Earlier  
Posted Nick Remark
#openstack-nova - 2020-12-11
11:50:10 sean-k-mooney the order shoudl be 0ab8a03c782851d4019c90e6159bd1f83c0658e6 then 17ba10b75961ef667e45b9dedbac34212a7e7e9b then your patch
11:51:15 lyarwood hemanth_n: ack, I can try to do this later today for you
11:51:31 hemanth_n sean-k-mooney, lyardwood: thanks for the quick responses once again
12:06:33 openstackgerrit Lee Yarwood proposed openstack/nova master: nova-grenade-multinode: Skip test_live_block_migration_paused https://review.opendev.org/c/openstack/nova/+/766720
12:07:04 lyarwood gibi: ^ lets skip that test until multinode grenade moves to zuulv3 and Focal
12:08:59 brinzhang hi team, when do we support GET /allocation_candidates?resources=CUSTOM_PMEM_NAMESPACE_xxGB? in nova and placement?
12:09:52 brinzhang I know support this when we accept the PMEM support patch in Nova, but Icannot find the placement changes, anyone know?
12:11:49 sean-k-mooney brinzhang: there was no placment change required
12:11:51 brinzhang we hit an error: GET /allocation_candidates?limit=1000&resources=CUSTOM_PMEM_NAMESPACE_10GB%3A1%2CDISK_GB%3A30%2CMEMORY_MB%3A2048%2CVCPU%3A2 HTTP/1.1" 400
12:12:49 sean-k-mooney brinzhang: what was the actul error in the placement logs
12:13:55 sean-k-mooney its asking for CUSTOM_PMEM_NAMESPACE_10GB:1,DISK_GB:30,MEMORY_MB:2048,VCPU:24
12:14:15 sean-k-mooney that looks like a perfectly normal request to me
12:14:24 stephenfin brinzhang: what's configured in '[libvirt] pmem_namespaces' in nova.conf? You should have an entry like 'pmem_namespaces = 10GB:ns0,...
12:14:33 brinzhang sean-k-mooney: please see: Invalid resource class in resources parameter: No such resource class CUSTOM_PMEM_NAMESPACE_30GB
12:14:52 stephenfin then nova isn't reporting the inventory
12:15:14 sean-k-mooney yep nova should create the inventory and register teh resouce class
12:15:37 stephenfin either you don't have it configured, your configuration is wrong, or you're using nova < Train
12:15:45 brinzhang stephenfin: pmem_namaspaces = 10GB:ns10_0|ns10_1|ns10_2|ns10_3,30GB:ns30_0|ns30_1|ns30_2|ns30_3
12:15:54 brinzhang this is our config in nova.conf
12:16:31 brinzhang stephenfin: does the config is right?
12:17:13 stephenfin it looks okay
12:17:16 stephenfin and what does 'openstack resource provider inventory list $COMPUTE_ID' show?
12:17:18 sean-k-mooney based on the docs yes
12:18:05 sean-k-mooney brinzhang: you precreated the namespaces right?
12:18:15 sean-k-mooney you have to do that manually nova will not do that for you
12:18:34 brinzhang yes, we precreated the namespaces manully
12:19:00 brinzhang s/manully/manually
12:19:01 sean-k-mooney ok then ya can you provide the output of the command stephenfin asked for
12:19:11 brinzhang let my try
12:20:37 openstackgerrit Lee Yarwood proposed openstack/nova stable/train: DNM - Testing https://review.opendev.org/c/openstack/devstack/+/766622 https://review.opendev.org/c/openstack/nova/+/766723
12:21:00 sean-k-mooney if the inventories are not present you need to check the nova comptue agent to see if the update_avaiable_resoue periodic task is failing and if so why
12:22:11 gibi lyarwood: +2 on the skipping the failing live migration test
12:37:53 brinzhang sean-k-mooney, stephenfin: yes, it's empty, Ishould check the update_avaiable_resoue periodic task, thanks.
12:38:11 brinzhang I will check it later ^
12:44:44 openstackgerrit Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from List SG API https://review.opendev.org/c/openstack/nova/+/766726
12:45:16 openstackgerrit Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from List SG API https://review.opendev.org/c/openstack/nova/+/766726
12:50:44 openstackgerrit Alexandre arents proposed openstack/nova master: libvirt: make cross cell resize spawn from snapshot image https://review.opendev.org/c/openstack/nova/+/765561
12:53:18 openstackgerrit Merged openstack/nova master: Disable qos resize tempest test for nova-multi-cell job https://review.opendev.org/c/openstack/nova/+/766471
12:59:49 gibi ^^ \o/
13:02:47 openstackgerrit Alexandre arents proposed openstack/nova master: libvirt: make cross cell resize spawn from snapshot image https://review.opendev.org/c/openstack/nova/+/765561
13:08:15 openstackgerrit Lee Yarwood proposed openstack/nova stable/victoria: nova-grenade-multinode: Skip test_live_block_migration_paused https://review.opendev.org/c/openstack/nova/+/765749
13:08:50 lyarwood whops
13:17:33 brinzhang gibi, stephenfin, sean-k-mooney: hope you can review cyborg shelve/unshelve support https://review.opendev.org/c/openstack/nova/+/729563
13:20:26 openstackgerrit Lee Yarwood proposed openstack/nova stable/victoria: lower-constraints: Bump packaging to 20.4 https://review.opendev.org/c/openstack/nova/+/765749
13:25:13 lyarwood brb in 30
13:36:42 gibi brinzhang: sorry, I won't get to it this week
13:45:06 openstackgerrit Alexandre arents proposed openstack/nova master: Revert "nova-multi-cell: Skip test_cold_migrate_unshelved_instance" https://review.opendev.org/c/openstack/nova/+/765933
13:56:57 openstackgerrit Lee Yarwood proposed openstack/nova stable/train: [stable-only] Cap bandit to 1.6.2 https://review.opendev.org/c/openstack/nova/+/766171
13:59:29 lyarwood was anyone working on ERROR: No matching distribution found for hacking<3.1.0,>=3.0.1 in stable/ussuri?
14:01:15 gibi lyarwood: I'm not aware of such work
14:01:34 openstack Launchpad bug 1907756 in OpenStack Compute (nova) "ERROR: No matching distribution found for hacking<3.1.0,>=3.0.1" [Undecided,New]
14:01:34 lyarwood https://bugs.launchpad.net/nova/+bug/1907756 kk thanks, we have a bug at least :)
14:06:42 kashyap sean-k-mooney: Fair point on the upgrade impact and fixing the version history. Thanks for the quick look
14:10:39 lyarwood http://paste.openstack.org/show/800970/ excellent, pip sorted out the awful error message in the latest version
14:10:56 kashyap sean-k-mooney: Do you have an example of how 'placement-manage' can register a trait?
14:14:18 openstackgerrit Lee Yarwood proposed openstack/nova stable/ussuri: [stable-only] Bump hacking to 3.0.1 in lower-contraints.txt https://review.opendev.org/c/openstack/nova/+/766738
14:16:05 bauzas kashyap: not placement-manage
14:16:12 bauzas kashyap: osc-placement rather
14:16:28 bauzas you ask for a custom trait, right?
14:16:42 kashyap Ah, okay; then in the review here, Sean meant 'osc-placement', then: https://review.opendev.org/c/openstack/nova-specs/+/759731
14:17:00 bauzas kashyap: see this example https://docs.openstack.org/nova/latest/admin/virtual-gpu.html#optional-provide-custom-traits-for-multiple-gpu-types
14:17:00 kashyap bauzas: See this part of Sean's comment:
14:17:02 kashyap [quote]
14:17:03 kashyap run to register the new standard traits before nova is upgraded."
14:17:03 kashyap on the host runing plamcent and the placement-manage executable to be
14:17:03 kashyap ... this will require os-traits to be installed
14:17:05 kashyap [/quote]
14:17:14 kashyap bauzas: /me clicks
14:17:41 bauzas ah, this
14:18:39 kashyap I'm placement-unaware -- so is there an example of "placement-manage executable to be run to register the new standard traits"
14:19:26 openstackgerrit Lee Yarwood proposed openstack/nova stable/ussuri: [stable-only] lower-contraints: Bump hacking to 3.0.1 https://review.opendev.org/c/openstack/nova/+/766738
14:20:28 kashyap bauzas: The only examples of "placement-manage" I see are during data migration here: https://docs.openstack.org/placement/latest/admin/upgrade-to-stein.html
14:20:49 kashyap [Scroll to the bottom to see the 'placement-manage' invocation]
14:30:21 kashyap bauzas: Ah-ha! I know what Sean meant: to sync the 'os-traits' into Placement: `placement-manage os-traits sync`
14:30:56 bauzas correct, sorry I forgot to answer
14:31:09 bauzas you need to tell you support a new os-trait
14:31:12 kashyap No problem
14:32:49 bauzas sean-k-mooney: dumb question, but as I copied from https://review.opendev.org/c/openstack/nova/+/656885/7/nova/network/neutron.py, do you know what we should verify the multi-provider extension given we only call the segments API extension ?
14:34:16 openstackgerrit Kashyap Chamarthy proposed openstack/nova-specs master: Re-propose "Secure Boot support for KVM & QEMU guests" for Wallaby https://review.opendev.org/c/openstack/nova-specs/+/759731
14:34:26 kashyap gibi: sean-k-mooney: Fixed --^ Hope that looks okay
14:37:39 sean-k-mooney am you repeaded placment manage
14:37:45 sean-k-mooney service by ``placement-manage os-traits sync``placement-manage`` .
14:38:01 sean-k-mooney but ya other wise it looks good to me
14:43:31 gibi sean-k-mooney: does placement-manager os-traits sync exists?
14:44:53 sean-k-mooney actully im not sure
14:45:03 sean-k-mooney there is dbsync
14:45:13 gibi yeah I only see db sync
14:45:14 sean-k-mooney perhaps placment jsut does the traits import on startu
14:45:19 sean-k-mooney *startup
14:46:42 gibi sean-k-mooney: it seems it does it at startup https://github.com/openstack/placement/blob/10146978ea2d7eace1ffdf6552eafd22360e37dd/placement/deploy.py#L133
14:47:26 sean-k-mooney ok cool in this case i guess you just need to install the new os-traits and restart
14:48:22 gibi yepp
14:49:42 kashyap sean-k-mooney: Sigh; fixing :)
14:50:44 kashyap gibi: Pretty sure I saw the "os-traits sync" somewhere -- let me dig it up
14:51:55 kashyap https://specs.openstack.org/openstack/nova-specs/specs/pike/implemented/resource-provider-traits.html
14:52:25 gibi kashyap: I did not found a corresponding implementation
14:53:01 kashyap gibi: Ah, okay. Then, what do you suggest here? :-) Just a 'placement-manage db sync' is okay?
14:53:25 gibi kashyap: as sean-k-mooney said, upgrade the os-traits lib and then restat the placement service
14:53:43 gibi placement service at startup will sync the traits from os-traits into the placement db
14:53:45 kashyap Okay; I'll adjust the text to that effect; thank you!

Earlier   Later