| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-10-27 | |||
| 16:47:30 | bauwser | jianghuaw: sorry, I was super busy prepping the new rev | |
| 16:47:40 | bauwser | jianghuaw: so, basically, we have the same problem with libvirt | |
| 16:47:52 | bauwser | jianghuaw: for example, we have a list of supported types | |
| 16:48:04 | bauwser | jianghuaw: then, the operator selects a specific type | |
| 16:48:16 | bauwser | (if you see my change, we only support one) | |
| 16:48:33 | bauwser | jianghuaw: so, the inventory is saying N+M | |
| 16:48:50 | bauwser | where N is the allocatable mdevs, and M is the already allocated ones | |
| 16:49:38 | bauwser | if the operator modifies that, then it's creating a problem because when you create a mdev by a type, the other types for the same device are getting 0 allocable mdevs | |
| 16:51:13 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: WIP: libvirt: Provide VGPU inventory for a single GPU type https://review.openstack.org/514763 | |
| 16:51:23 | bauwser | jianghuaw: new revision ^ | |
| 16:51:38 | jianghuaw | yes. I see. It can't create mdev for the other types once created a specific type of mdev. | |
| 16:51:51 | jianghuaw | looking at it | |
| 16:53:18 | jianghuaw | bauwser, no. my problem is different: https://github.com/openstack/nova/blob/master/nova/compute/resource_tracker.py#L807 | |
| 16:53:38 | jianghuaw | it will return if the compute_node resource is not changed. | |
| 16:54:51 | jianghuaw | bauwser, at the moment it only compare some resource fields of compute-node's: e.g. vcpu, memory. | |
| 16:54:52 | bauwser | jianghuaw: so, what's your concern ? | |
| 16:55:44 | jianghuaw | the vGPU resource is not saved in compute-node. so my concern is the vGPU inventory may be not be updated if the other resource is not changed. | |
| 16:56:01 | bauwser | a-ja | |
| 16:56:47 | jianghuaw | bauwser, as vGPU is only reported by get_inventory() | |
| 16:57:19 | bauwser | jianghuaw: right, good point | |
| 16:57:39 | bauwser | jianghuaw: so we should somehow signal those the old way | |
| 16:58:11 | bauwser | not saying to add the VGPU resources in the Computenode object :) | |
| 16:58:29 | bauwser | just having a way to signal it thru update_available_resources | |
| 16:58:31 | openstackgerrit | Eric Fried proposed openstack/nova master: Use ksa adapter for placement conf & requests https://review.openstack.org/492247 | |
| 16:58:43 | bauwser | or other way, I dunno | |
| 16:58:48 | bauwser | I need to think about it | |
| 16:59:17 | jianghuaw | I see. what I was thinking is to invoked always get_inventory(). | |
| 16:59:56 | jianghuaw | but I'm not sure if that's acceptable. | |
| 17:01:58 | jianghuaw | bauwser, here it defines update_inventory https://github.com/openstack/nova/blob/master/nova/compute/provider_tree.py#L233 | |
| 17:02:18 | bauwser | jianghuaw: one way to see that is that we could call get_inventory() first, pass it to _resource_change, keep it in memory and compare it at the next _resource_change run | |
| 17:02:39 | jianghuaw | but it requires some way to determine how and when to get the new inventory. | |
| 17:02:47 | bauwser | because _resource_change is just for making sure we don't update the DB every 60 secs | |
| 17:03:24 | openstackgerrit | Merged openstack/nova master: rp: move RP._set_traits() to module scope https://review.openstack.org/509028 | |
| 17:03:30 | bauwser | jianghuaw: for Queens, I really would like my series to not be depending on n-r-p | |
| 17:04:40 | jianghuaw | bauwser, the n-r-p seems in good progress. isn't it? | |
| 17:04:57 | jianghuaw | maybe we can use n-r-p? | |
| 17:07:31 | jianghuaw | bauwser, anyway I also agree that passing it to _resource_change looks like a good solution. | |
| 17:07:59 | bauwser | jianghuaw: I can write that change | |
| 17:08:09 | jianghuaw | I'd like to see the viewpoint from leakypipes also. | |
| 17:08:30 | bauwser | I need to dinner, but I'd be around by the next 2 hours | |
| 17:08:31 | jianghuaw | bauwser, yes you can go ahead with the change. | |
| 17:08:49 | bauwser | jianghuaw: which TZ are you in ? | |
| 17:09:07 | jianghuaw | I'm in China. | |
| 17:09:12 | jianghuaw | so 1:09 am:-) | |
| 17:09:59 | jianghuaw | please go ahead with your dinner. We can chat further on next Monday. | |
| 17:10:07 | jianghuaw | :-) | |
| 17:13:21 | bauwser | ++ | |
| 17:13:28 | bauwser | jianghuaw: and go to bed | |
| 17:13:30 | bauwser | ! | |
| 17:14:37 | jianghuaw | bauwser, cheers. wish you a nice weekend:-) | |
| 17:16:58 | openstackgerrit | Eric Fried proposed openstack/nova master: Parse granular resources/traits from extra_specs https://review.openstack.org/515151 | |
| 17:16:59 | openstackgerrit | Eric Fried proposed openstack/nova master: Granularize resources_from_{flavor|request_spec} https://review.openstack.org/515223 | |
| 17:18:29 | jianghuaw | bauwser, when you come back, hope you have seconds to look at this patch for get_inventory in xenapi: https://review.openstack.org/#/c/339204 | |
| 17:19:31 | jianghuaw | bauwser, also I have a patching to report vGPU resources from XenAPI: https://review.openstack.org/#/c/512965/ | |
| 17:20:21 | jianghuaw | I'm preparing another patch to create vGPU inventory. | |
| 17:24:42 | openstackgerrit | Ed Leafe proposed openstack/nova master: Only filter/weigh hosts once if scheduling a single instance https://review.openstack.org/513931 | |
| 17:24:42 | openstackgerrit | Ed Leafe proposed openstack/nova master: Add Selection objects https://review.openstack.org/499239 | |
| 17:24:43 | openstackgerrit | Ed Leafe proposed openstack/nova master: Return Selection objects from the scheduler driver https://review.openstack.org/495854 | |
| 17:24:43 | openstackgerrit | Ed Leafe proposed openstack/nova master: Change RPC for select_destinations() https://review.openstack.org/510159 | |
| 17:24:44 | openstackgerrit | Ed Leafe proposed openstack/nova master: Move the claim_resources method to scheduler utils https://review.openstack.org/511357 | |
| 17:24:44 | openstackgerrit | Ed Leafe proposed openstack/nova master: Make conductor pass and use host_lists https://review.openstack.org/511358 | |
| 17:31:39 | jianghuaw | leakypipes, thanks for the +2. | |
| 17:32:30 | leakypipes | jianghuaw: np | |
| 17:32:38 | leakypipes | jianghuaw: now go to bed! :) | |
| 17:33:04 | jianghuaw | :) | |
| 17:45:31 | openstackgerrit | Merged openstack/nova master: Include /resource_providers/uuid/allocations link https://review.openstack.org/499826 | |
| 17:50:43 | openstackgerrit | Eric Fried proposed openstack/nova master: Granularize resources_from_{flavor|request_spec} https://review.openstack.org/515223 | |
| 18:35:30 | mriedem | figleaf: superdan: leakypipes: cdent: how would you guys all love to have a thrilling hangout to make sure we're on the same page about https://review.openstack.org/#/c/510159/ ? | |
| 18:35:36 | mriedem | assuming cdent not so much given friday night fun time | |
| 18:35:54 | superdan | you mean now? | |
| 18:36:05 | cdent | i’m around actually, but haven’t been following along today | |
| 18:36:20 | mriedem | superdan: i'm game, i don't think this would be a long call | |
| 18:36:22 | mriedem | else it waits | |
| 18:37:40 | mriedem | i think in at least one retrospective we said, 'when in doubt, hangout' | |
| 18:37:55 | mriedem | and, 'if it's brown, flush it down' | |
| 18:38:12 | figleaf | mriedem: I can do it in the next 15 min; then I have to run out for a bit | |
| 18:39:23 | mriedem | figleaf: as in you're free at 2pm? | |
| 18:39:38 | mriedem | or out at 2pm? | |
| 18:39:42 | figleaf | mriedem: no, I'm free now. Gotta be somewhere at 2 | |
| 18:40:44 | mriedem | https://hangouts.google.com/call/M3gJOjmdi0q0Ayha5KkYAAEE | |
| 18:41:49 | superdan | cdent: you're breathing.. heavily | |
| 18:48:50 | leakypipes | mriedem: cool with me. | |
| 18:55:53 | figleaf | superdan: you have that effect on people | |
| 18:56:11 | superdan | I'm going to pretend you didn't say that | |
| 18:57:05 | mriedem | figleaf: ok left a summary in the patch | |
| 18:57:08 | mriedem | https://review.openstack.org/#/c/510159/ | |
| 18:57:08 | mriedem | thanks | |
| 19:34:54 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: add nested resource providers https://review.openstack.org/377138 | |
| 19:34:55 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: allow filter providers in tree https://review.openstack.org/377215 | |
| 19:34:55 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: adds REST API for nested providers https://review.openstack.org/384807 | |
| 19:34:56 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: update client to set parent provider https://review.openstack.org/385693 | |
| 19:37:05 | openstackgerrit | Eric Fried proposed openstack/nova master: ResourceRequest.to_querystring() https://review.openstack.org/515811 | |
| 19:37:58 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: adds REST API for nested providers https://review.openstack.org/384807 | |
| 19:37:58 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: update client to set parent provider https://review.openstack.org/385693 | |
| 19:45:16 | fried_rice | leakypipes Would you mind skimming the next couple patches in the series? I can dig having a RequestGroup class, but I'd like to keep the encompassing thing as a class (rather than a plain list of RequestGroup) so it do things like merge all the resources and produce a querystring. | |
| 19:45:36 | fried_rice | s/so it do/so it can do/ | |
| 19:47:37 | leakypipes | fried_rice: yup, I definitely will go through the remainder of the series. Probably won't be today, though... | |
| 19:48:10 | leakypipes | fried_rice: you're high in my queue, though. | |
| 19:48:37 | fried_rice | leakypipes Cool man. | |
| 20:04:26 | bauwser | leakypipes: around ? | |