Posted	Nick	Remark
#openstack-nova - 2019-11-04
20:10:46	mriedem	s'ok
20:11:10	mriedem	that rollback code is all gorpy
20:12:45	artom	mriedem, fixed
20:12:53	artom	mriedem, lemme look it over once again, then +1
20:13:04	mriedem	sure, thanks
20:13:39	artom	I was busy shitting all over https://review.opendev.org/#/c/512815/8
20:13:42	artom	Politely
20:18:10	openstackgerrit	Merged openstack/nova stable/rocky: Avoid error 500 on shelve task_state race https://review.opendev.org/692631
20:21:12	eandersson	btw anyone got any experience with the keep alive issue and openstack services? I noticed a bug or two reported for nova and failures
20:21:58	eandersson	> 'Connection aborted.', BadStatusLine("''",)
20:22:13	eandersson	A few of the bug reports indicates that disabling keepalive would resolve some of these.
20:26:00	efried	mriedem: One question on the qos patches (I'm on the bottom one rn): It looks like some of the code is being hit for both evacuate and rebuild. It wouldn't do any harm for the latter, since it's just populating the rg/rp mappings. But just wanted to confirm that's what I'm seeing?
20:26:05	mriedem	artom: oh wow that's fun, smells like starlingx
20:26:29	artom	mriedem, yeah eh?
20:26:32	mriedem	starlingx did a lot of hacks with libvirt channels
20:26:37	efried	mriedem: specifically these changes? https://review.opendev.org/#/c/688387/8/nova/conductor/manager.py
20:27:45	mriedem	artom: good comments on that patch btw
20:27:46	mriedem	all valid poitns
20:27:48	mriedem	*points
20:27:53	artom	I'm useful!
20:28:04	mriedem	makes up for your lack of if/elif knowledge :)
20:28:08	artom	Hahaha
20:28:38	mriedem	efried: hmm, weird
20:28:42	mriedem	s/weird/where/?
20:28:51	mriedem	https://review.opendev.org/#/c/688387/8/nova/compute/manager.py only does the port mapping stuff if evacuate
20:29:05	mriedem	same in https://review.opendev.org/#/c/688387/8/nova/conductor/manager.py
20:29:20	mriedem	or are you asking, why do we only do that for evacuate?
20:29:26	efried	No
20:29:56	efried	in one file the bool is called 'evacuate', so I can buy that it's only happening for evacuate. In the other it's called 'recreate' and the comments imply we're hitting the path for both evacuate and rebuild.
20:30:11	mriedem	in compute manager that evacuate variable used to be called recreate as well,
20:30:13	efried	so either it's the same and the var name and comments are confusing, or...
20:30:16	mriedem	i changed that awhile back,
20:30:25	mriedem	it's still recreate in conductor b/c no one has renamed the variable for clarity yet
20:30:53	efried	okay. So recreate is evacuate.
20:31:06	efried	and... `not recreate` is rebuild, which is a kind of recreate, but whatever.
20:31:16	mriedem	right
20:31:19	mriedem	hence the confusion
20:31:26	efried	thanks for clarifying.
20:31:36	efried	someone should rename that variable.
20:31:49	mriedem	https://review.opendev.org/#/c/508190/
20:31:56	mriedem	that's me doing it in compute
20:33:39	efried	ah, so you didn't actually try to rename it all the way through the stack. That would indeed be hairy.
20:33:58	mriedem	right, i wasn't going to touch rpc stuff for that
20:34:10	mriedem	and the driver interface is likely nbd but it's at least an email
20:34:33	mriedem	there is no in-tree driver that implements that anymore anyway, it was around for the old baremetal driver that eventually became ironic
20:39:33	openstackgerrit	Eric Fried proposed openstack/nova master: cond: rename 'recreate' var to 'evacuate' https://review.opendev.org/692900
20:39:35	efried	mriedem: ^
20:51:16	mriedem	comments inline
21:03:52	openstackgerrit	Eric Fried proposed openstack/nova master: cond: rename 'recreate' var to 'evacuate' https://review.opendev.org/692900
21:04:47	efried	done and done
21:06:21	efried	and +A on gibi_ptg's patches.
21:07:29	mriedem	sweet
21:09:17	efried	Gonna go do some more damage on osc, then come back to vtpm
21:24:49	mriedem	heave ho
21:25:23	openstackgerrit	Matt Riedemann proposed openstack/nova master: Follow up to I3e28c0163dc14dacf847c5a69730ba2e29650370 https://review.opendev.org/692856
21:25:24	openstackgerrit	Matt Riedemann proposed openstack/nova master: Pass exception through TaskBase.rollback https://review.opendev.org/692689
21:25:24	openstackgerrit	Matt Riedemann proposed openstack/nova master: Execute CrossCellMigrationTask from MigrationTask https://review.opendev.org/635668
21:25:25	openstackgerrit	Matt Riedemann proposed openstack/nova master: Refresh instance in MigrationTask.execute Exception handler https://review.opendev.org/669012
21:25:25	openstackgerrit	Matt Riedemann proposed openstack/nova master: Plumb allow_cross_cell_resize into compute API resize() https://review.opendev.org/635684
21:25:26	openstackgerrit	Matt Riedemann proposed openstack/nova master: Filter duplicates from compute API get_migrations_sorted() https://review.opendev.org/636224
21:25:26	openstackgerrit	Matt Riedemann proposed openstack/nova master: Start functional testing for cross-cell resize https://review.opendev.org/636253
21:25:27	openstackgerrit	Matt Riedemann proposed openstack/nova master: Handle target host cross-cell cold migration in conductor https://review.opendev.org/642591
21:25:27	openstackgerrit	Matt Riedemann proposed openstack/nova master: Validate image/create during cross-cell resize functional testing https://review.opendev.org/642592
21:25:28	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add zones wrinkle to TestMultiCellMigrate https://review.opendev.org/643450
21:25:28	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add negative test for cross-cell finish_resize failing https://review.opendev.org/643451
21:25:29	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add negative test for prep_snapshot_based_resize_at_source failing https://review.opendev.org/669013
21:25:29	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add confirm_snapshot_based_resize_at_source compute method https://review.opendev.org/637058
21:25:30	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add ConfirmResizeTask https://review.opendev.org/637070
21:25:30	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add confirm_snapshot_based_resize conductor RPC method https://review.opendev.org/637075
21:25:31	openstackgerrit	Matt Riedemann proposed openstack/nova master: Confirm cross-cell resize from the API https://review.opendev.org/637316
21:25:31	openstackgerrit	Matt Riedemann proposed openstack/nova master: Add revert_snapshot_based_resize_at_dest compute method https://review.opendev.org/637630
21:25:32	openstackgerrit	Matt Riedemann proposed openstack/nova master: Deal with cross-cell resize in _remove_deleted_instances_allocations https://review.opendev.org/639453
21:30:54	openstackgerrit	melanie witt proposed openstack/nova stable/stein: Add regression test for bug 1824435 https://review.opendev.org/692906
21:30:54	openstack	bug 1824435 in OpenStack Compute (nova) stein "fill_virtual_interface_list migration fails on second attempt" [Medium,Triaged] https://launchpad.net/bugs/1824435
21:30:54	openstackgerrit	melanie witt proposed openstack/nova stable/stein: Remove redundant call to get/create default security group https://review.opendev.org/692907
21:48:43	mriedem	melwitt: i saw this today https://i.imgur.com/beq9YYf.jpg
21:50:06	melwitt	lol yesss
22:06:28	mriedem	dansmith: another spec which we've basically said no to in the past but the pushback is "we have some customers that want nova to do this" https://review.opendev.org/#/c/672400/
22:06:41	mriedem	because pre-creating a port is hard
22:09:05	mriedem	if enough diverse users/vendors came forward asking for this like people did for passing through volume type on boot from volume then maybe it's a different discussion
22:09:41	dansmith	well, I haven't read that other than the commit message, but..
22:09:45	dansmith	I'm pretty conflicted on these
22:10:19	dansmith	since we're split, it really seems like an impossible-to-win scenario chasing every attribute of the other service, especially when they version their api so differently from us
22:10:42	mriedem	"we're split" meaning we (nova) from neutron?
22:10:45	dansmith	but I also really sympathize with simple things being hard because of how we decided in 2013 to segregate the project
22:10:46	dansmith	yes
22:11:49	mriedem	i'm pretty sure vnic_type is going to be in any neutron deployment because we rely on that extension pretty heavily https://docs.openstack.org/api-ref/network/v2/index.html#port-binding-extended-attributes
22:12:38	mriedem	alright, i won't block it
22:12:51	mriedem	i'm fairly certain starlingx had that in their nova fork already
22:13:00	mriedem	maybe sean-k-mooney would know if red hat wants that as well
22:13:06	dansmith	I'm not saying we shouldn't block it
22:13:22	dansmith	I'm saying we're screwed either way
23:00:32	openstackgerrit	Merged openstack/nova master: Allow evacuating server with port resource request https://review.opendev.org/688387
23:00:38	openstackgerrit	Merged openstack/nova master: Enable evacuation with qos ports https://review.opendev.org/688688
23:11:50	efried	dansmith: help me here, I swear somewhere we had a functional test to validate the redirects in our .htaccess file??
23:14:24	openstackgerrit	melanie witt proposed openstack/nova stable/stein: Add regression test for bug 1824435 https://review.opendev.org/692906
23:14:24	openstack	bug 1824435 in OpenStack Compute (nova) stein "fill_virtual_interface_list migration fails on second attempt" [Medium,In progress] https://launchpad.net/bugs/1824435 - Assigned to melanie witt (melwitt)
23:14:24	openstackgerrit	melanie witt proposed openstack/nova stable/stein: Remove redundant call to get/create default security group https://review.opendev.org/692907
23:14:24	openstackgerrit	melanie witt proposed openstack/nova stable/stein: Add integration testing for heal_allocations https://review.opendev.org/692923
23:14:55	efried	found it

Earlier   Later