Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-02-12
15:28:15 stephenfin cdent: It's not even solving one of them yet. I probably have it configured incorrectly...again
15:28:33 bauzas mriedem: and the fact that if you have an image that is isolated, but then you create a volume from that image and then boot from that volume, Nova won't guarantee your host will be excluded
15:28:50 stephenfin bauzas: Don't be dissing the Kong now, ya hear ;)
15:29:15 mriedem bauzas: well, it will depending on the other config options
15:29:21 bauzas mriedem: exactly
15:29:47 bauzas mriedem: but consider the case where I have an image that is restricted to host1
15:29:50 mriedem i'd say, if you care, use AggregateImagePropertiesIsolation
15:30:13 mriedem actually,
15:30:20 mriedem AggregateImagePropertiesIsolation likely doesn't work for volume-backed instances properly either
15:30:30 bauzas mriedem: if I'm making a volume from that image and then booting from it, then the instance will be exclused from that host if strict_host_blah=True
15:31:09 mriedem for a volume-backed instance, the only image properties we store in the request spec are the ones from the volume_image_metadata https://github.com/openstack/nova/blob/master/nova/utils.py#L1033
15:31:35 mriedem and i don't know if that actually stores the properties from the backing image, and even if it does, they could be stale
15:32:16 mriedem so yeah, any filter that relies on image properties is likely semi-broken with boot from volume
15:32:44 openstackgerrit Sylvain Bauza proposed openstack/nova master: for doc clarity https://review.openstack.org/543494
15:33:44 openstackgerrit Chris Dent proposed openstack/nova master: Isolate config parse_args for placement https://review.openstack.org/543495
15:35:46 AJaeger mriedem, the novaclient reviews I mentioned earlier are: review https://review.openstack.org/543115 , https://review.openstack.org/#/c/543114, https://review.openstack.org/#/c/541370/ - the first two are needed to fix the jobs, the third is needed for Zuul v3.
15:37:59 openstackgerrit Sylvain Bauza proposed openstack/nova master: doc: BFV instances and IsolatedHostsFilter https://review.openstack.org/543494
15:38:16 bauzas mriedem: just made a first try for doc ^
15:40:41 AJaeger thanks, mriedem
15:41:26 bauzas stephenfin: sahid: do you have any idea why the NUMATopologyFilter needs to set some limits for CPU and RAM ? https://github.com/openstack/nova/blob/master/nova/scheduler/filters/numa_topology_filter.py#L92-L93
15:41:33 bauzas context is https://bugs.launchpad.net/nova/+bug/1748858
15:41:34 openstack Launchpad bug 1748858 in OpenStack Compute (nova) "allocation_ratio set by aggregate metadata don't work in NUMATopologyFilter" [Undecided,New] - Assigned to yangjie (yang.jie)
15:42:13 bauzas stephenfin: sahid: the NUMATopologyFilter should only care in my opinion to enforce the numa limits, not any other resource usage
15:42:34 stephenfin bauzas: Not off the top of my head, no. That whole 'NUMATopologyLimits' object is a weird one I've yet to grasp properly. Hopefully sahid knows more
15:49:19 openstackgerrit Matt Riedemann proposed openstack/nova master: Fix typo in NUMATopologyFilter docs https://review.openstack.org/543505
15:56:17 mriedem bauzas: fwiw, the image properties are stored on the volume's volume_image_metadata dict: http://paste.openstack.org/show/670139/
15:56:27 mriedem i add foo=bar to the image i used to create that bootable volume
15:56:52 mriedem so AggregateImagePropertiesIsolation would only be as good as how accurate that metadata is
15:56:58 bauzas mriedem: are you persuing me to look up the volume if BFV ?
15:57:05 mriedem since nova doesn't re-fetch the image to get the latest meta
15:57:17 mriedem persuing you how?
15:57:43 bauzas the fact is that filters don't lookup the volume image metadata, right?
15:57:49 bauzas they just assume every instance has an image
15:58:10 bauzas well, they rather assume every RequestSpec has an image
15:58:18 mriedem the filter assumes the request spec image has the details from when the api created the request spec
15:58:22 mriedem yes that
15:58:48 mriedem and the request spec for a volume-backed instance will have the image meta stored in the volume from the time the volume was created
15:59:10 bauzas mriedem: if we create a fake image based on that, right?
15:59:13 mriedem but my point is, i could create a volume with some image, then change properties on that image, and then create a server from the volume and the image meta in the volume would be stale
15:59:21 bauzas I agre
15:59:24 bauzas I got your point
15:59:52 bauzas if we want to support BFV instances correctly, we would necessarly have to lazy-load things on the fly if we need to access the "image"
16:00:18 bauzas hence the "are you persuing me to lookup the volume"
16:00:42 mriedem no i'm not suggesting that,
16:01:05 mriedem i was saying, if we cared, we could get the image_id from the volume meta, and lookup the image at server create time to get the current image meta from glance
16:01:10 mriedem rather than the possibly stale image meta from the volume
16:01:27 mriedem but, i'm not suggesting we do that either
16:01:41 mriedem i guess we should just assume that the image meta with the volume is like a snapshot
16:01:55 mriedem if the image meta is changed after the volume is created, meh
16:02:25 mriedem if it was changed in such a substantial way that it would impact nova scheduling for the instance, then it should probably be a different image
16:03:48 bauzas that looks like the best, indeed
16:07:26 bauzas mriedem: https://bugs.launchpad.net/nova/+bug/1748858
16:07:28 openstack Launchpad bug 1748858 in OpenStack Compute (nova) "allocation_ratio set by aggregate metadata don't work in NUMATopologyFilter" [Undecided,Won't fix] - Assigned to yangjie (yang.jie)
16:07:56 bauzas mriedem: we can clarify the doc, but I don't feel we should open a can of worms about aggregate ratios
16:10:49 openstackgerrit Eric Berglund proposed openstack/nova master: WIP: PowerVM Driver: Snapshot https://review.openstack.org/543023
16:12:15 mriedem bauzas: i don't either
16:13:19 openstackgerrit Merged openstack/nova-specs master: Fix the names on some specs https://review.openstack.org/542504
16:15:13 openstackgerrit Merged openstack/nova-specs master: Move queens implemented specs https://review.openstack.org/542505
16:18:00 openstackgerrit Eric Berglund proposed openstack/nova master: Use dict.get() when accessing capabilities dict https://review.openstack.org/539715
16:38:57 openstackgerrit Bence Romsics proposed openstack/osc-placement master: PUT dict format to /allocations (v1.11, v1.12) https://review.openstack.org/542819
16:40:13 openstackgerrit sahid proposed openstack/nova master: libvirt: disconnect volume from host during detach https://review.openstack.org/515008
16:46:28 openstackgerrit sahid proposed openstack/nova master: hardware: only take into account small pages https://review.openstack.org/532168
16:55:15 gibi mriedem: Would it be OK with you cancelling the next 3 notification subteam meeting? This week I don't feel we have things to talk about and next week I will already travelling to Dublin and then there will be PTG week.
16:56:07 mriedem gibi: works for me
16:56:46 gibi mriedem: OK, cool
17:04:37 mgariepy is there a way to set a flavor to set the kvm=off option or i really need to pass the key from glance image ?
17:05:16 mgariepy referring to : https://review.openstack.org/#/c/459753
17:06:04 openstackgerrit Matt Riedemann proposed openstack/nova master: Move the nova-next job in-tree and update it https://review.openstack.org/541474
17:06:04 openstackgerrit Matt Riedemann proposed openstack/nova master: Test websocketproxy with TLS in the nova-next job https://review.openstack.org/513160
17:06:48 mriedem mgariepy: looking at the code https://review.openstack.org/#/c/459753/18/nova/virt/libvirt/driver.py@4184 it only checks the image meta, not flavor extra specs
17:07:39 mgariepy thanks mriedem i'll add the key to an image would it be a lot of work to add that to flavor extra specs ?
17:08:35 mriedem probably not a lot of work, no, but generally i think we prefer image meta since those are standardized and flavor extra specs are not
17:09:33 mriedem i guess the flavor extra spec could maybe make sense in this case if you're using a GPU trait on the flavor or something
17:09:51 mriedem https://docs.openstack.org/nova/latest/admin/virtual-gpu.html#configure-a-flavor-controller
17:10:12 mriedem this is getting dangerously into bauzas territory though
17:12:16 mgariepy i define my flavor to have the gpus via pci_passthrough:alias would have liked to have the hidden kvm stuff on that flavor as well.
17:12:30 mgariepy but as long as it works i guess i'll use the workaround.
17:13:24 openstackgerrit Lee Yarwood proposed openstack/nova stable/pike: Do not attempt volume swap when guest is stopped/suspended https://review.openstack.org/533733
17:15:11 mriedem mgariepy: ok, i guess it's unclear to me if that's only an issue for windows guests as the commit message implies
17:15:33 mgariepy i need that for linux guest as well.
17:15:33 mriedem in that case it makes sense to me to be tied to the image
17:15:37 mriedem oh
17:15:56 mgariepy nvidia driver don't like kvm guest.
17:16:47 mriedem i'm just amazed the thing is documented https://docs.openstack.org/python-glanceclient/latest/cli/property-keys.html
17:16:52 jaypipes efried: where is that Numbered Request Group Affinity spec?
17:17:17 efried jaypipes: It's not a spec... yet. It's a subsection under Alternatives.
17:17:39 efried jaypipes: I'm considering writing a spec for it, though. It would be largely the same as the one you're looking at.
17:57:19 openstackgerrit Merged openstack/nova-specs master: Granular Resource Request Syntax (Rocky) https://review.openstack.org/540179
17:58:34 edleafe jaypipes: Did we ever answer the question about whether we need "distance" to be continuous or binary? IOW, do we need "at least 2 away" or "closer than 5", or do we just need "is/isn't in the same grouping"?
17:58:52 edleafe jaypipes: you mentioned some HA use cases that I'm not familiar with
18:08:14 openstackgerrit Merged openstack/nova master: Add regression test for BFV+IsolatedHostsFilter failure https://review.openstack.org/543257
18:08:31 openstackgerrit Merged openstack/nova master: Handle volume-backed instances in IsolatedHostsFilter https://review.openstack.org/543263
18:08:56 openstackgerrit Merged openstack/nova master: Fix docs for IsolatedHostsFilter https://review.openstack.org/543264
18:13:22 jaypipes edleafe: well, the HA use cases are more about anti-affinity... i.e. "make sure the three nodes of my cluster are running in separate failure domains (or groups, or racks, or whatever)"
18:14:57 edleafe jaypipes: thanks. Are there any cases that you know of that don't boil down to yes/no? Like, are X and Y in the same "thing" (rack, row, etc.) or aren't they
18:15:36 openstackgerrit Lee Yarwood proposed openstack/nova stable/pike: libvirt: Block swap_volume attempts with encrypted volumes https://review.openstack.org/543569
18:17:38 jaypipes edleafe: yeah. the whole "soft affinity" is what those use cases are about.
18:17:52 jaypipes edleafe: in "soft affinity", it's a preference vs. yes/no
18:18:22 jaypipes edleafe: i.e. "place these things *as near to each other* as possible" or "place these related things as far away as possible"
18:18:51 jaypipes edleafe: in the soft affinity cases, the distance being relative and not yes/no enables that sorting of potential destinations

Earlier   Later