |
| 20:13:39 | mriedem | shrug, i personally don't care for the wishlist | |
| 20:13:53 | mriedem | because it just seems really random | |
| 20:13:55 | melwitt | is the proposed spec all just feature parity stuff? | |
| 20:14:05 | edmondsw | I'm happy to abandon this and start using specless blueprints this release | |
| 20:14:05 | efried | melwitt: yes | |
| 20:14:05 | mriedem | and i see adding support for hot plugging vifs was added after PS3 | |
| 20:14:13 | efried | ++ | |
| 20:14:31 | mriedem | there are no actual design details in the spec, so it's not really useful imo | |
| 20:14:34 | sean-k-mooney | mriedem: so just to follow on from that, does that mean if i delete a neutron port that is bound to an instance that automatically results in a notifcation to nova to call detach interface on the virt driver? | |
| 20:14:36 | edmondsw | yep | |
| 20:14:36 | mriedem | so specless bp is fine | |
| 20:14:47 | melwitt | sounds reasonable to me | |
| 20:14:48 | mriedem | sean-k-mooney: yup | |
| 20:15:12 | efried | Sweet. Swat that paperwork down! | |
| 20:15:51 | sean-k-mooney | mriedem: hum ok is that documented anywhere. it kind of makes sense i just would not have expected neutron to allow you to delete teh port in that case | |
| 20:16:04 | edmondsw | mriedem to the evacuate question... I'm honestly not sure if there's anything more to it than a) supporting moving to another host in general and b) updating the support matrix | |
| 20:16:21 | jaypipes | dansmith: will try to answer it first thing in the morning... been a day. having a beer. | |
| 20:16:34 | dansmith | jaypipes: oh you said you were getting to work, so I figured you were around | |
| 20:16:54 | dansmith | mah bad, tomorrow is fine | |
| 20:17:57 | melwitt | getting to work cleaning dog poodoo | |
| 20:18:07 | mriedem | edmondsw: someone on your dev or qa team could actually test it first | |
| 20:18:11 | mriedem | to flush out any obvious problems | |
| 20:18:21 | edmondsw | mriedem oh definitely | |
| 20:18:29 | edmondsw | was just talking from a code perspective | |
| 20:18:58 | mriedem | sean-k-mooney: good question; i didn't trace the requests in the tempest job to see if the port was actually detached before it was deleted, or what the order was there, | |
| 20:19:15 | mriedem | sean-k-mooney: should be pretty easy to find out - create a server from a pre-existing port and try to delete the port | |
| 20:19:24 | mriedem | i don't have a devstack handy | |
| 20:19:32 | mriedem | i know you can't delete an attached volume | |
| 20:19:48 | sean-k-mooney | mriedem: ill give it a try i think i have an env running | |
| 20:23:26 | openstackgerrit | Eric Berglund proposed openstack/nova master: DNM: EXPERIMENTAL: Set proc_units_factor to 0.1 in VMBuilder init https://review.openstack.org/554688 | |
| 20:23:49 | sean-k-mooney | mriedem: so ya i can create the port, boot with it then delete it and the interface gets detatched from the running vm | |
| 20:23:59 | sean-k-mooney | mriedem: im just going to check the xml to confirm | |
| 20:24:35 | sean-k-mooney | mriedem: yep the vm no has no nics | |
| 20:25:08 | sean-k-mooney | i wonder if the revers works. can i create a neutron port and then set its owner to a vm and trigger an attache | |
| 20:27:04 | cfriesen | So in the discussion around https://review.openstack.org/#/c/552924/ something interesting has come up...currently vcpus and 4KB pages can be consumed on a compute-node bases (where it floats across the whole compute node) or on a numa-node basis (where it's constrained to a single numa node) depending on whether the instance has a numa topology or not. (https://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver. | |
| 20:27:06 | cfriesen | py#L4435) This makes it hard to reason about resource tracking. | |
| 20:30:54 | mriedem | sean-k-mooney: no the reverse doesn't work, | |
| 20:31:05 | mriedem | not might get a network-vif-plugged event but it won't know what to do with it and just log a warning | |
| 20:31:10 | mriedem | *nova | |
| 20:31:33 | sean-k-mooney | mriedem: ya just tried it. | |
| 20:31:40 | mriedem | the fact you can delete a port while it's attached is pretty scary | |
| 20:31:43 | sean-k-mooney | i create another port and did openstack port set --device 55a08a3b-fd27-495f-85da-39eedd5fde9c --device-owner "compute:nova" --host "ubuntu" myport | |
| 20:32:01 | sean-k-mooney | it ended up binding the port on the host correctly | |
| 20:32:18 | sean-k-mooney | im going to do a hard reboot but i think it will then update the xml correctly | |
| 20:32:31 | mriedem | in the compute api we check instance states when detaching a port, | |
| 20:32:34 | edleafe | dansmith: sorry, was away. Yes, the idea for member_of was "member of at least one of these". The typical use case would be only 1 agg. | |
| 20:32:39 | mriedem | but the networking api wouldn't have that | |
| 20:32:47 | mriedem | so you could detach ports by deleting them while an instance is migrating | |
| 20:33:14 | mriedem | which seems, not good | |
| 20:34:02 | cfriesen | does it mess up our accounting? | |
| 20:34:59 | cfriesen | why is "while it's migrating" different than "while it's running"? | |
| 20:35:34 | sean-k-mooney | mriedem: strange the hard reboot did not cause the interface to show up either. i must be missing something that the attach/detach whould have done | |
| 20:36:24 | dansmith | edleafe: yeah, so that's kinda what I was thinking at first, but then I tried to apply it to two things at once, so: | |
| 20:36:42 | dansmith | edleafe: imagine I have a tenant-restricted aggregate and the user boots with a requested AZ | |
| 20:37:08 | dansmith | edleafe: if I just concatenate the tenant and az list, then I won't restrict them properly | |
| 20:37:18 | efried | cfriesen: Are you suggesting that somewhere in the stack there will be code responsible for translating the allocation_request for [PCPU:2 from NUMA_RP_1] to [PCPU:2 from NUMA_RP_1 + VCPU:2 from compute node RP] ?? | |
| 20:37:37 | mriedem | cfriesen: it was an example of a bad time to detach a port | |
| 20:37:49 | cfriesen | efried: I'm suggesting that our logic right now is not rigorous. :) | |
| 20:37:53 | dansmith | edleafe: I can restrict the tenant aggs to only the one that matches the az, but they might not be the same size (and thus are different) | |
| 20:38:00 | sean-k-mooney | right we could have already caluated the new xml and miss the port is being removed | |