Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2022-09-30
22:38:15 atmark https://paste.openstack.org/show/b9tx6I2rQ3tUtsGBR98g/
22:38:20 sean-k-mooney what im wondering is does that exist and is the cell_id set
22:38:32 sean-k-mooney and if so is that cell in your cell_mappings table
22:38:54 sean-k-mooney ok so cell_id is 6
22:39:11 sean-k-mooney the cell mappings
22:39:16 sean-k-mooney has potitally password inf
22:39:22 sean-k-mooney so dont paste that
22:39:46 atmark 6 exist
22:39:49 sean-k-mooney but does the db connection string for cell id 6 end in nova
22:39:49 atmark https://paste.openstack.org/show/bh9PvsUNrxNuYSRsATga/
22:40:08 atmark 6 and 9
22:40:13 sean-k-mooney yep
22:40:20 sean-k-mooney does 6 point to the db where the instance iss
22:40:25 sean-k-mooney i.e. nova in your case
22:40:48 sean-k-mooney it shoudl end in :3306/nova
22:41:10 sean-k-mooney this is how we know which database to look in to delete the instace at the cell level
22:41:39 atmark https://paste.openstack.org/show/blXqFYFedaFGS9U7igPG/
22:41:50 atmark cell0
22:42:06 atmark it's pointing to nova_cell0
22:42:24 sean-k-mooney ya so cell0 is where it should be if it fail
22:42:43 sean-k-mooney can you ceck cell 0 and see if its also in the instance table there
22:43:48 atmark yup it's here
22:44:05 sean-k-mooney ok so its in both nova and nova_cell0
22:44:10 sean-k-mooney thats really strange
22:44:14 atmark correct
22:44:40 sean-k-mooney so the quickest way to fix this is just to delete that one record
22:44:50 sean-k-mooney but it shoudl only ever end up in one of those two dbs
22:45:07 sean-k-mooney any instnace that failed before it was schduled to a host should end up in nova_cell0
22:45:14 atmark delete in both db?
22:45:38 atmark probably safe to delete in both right
22:45:44 sean-k-mooney you could try just removing it in nova
22:45:57 sean-k-mooney and then delete again but ya it should be safe to delete in both
22:46:27 sean-k-mooney i dont know of a failrue mode like this unless you change you cell mapping at some point
22:46:44 sean-k-mooney normally cell 0 get id 1 an cell_1 gets id 2
22:47:05 sean-k-mooney although there is no significance to the speicic id
22:47:13 atmark iirc, we did played with mappings because something failed
22:47:40 sean-k-mooney ack so maybe this change at some point
22:48:07 sean-k-mooney i have never seen this specific issue before
22:48:23 atmark yep. anywany, this is for monday. don't wanna cause something bad
22:48:40 atmark thanks for the help
22:49:18 atmark we have 3 prod envs, only 1 env is exhibiting this issue
22:49:25 sean-k-mooney ack
22:49:33 atmark all same version
22:50:21 sean-k-mooney its my favorite way to install and manage openstack
22:50:34 sean-k-mooney im just sad that the zed release is not avlaible currently
#openstack-nova - 2022-10-01
08:59:46 opendevreview Takashi Natsume proposed openstack/nova master: Add a hacking rule for the setDaemon method https://review.opendev.org/c/openstack/nova/+/854653
#openstack-nova - 2022-10-03
03:13:40 abdi OpenStack Nova Team. I filed the following bug against tempest: https://bugs.launchpad.net/tempest/+bug/1989232. However, it has not got a lot of attention from the tempest side. My ongoing investigation suggests nova may be the culprit. Could you review my last posting on the bug and comment?
08:06:37 opendevreview Amit Uniyal proposed openstack/nova master: Adds check if resized to swap zero https://review.opendev.org/c/openstack/nova/+/857339
09:32:12 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: compute: enhance compute evacuate instance to support target state https://review.opendev.org/c/openstack/nova/+/858383
09:32:13 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: api: extend evacuate instance to support target state https://review.opendev.org/c/openstack/nova/+/858384
09:45:16 auniyal Hi gibi
09:45:29 auniyal I have a doubt in this comment, can you please confirm, the changes
09:45:29 auniyal https://review.opendev.org/c/openstack/nova/+/791135/7
09:46:13 auniyal last comment in this patch
09:55:19 gibi auniyal: what I ment there is that you have to restore the state of the instance object to point to the source host. And that needs not just the instance.host but also the instnace.node to be set properly. You can copy the instance current node at the start of the post_live_migration_at_destination call in a local variabla and then use that to restore instance.node in the finally block
10:10:28 sean-k-mooney gibi: no
10:10:48 sean-k-mooney gibi: post live migration is after we are allowed to rollback
10:10:54 gibi sorry
10:10:57 gibi I mixed source det
10:10:58 gibi dest
10:11:03 sean-k-mooney oh ok
10:11:21 gibi so it is not enough to set the instance.host but need to set the instance.node
10:11:24 gibi to the dest
10:11:27 sean-k-mooney yes
10:11:38 gibi but then auniyal you have to look up the dest node
10:12:21 sean-k-mooney well your running on the test at this point
10:12:26 auniyal isn't dest node is same as dest host ? or dest
10:12:31 sean-k-mooney so you just need to set it im not sure you need to look it up
10:12:41 sean-k-mooney auniyal: for libvirt yes
10:12:44 sean-k-mooney in general no
10:12:52 sean-k-mooney actully no
10:13:01 sean-k-mooney even for libvirt they can be different
10:13:19 sean-k-mooney although i think that owuld be an error
10:15:02 sean-k-mooney node_name = compute_node.hypervisor_hostname
10:15:28 sean-k-mooney where as instance.host = CONF.host
10:16:01 auniyal queston, what is diff between host and node, till now, I thought the instance.host is where compute service is running, (node is as general we say in any network tree/mapping (machine))
10:16:02 sean-k-mooney by default those shoudl be the same but if you manually set [DEFAULT]/host
10:16:09 sean-k-mooney then they can be differnt
10:16:26 sean-k-mooney node is the name of the comptue node
10:16:33 sean-k-mooney host is the name of the compute service
10:16:58 sean-k-mooney for non clustered drivers like libvirt they are normally the same
10:17:52 sean-k-mooney for clustered dirvers like ironic where 1 compute service manages many baremetal servers
10:17:55 sean-k-mooney they are differnet
10:18:31 sean-k-mooney you can look up the name like this
10:18:33 sean-k-mooney https://github.com/openstack/nova/blob/1025c9879341d44db33c4cc501435364dd185a9e/nova/compute/manager.py#L9263-L9267
10:19:58 auniyal ack
11:48:15 opendevreview Andre Aranha proposed openstack/nova master: Replace Centos 8 jobs for Centos 9 https://review.opendev.org/c/openstack/nova/+/858272
11:49:47 opendevreview Andre Aranha proposed openstack/nova master: Remove the periodic Centos 8 job https://review.opendev.org/c/openstack/nova/+/858272
12:02:36 opendevreview Andre Aranha proposed openstack/nova stable/yoga: Test setting the nova job to centos-9-stream https://review.opendev.org/c/openstack/nova/+/860087
12:39:35 opendevreview Maksim Malchuk proposed openstack/nova stable/xena: Fix to implement 'pack' or 'spread' VM's NUMA cells https://review.opendev.org/c/openstack/nova/+/829804
14:48:13 atmark Hello. Getting this error when trying to temporarily increase allocation_ratio in placement via CLI
14:48:23 atmark JSON does not validate: 'total' is a required property Failed validating 'required' in schema['properties']['inventories']['patternProperties']['^[A-Z0-9_]+$']:
14:48:53 atmark ./openstack resource provider inventory set c80259e6-67fb-47a0-b04f-364b2f2b2969 --resource MEMORY_MB:allocation_ratio=1.0
16:14:10 gibi atmark: either define the total value in --resource too, or use --amend to instruct the client to only change the allocation_ratio
17:19:49 atmark gibi: --amend works, thanks
17:44:03 gibi atmark: :)
#openstack-nova - 2022-10-04
08:15:39 bauzas gibi: your help is needed for a janitory change https://review.opendev.org/c/openstack/nova-specs/+/856173
08:20:33 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova-specs master: spec: allowing target state for evacuate https://review.opendev.org/c/openstack/nova-specs/+/857838
08:36:51 gibi bauzas: done
08:37:01 bauzas thanks

Earlier   Later