Posted	Nick	Remark
#openstack-nova - 2021-04-08
08:25:04	nautik	just wanted to make sure I got this right: you are saying that if the host has been shut down, disabling nova is useless as long as it seen as "Down", is that correct?
08:27:57	gibi	nautik: yes, if the controller already detected that the host down, then it will not forward requests to it.
08:32:40	nautik	gibi: and servers will not be scheduled by placement either I guess?
08:33:15	nautik	(scheduler/placement)
08:34:36	gibi	nautik: yes, it is the expected behavior that no server is scheduled to down compute hosts
08:36:06	nautik	ok, so I will discuss on Masakari side then; because masakari starts by trying to disable the compute service, which I guess was to try to cover the time when the control plane has not yet seen the host as down
08:37:00	gibi	do you know why masakari trying to disable the host?
08:37:46	lucasagomes	hi folks, now that we have wallaby branched, can someone take a look at https://review.opendev.org/c/openstack/nova/+/776419, https://review.opendev.org/c/openstack/nova/+/776934 and https://review.opendev.org/c/openstack/nova/+/776944
08:38:01	gibi	nautik: I have to drop offline for 30 mins but I will be back after
08:38:07	lucasagomes	this patches will make sure the nova gate is not broken when we flip the default network backend in devstack to OVN
08:38:13	lucasagomes	these patches*
08:38:20	lucasagomes	sean-k-mooney, ^ if u have some time
08:38:25	nautik	gibi: nope I don't know ^^
08:43:16	openstackgerrit	Lee Yarwood proposed openstack/nova stable/ussuri: libvirt: Skip encryption metadata lookups if secret already exists on host https://review.opendev.org/c/openstack/nova/+/765770
08:43:20	nautik	gibi: but masakari is not the one directly powering off the host, it relies on pacemaker/corosync to do it for him and has some "wait for stonith" sleep configurations
08:44:24	nautik	then it waits for the host to be seen as down in nova to handle servers evacuation
08:45:12	lyarwood	wait is this instanceha?
08:45:25	lyarwood	(a downstream OSP feature)
08:46:52	lyarwood	https://bugs.launchpad.net/nova/+bug/1920977/comments/1 from bauzas is correct, you should be using the force down API, AFAIK that's what we do in the instanceHA feature downstream within OSP
08:46:54	openstack	Launchpad bug 1920977 in OpenStack Compute (nova) "Error 504 when disabling a nova-compute service recently down" [Undecided,Invalid]
08:47:54	nautik	lyarwood: would you have a link about the instanceHA OSP feature ?
08:48:34	nautik	I was referring to https://docs.openstack.org/masakari/latest/ but comparing would be interesting
08:50:19	lyarwood	nautik: https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.1/html/high_availability_for_compute_instances/index that's the high level product docs
09:01:33	lyarwood	nautik: https://github.com/ClusterLabs/fence-agents/pull/303 <- this is how we fixed this behaviour change
09:01:41	lyarwood	nautik: I think that's what you're seeing right?
09:07:03	nautik	lyarwood: awesome, that's exactly what I needed! Thank you!
09:07:22	nautik	I'll discuss this further on Masakari's side :)
09:07:36	lyarwood	nautik: np
09:40:06	gibi	lyarwood: thanks for the help to nautik while I was off
09:44:58	lyarwood	gibi: np, we hit the same issues a while ago downstream, just took a while to find the bug and code links
09:46:08	gibi	lyarwood: does the whole disable a host needs an rpc call down to host is due to initiate the DISABLED trait management in placement?
09:47:16	lyarwood	gibi: yeah something like that, we just forced hosts down first and then disabled
09:47:23	lyarwood	gibi: avoiding the call
09:48:15	gibi	thanks
09:48:36	gibi	it feels a bit backward that we cannot disable a compute without the compute being up :)
10:40:10	openstackgerrit	Wenping Song proposed openstack/nova-specs master: Repropose vGPU support spec https://review.opendev.org/c/openstack/nova-specs/+/780452
10:49:33	lyarwood	elod: https://review.opendev.org/c/openstack/nova/+/751364/ would you mind hitting these again
10:49:49	lyarwood	elod: sean-k-mooney has some backports for train that would be made easier with these landed
10:50:45	elod	lyarwood: sure, will do
10:50:46	sean-k-mooney	they would yes though i woudl prefer not to have to respine it again to pic them up since i have already got it working without them
10:51:38	sean-k-mooney	a week ago i would have said yes hold hartedly but i already spent the day and a half getting everything to work
10:51:59	lyarwood	sean-k-mooney: well either way we will need to respin something
10:52:32	sean-k-mooney	why did you request changes?
10:53:05	sean-k-mooney	i dont see any comments on the review
10:53:25	lyarwood	sean-k-mooney: well if https://review.opendev.org/c/openstack/nova/+/751364/ lands you don't need all the inline changes
10:53:30	lyarwood	sean-k-mooney: no not yet
10:54:09	sean-k-mooney	true thoug i was hoping they but they also dont hurt
10:54:27	lyarwood	yeah I guess everything is inlined so there's no conflicts
13:32:28	openstackgerrit	Lucas Alvares Gomes proposed openstack/nova master: [OVN] Explicitly set grenade job to ML2/OVS https://review.opendev.org/c/openstack/nova/+/776934
13:33:24	openstackgerrit	Lucas Alvares Gomes proposed openstack/nova master: [OVN] Explicitly set nova-next job to ML2/OVS https://review.opendev.org/c/openstack/nova/+/776944
13:34:19	openstackgerrit	Lucas Alvares Gomes proposed openstack/nova master: DO NOT REVIEW: Test OVN devstack module https://review.opendev.org/c/openstack/nova/+/776945
15:50:15	gibi	nova weekly meeting starts in 10 minutes in #openstack-meeting-3
15:56:03	gmann	gibi: did you update the ptg etherpad for nova+keystone sessions timing? https://etherpad.opendev.org/p/nova-xena-ptg
15:58:28	gibi	gmann: I did but I guess I was disconnected and it wasnt saved I fixed it now
15:58:59	gmann	gibi: i see now, thanks
15:59:09	gibi	thanks for the headsup
20:04:13	openstackgerrit	Merged openstack/nova master: docs: Change 'add_stylesheet' to 'add_css_file' https://review.opendev.org/c/openstack/nova/+/781529
20:04:43	openstackgerrit	Merged openstack/nova master: docs: Add CPU models guide https://review.opendev.org/c/openstack/nova/+/781438
22:04:45	openstackgerrit	Merged openstack/nova master: compute: Reject requests to commit intermediary snapshot of an inactive instance https://review.opendev.org/c/openstack/nova/+/781138
22:33:32	openstackgerrit	Ghanshyam proposed openstack/nova master: Switch to new rolevar for run-temepst role https://review.opendev.org/c/openstack/nova/+/775040
#openstack-nova - 2021-04-09
01:16:14	openstackgerrit	Wenping Song proposed openstack/nova-specs master: Repropose vGPU support spec https://review.opendev.org/c/openstack/nova-specs/+/780452
01:32:18	openstackgerrit	Wenping Song proposed openstack/nova-specs master: Repropose vGPU support spec https://review.opendev.org/c/openstack/nova-specs/+/780452
01:55:21	openstackgerrit	Merged openstack/nova master: libvirt: Simplify device_path check in _detach_encryptor https://review.opendev.org/c/openstack/nova/+/778463
01:55:46	openstackgerrit	Merged openstack/nova master: trivial: fix word duplication in api ref https://review.opendev.org/c/openstack/nova/+/782028
07:03:09	openstackgerrit	Qiu Fossen proposed openstack/nova-specs master: Repropose copying pmem devices data spec https://review.opendev.org/c/openstack/nova-specs/+/785563
08:18:53	openstackgerrit	Xinran WANG proposed openstack/nova-specs master: Repropose smartnic support spec https://review.opendev.org/c/openstack/nova-specs/+/783632
08:21:04	xinranwang	gibi: Hi gibi, I have reproposed the smartnic spec, please review it when you got time. https://review.opendev.org/c/openstack/nova-specs/+/783632
08:49:26	openstackgerrit	Qiu Fossen proposed openstack/nova-specs master: Repropose copying pmem devices data spec https://review.opendev.org/c/openstack/nova-specs/+/785563
09:04:22	openstackgerrit	Lee Yarwood proposed openstack/nova master: zuul: Replace grenade and nova-grenade-multinode with grenade-multinode https://review.opendev.org/c/openstack/nova/+/778885
09:04:27	openstackgerrit	Lee Yarwood proposed openstack/nova master: zuul: Remove nova-dsvm-multinode-base https://review.opendev.org/c/openstack/nova/+/778908
09:06:22	openstackgerrit	Lee Yarwood proposed openstack/nova stable/wallaby: compute: Reject requests to commit intermediary snapshot of an inactive instance https://review.opendev.org/c/openstack/nova/+/785576
09:06:57	openstackgerrit	Lee Yarwood proposed openstack/nova stable/wallaby: libvirt: Simplify device_path check in _detach_encryptor https://review.opendev.org/c/openstack/nova/+/785577
09:07:35	openstackgerrit	Lee Yarwood proposed openstack/nova stable/victoria: compute: Reject requests to commit intermediary snapshot of an inactive instance https://review.opendev.org/c/openstack/nova/+/785579
09:16:36	openstackgerrit	Lee Yarwood proposed openstack/nova stable/ussuri: compute: Reject requests to commit intermediary snapshot of an inactive instance https://review.opendev.org/c/openstack/nova/+/785582
09:17:26	openstackgerrit	Lee Yarwood proposed openstack/nova stable/train: compute: Reject requests to commit intermediary snapshot of an inactive instance https://review.opendev.org/c/openstack/nova/+/785583
09:18:39	openstackgerrit	Lee Yarwood proposed openstack/nova stable/victoria: libvirt: Simplify device_path check in _detach_encryptor https://review.opendev.org/c/openstack/nova/+/785584
09:19:14	openstackgerrit	Lee Yarwood proposed openstack/nova stable/ussuri: libvirt: Simplify device_path check in _detach_encryptor https://review.opendev.org/c/openstack/nova/+/785585
09:32:18	openstackgerrit	Lee Yarwood proposed openstack/nova stable/train: libvirt: Remove native LUKS compat code https://review.opendev.org/c/openstack/nova/+/766210
09:32:19	openstackgerrit	Lee Yarwood proposed openstack/nova stable/train: libvirt: Skip encryption metadata lookups if secret already exists on host https://review.opendev.org/c/openstack/nova/+/765771
10:56:01	openstackgerrit	Lee Yarwood proposed openstack/nova master: WIP zuul: Add evacuation migration record assertions to run-evacuate-hook https://review.opendev.org/c/openstack/nova/+/785618
11:32:53	openstackgerrit	Lee Yarwood proposed openstack/nova stable/ussuri: Rebase qcow2 images when unshelving an instance https://review.opendev.org/c/openstack/nova/+/785625
11:39:52	openstackgerrit	Lee Yarwood proposed openstack/nova stable/train: Rebase qcow2 images when unshelving an instance https://review.opendev.org/c/openstack/nova/+/785627
12:36:31	lyarwood	/ j #tripleo
12:36:33	lyarwood	argh
14:29:10	openstackgerrit	Lee Yarwood proposed openstack/nova master: glance: Improve [glance]enable_rbd_download and associated options help text https://review.opendev.org/c/openstack/nova/+/785656
14:33:43	openstackgerrit	sean mooney proposed openstack/nova master: Remove RBD swap disk if resizing to a flavor without swap https://review.opendev.org/c/openstack/nova/+/534711
15:08:22	openstackgerrit	Merged openstack/nova master: [OVN] Adapt the live-migration job scripts to work with OVN https://review.opendev.org/c/openstack/nova/+/776419
15:13:15	lyarwood	gibi: https://bugs.launchpad.net/nova/+bug/1923206 - FYI this has come up downstream and upstream in TripleO
15:13:18	openstack	Launchpad bug 1923206 in OpenStack Compute (nova) "libvirt.libvirtError: internal error: unable to execute QEMU command 'device_del': Device $device is already in the process of unplug" [Undecided,New] - Assigned to Lee Yarwood (lyarwood)
15:13:31	lyarwood	gibi: I'm looking at fixing it now ahead of your rework series landing in Xena
15:13:50	lyarwood	gibi: it's just awkward given the blind retry logic that we had previously
15:14:48	gibi	lyarwood: ack, sorry that I had no time to go back to the bugfix
15:16:36	lyarwood	gibi: np I'm going to work on this and then I can try to push that along again early in Xena
15:16:44	lyarwood	gibi: as once again it would likely avoid this situation
15:18:58	sean-k-mooney	lyarwood: did you make the parmater tuneable?
15:19:14	sean-k-mooney	lyarwood: could they just increase the retry interval
15:19:27	sean-k-mooney	to say 30 seconds
15:20:37	lyarwood	sean-k-mooney: with this it's the initial interval so no I don't think we can, it's hardcoded in the decorator iirc
15:21:19	sean-k-mooney	ok i tought you added workaround config options for this when we increased them
15:21:29	sean-k-mooney	but looks like we did not

Earlier   Later