Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-03-30
15:20:52 mriedem sdague: maybe? i was going to try and spend some time reading your limits spec
15:21:04 ralonsoh mriedem: this is still in progress. Now I can't tell you (I don't know) which is the current status
15:21:12 sdague mriedem: yeh, the limits one is more important
15:21:28 vladikr jaypipes, bauzas, hey could you please take a look at the xen vgpu spec, when you have a moment. https://review.openstack.org/#/c/450122 - I'm trying to convince the author to make the spec more generic, so we could implement the libvirt side of the story under the same spec
15:21:37 ralonsoh mriedem, bauzas: of course, intel 3rd party CI is the right place to put those tests
15:21:42 bauzas vladikr: sure thing
15:21:55 vladikr bauzas, thanks
15:22:12 bauzas ralonsoh: mriedem: I didn't want to block that spec, but I do wonder why we couldn't have functional in-tree tests for that
15:22:13 jaypipes vladikr: yup, it's in my queue. btw, jianghuaw__ is the author :)
15:22:37 vladikr jaypipes, thanks :)
15:22:38 mriedem bauzas: if we can, then great
15:22:49 bauzas ralonsoh: mriedem: I know it's a huge implementation, but we could use some fixtures for that
15:22:56 mriedem assuming the functional tests don't incorrectly mock out things
15:23:23 bauzas mriedem: for example, some NFVLibvirt fixture providing the capabilities
15:23:28 ralonsoh bauzas, mriedem: that could be the easiest and fastest option
15:24:04 bauzas plus a NFVNeutronPCIportFixture :p
15:24:42 ralonsoh bauzas, mriedem: I need to take a look at SRIOVServersTest
15:43:02 openstackgerrit Jiri Suchomel proposed openstack/nova master: Retry few times when creating the initial connection to libvirtd. https://review.openstack.org/416928
15:50:05 openstackgerrit Roman Podoliaka proposed openstack/osc-placement master: WIP: implement API version 1.0 https://review.openstack.org/451877
16:01:23 kaisers__ mdbooth: ping?
16:08:46 esberglu I am upgrading my multinode environment from newton to ocata. Currently I have to deploy the control, then the compute, then go back to the control and run
16:08:48 esberglu "nova-manage cell_v2 simple_cell_setup"
16:08:52 esberglu Otherwise I get an error about no host to cell mapping and can't deploy any instances
16:08:56 esberglu Not sure if there is some sort of devstack configuration that I am missing or if this the correct way to go about doing it
16:10:09 dansmith esberglu: you should be able to run simple_cell_setup prior to doing anything with the computes
16:10:42 dansmith esberglu: this bit of upgrading devstack is actually in grenade though.. are you talking about a real setup, or literally upgrading your devstack?
16:11:48 esberglu Not sure I understand your question
16:19:55 mriedem sdague: done https://review.openstack.org/#/c/440815
16:20:12 mriedem sdague: i think we'll need to talk about the rollout / upgrade plan at the forum
16:20:17 mriedem but i'm on board with the general direction
16:20:21 mriedem good story telling old man
16:20:34 mriedem there are some wording cleanups that could happen in a follow up for clarification, but otherwise lgtm
16:40:14 openstackgerrit Samantha Blanco proposed openstack/nova master: Make flavor-rxtx policy more granular https://review.openstack.org/444036
16:49:58 openstackgerrit Dan Smith proposed openstack/nova master: Teach HostAPI about cells https://review.openstack.org/442162
16:54:32 alextricity25 has anyone seen this before:
16:54:43 alextricity25 /openstack/log/605056-infra02_nova_api_os_compute_container-a15d94c3/nova/nova-api-os-compute.log:2017-03-29 19:01:07.237 10449 DEBUG nova.api.openstack.wsgi [req-4219ada0-490c-46c5-870f-27823c4d6e2c e1eeaf7a6cec4756aa2880efda64448f fdf3fe3df3f5428a8b30daa70736bc6a - default default] Returning 400 to user: Block Device Mapping is Invalid: failed to get volume c82eb821-ec3a-4ce6-8abc-12466d1c7fe0. __call__ /opens
17:08:05 bauzas dansmith: heh, fun tweet
17:08:18 bauzas dansmith: but honestly, the scheduler is good :)
17:08:39 bauzas dansmith: the real problem is not about the scheduler, rather about how the resources are asked :)
17:08:45 dansmith I seriously want to burn it all right now
17:08:49 dansmith and I mean ALL of it
17:08:59 bauzas why ? because of my comments ? :p
17:09:36 dansmith no, because it's terrible
17:10:03 bauzas well, it took me 1 year to understand all the scheduler modules :p
17:10:16 bauzas and then, I discovered the RT :p
17:11:03 bauzas tbh, I'd prefer to kill the RT *before* killing the scheduler :p
17:11:11 bauzas I mean, filters are nice
17:11:26 bauzas of course, we have problems with the HostManager :)
17:23:50 openstackgerrit Lee Yarwood proposed openstack/nova master: compute: Move detach logic from manager into driver BDM https://review.openstack.org/439520
17:29:55 sajeesh ping jaypipes
17:30:27 jaypipes sajeesh: pong.
17:31:27 sajeesh Hi jaypipes..is it finalised that quota limits will be stored in keystone only and not in placement
17:42:43 melwitt sajeesh: there is a spec proposed to keystone for quota limits https://review.openstack.org/#/c/440815/
17:45:01 jaypipes melwitt: sajeesh has commented already on that.
17:45:09 jaypipes I have it on my long list of todos...
17:45:21 melwitt oh, okay
17:45:47 jaypipes but basically, sajeesh, the idea is to have the service projects (i.e. NOT keystone) do the usage checks and to have Keystone only store the absolute limits.
17:46:06 jaypipes guh.
17:47:42 sdague jaypipes: so I just reved that spec, it would be good to make sure it still jives with your thinking, as we were going to do basically go / no go tomorrow on moving forward on it - https://review.openstack.org/#/c/440815/
17:47:50 melwitt oh, I see, he was thinking usage and limits had to be in the same service maybe
17:47:56 sdague melwitt: nope
17:48:06 sajeesh sorry...jaypipes...I couldn't chat as I am having some problem with internet connection
17:48:13 sdague https://review.openstack.org/#/c/440815/5/specs/keystone/backlog/unified-limits.rst@117
17:48:18 sdague melwitt: ^^^
17:48:27 sdague hopefully that bit makes it clear
17:48:36 melwitt yeah, I know. I was saying I think maybe sajeesh was thinking they had to be both in the same place
17:48:54 sajeesh yes....
17:49:00 sdague melwitt: ah, sorry, I misread your sentence
17:49:07 melwitt (based on his comment on the spec)
17:49:13 sdague and missed the pronoun switch from I -> he
17:50:43 melwitt no worries
17:51:30 dansmith sdague: omg.. pronouns are SO important these days and you're just ignoring them? tres bad!
17:54:56 openstackgerrit Steve Noyes proposed openstack/nova master: Do not attempt volume swap when guest is stopped/suspended https://review.openstack.org/389798
17:59:00 jaypipes sajeesh: did you see my last comment? it was this:
17:59:02 jaypipes but basically, sajeesh, the idea is to have the service projects (i.e. NOT keystone) do the usage checks and to have Keystone only store the absolute limits.
17:59:21 jaypipes sdague: yes, will review shortly... dealing with some Mirantis $work things right this sec
18:00:23 jaypipes looks like we have a case of the net-sh*ts today.
18:00:28 sajeesh ok
18:34:35 openstackgerrit Anusha Unnam proposed openstack/nova master: Add description to policies in services.py https://review.openstack.org/449782
18:35:13 openstackgerrit Samantha Blanco proposed openstack/nova master: Make flavor-rxtx policy more granular https://review.openstack.org/444036
18:35:53 openstackgerrit Gábor Antal proposed openstack/nova master: Move rebuild notification tests into separate method https://review.openstack.org/451789
18:57:03 openstackgerrit Sujitha proposed openstack/nova master: Add BDM to InstancePayload https://review.openstack.org/448779
20:03:05 efried mriedem Good time to catch up on code reviews, so your deltoids don't get burned up.
20:03:34 mriedem let me drink some eggs first
20:04:14 efried Purely selfless comment, of course. Just looking after our noble and eminent PTL.
20:06:17 jaypipes sdague: ok, phew, finally finished the limits spec review.
20:10:37 melwitt jaypipes: I've made similar comments on terminology on other reviews. like to keep it as "limits" and "usage" and make it clear about the difference. "quota" is overloaded and confusing
20:10:44 mriedem yuh dun blowed up all ma stats
20:13:59 xgerman hi,
20:14:27 xgerman wonder what the status of soft-anti-affinity is? I am on master nd it doesn’t work
20:14:37 xgerman https://www.irccloud.com/pastebin/zfgU1sJu/
20:17:03 sdague jaypipes: ok, most of the clarifications are good, I'll look at doing another quick edit to be consistent there. When it comes to why I think there needs to be a library, see the various quota rule sets here that were being explored - https://review.openstack.org/#/c/441203/
20:18:22 mriedem xgerman: i'd talk to gibi or ildikov
20:18:35 xgerman thanks
20:18:49 openstackgerrit Anusha Unnam proposed openstack/nova master: Add description to policies in servers.py https://review.openstack.org/451966
20:19:07 xgerman when I grepped my nova dir it was there so odd
20:19:42 ildikov xgerman: is that latest master?
20:19:52 xgerman pretty much
20:20:07 xgerman build my devstack 3-27
20:20:44 xgerman or is there something I need to enable - I am run ing with default settings

Earlier   Later