Earlier  
Posted Nick Remark
#openstack-nova - 2021-04-30
11:09:59 ierdem it says "Queue.declare: (404) NOT_FOUND". I can not reach the VMs. I don't know why rabbitmq queues deleted from the list, or why these queues can not recreated by service requests? Can you help me? How can I solve this queue problem and what is the reason of this? btw, this is nova-compute logs: http://paste.openstack.org/show/804880/ , Thanks
11:11:04 ierdem Rabbitmq and other services are working now, but problem still exist
11:11:05 ierdem Rabbitmq and other services are working now, but problem still exist
11:21:15 noonedeadpunk I'm not really sure that there's no problem with cluster now...
11:21:15 noonedeadpunk I'm not really sure that there's no problem with cluster now...
11:22:31 noonedeadpunk Does other services work properly? Or it's only nova that complain?
11:22:31 noonedeadpunk Does other services work properly? Or it's only nova that complain?
11:26:59 ierdem After restarting all services, nova-compute services seems work fine but neutron-server containers have the same error log now, "queue declare(404) not found". I could not inspect the all services for now, I am checking one by one but you probably right, there will be errors in cluster I think
11:26:59 ierdem After restarting all services, nova-compute services seems work fine but neutron-server containers have the same error log now, "queue declare(404) not found". I could not inspect the all services for now, I am checking one by one but you probably right, there will be errors in cluster I think
11:29:45 noonedeadpunk ierdem: jsut try running rabbitmq-install.yml -e rabbitmq_upgrade=true - it should fix issue with missing queues (with several mins of rabbit cluster downtime)
11:29:45 noonedeadpunk ierdem: jsut try running rabbitmq-install.yml -e rabbitmq_upgrade=true - it should fix issue with missing queues (with several mins of rabbit cluster downtime)
11:34:13 ierdem I am running now, thanks
11:34:13 ierdem I am running now, thanks
11:34:58 stephenfin lyarwood: Just a reminder to take a punt at https://review.opendev.org/c/openstack/nova/+/676209 if you could today
11:34:59 stephenfin lyarwood: Just a reminder to take a punt at https://review.opendev.org/c/openstack/nova/+/676209 if you could today
11:41:10 lyarwood stephenfin: ack still on my list, if not today it might have to wait until after PTO
11:41:10 lyarwood stephenfin: ack still on my list, if not today it might have to wait until after PTO
11:46:45 stephenfin (y)
11:46:45 stephenfin (y)
11:54:04 openstackgerrit Stephen Finucane proposed openstack/nova master: libvirt: Delegate OVS plug to os-vif https://review.opendev.org/c/openstack/nova/+/602432
11:54:04 openstackgerrit Stephen Finucane proposed openstack/nova master: libvirt: Delegate OVS plug to os-vif https://review.opendev.org/c/openstack/nova/+/602432
12:35:53 damiandabrowski hey! libvirt v6.0.0 fails to launch a domain when using qcow2 disks where the backing file format is not explicitly specified in the qcow2 metadata.
12:35:53 damiandabrowski We have merged a fix for newly created disks: https://review.opendev.org/c/openstack/nova/+/710790/
12:35:53 damiandabrowski But do we have any fix for already running VMs? Because now, they fails to spawn after reboot.
12:35:53 damiandabrowski hey! libvirt v6.0.0 fails to launch a domain when using qcow2 disks where the backing file format is not explicitly specified in the qcow2 metadata.
12:35:53 damiandabrowski We have merged a fix for newly created disks: https://review.opendev.org/c/openstack/nova/+/710790/
12:35:53 damiandabrowski But do we have any fix for already running VMs? Because now, they fails to spawn after reboot.
12:45:43 lyarwood damiandabrowski: I don't think we had to fix that as the libvirt devs eventually stopped enforcing that
12:45:44 lyarwood damiandabrowski: I don't think we had to fix that as the libvirt devs eventually stopped enforcing that
12:46:02 lyarwood damiandabrowski: which base OS is this?
12:46:03 lyarwood damiandabrowski: which base OS is this?
12:47:30 damiandabrowski <lyarwood "damiandabrowski: which base OS i"> ubuntu 18.04, libvirt 6.0.0-0ubuntu8.8~cloud0
12:47:30 damiandabrowski <lyarwood "damiandabrowski: which base OS i"> ubuntu 18.04, libvirt 6.0.0-0ubuntu8.8~cloud0
12:48:00 lyarwood kashyap: ^ the libvirt devs eventually removed that right?
12:48:00 lyarwood kashyap: ^ the libvirt devs eventually removed that right?
12:48:14 kashyap * kashyap reads
12:48:52 kashyap lyarwood: Yes, I recall; but I need to find trhe right commit
12:48:52 kashyap lyarwood: Yes, I recall; but I need to find trhe right commit
12:49:19 kashyap * kashyap digs up
13:08:52 kashyap damiandabrowski: lyarwood: There we go: "Allow format probing under special circumstances": https://github.com/libvirt/libvirt/commit/ae9e6c2a2b
13:08:52 kashyap damiandabrowski: lyarwood: There we go: "Allow format probing under special circumstances": https://github.com/libvirt/libvirt/commit/ae9e6c2a2b
13:09:02 kashyap damiandabrowski: lyarwood: That's the libvirt commit that relaxes the restriction
13:09:02 kashyap damiandabrowski: lyarwood: That's the libvirt commit that relaxes the restriction
13:09:50 lyarwood so you need >=6.1.0
13:09:51 lyarwood so you need >=6.1.0
13:09:53 kashyap damiandabrowski: That should be in 6.1.0 and above. (I checked via `git tag --contains`)
13:09:53 kashyap damiandabrowski: That should be in 6.1.0 and above. (I checked via `git tag --contains`)
13:09:56 kashyap lyarwood: Yep
13:09:56 kashyap lyarwood: Yep
13:10:31 lyarwood kashyap: I guess we could write a known issues releasenote about that
13:10:31 lyarwood kashyap: I guess we could write a known issues releasenote about that
13:11:06 kashyap lyarwood: Yeah; that's reasonable. I can volunteer to write one.
13:11:06 kashyap lyarwood: Yeah; that's reasonable. I can volunteer to write one.
13:13:54 lyarwood kashyap: awesome thanks
13:13:54 lyarwood kashyap: awesome thanks
13:16:31 damiandabrowski thanks a lot for the input!
13:16:31 damiandabrowski thanks a lot for the input!
13:16:40 kashyap I'd call it a "resolved known issue" :-)
13:16:40 kashyap I'd call it a "resolved known issue" :-)
14:09:39 dansmith gibi: but we should be checking for the *last* victoria version to determine if something is victoria, right?
14:09:39 dansmith gibi: but we should be checking for the *last* victoria version to determine if something is victoria, right?
14:12:59 noonedeadpunk hey there!
14:12:59 noonedeadpunk hey there!
14:13:48 gibi dansmith: when we says that Wallaby controller supports Victoria compute, doesn't that means it supports even first Victoria service version?
14:13:48 gibi dansmith: when we says that Wallaby controller supports Victoria compute, doesn't that means it supports even first Victoria service version?
14:14:30 gibi ~ the oldest thing we support is the oldest Victoria pieces
14:14:30 gibi ~ the oldest thing we support is the oldest Victoria pieces
14:15:03 noonedeadpunk We've faced nasty thing on bionic, regarding libvirt virStorageFileGetMetadataRecurse. While new images are created properly thanks to https://review.opendev.org/c/openstack/nova/+/710790/ I'm not sure about the way forward for existing ones
14:15:03 noonedeadpunk We've faced nasty thing on bionic, regarding libvirt virStorageFileGetMetadataRecurse. While new images are created properly thanks to https://review.opendev.org/c/openstack/nova/+/710790/ I'm not sure about the way forward for existing ones
14:15:20 noonedeadpunk as once VM stopps, it can't be started anymore
14:15:20 noonedeadpunk as once VM stopps, it can't be started anymore
14:16:06 noonedeadpunk I believe I'm not the first who comes with this, but yeah :(
14:16:06 noonedeadpunk I believe I'm not the first who comes with this, but yeah :(
14:16:24 dansmith gibi: that's not how we've ever done it for RPC, for example, but I guess it's fair.. it's just confusing though,
14:16:24 dansmith gibi: that's not how we've ever done it for RPC, for example, but I guess it's fair.. it's just confusing though,
14:16:38 lyarwood noonedeadpunk: it was discussed earlier in the channel, you need libvirt >= 6.1.0
14:16:39 lyarwood noonedeadpunk: it was discussed earlier in the channel, you need libvirt >= 6.1.0
14:16:45 dansmith because from the perspective of someone trying to figure out what it's looking for, the numbers don't match up
14:16:45 dansmith because from the perspective of someone trying to figure out what it's looking for, the numbers don't match up
14:16:46 lyarwood noonedeadpunk: we don't workaround this in nova
14:16:46 lyarwood noonedeadpunk: we don't workaround this in nova
14:16:56 dansmith gibi: but if that's the goal, then I guess it's okay
14:16:56 dansmith gibi: but if that's the goal, then I guess it's okay
14:17:26 noonedeadpunk lyarwood: yeah, I see that, but libvirt 6.1.0 is smth that seems needs to be built manually I guess
14:17:26 noonedeadpunk lyarwood: yeah, I see that, but libvirt 6.1.0 is smth that seems needs to be built manually I guess
14:17:32 noonedeadpunk as UCA doesn't provide that :(
14:17:32 noonedeadpunk as UCA doesn't provide that :(
14:17:40 lyarwood noonedeadpunk: then raise a bug with the UCA folks
14:17:40 lyarwood noonedeadpunk: then raise a bug with the UCA folks
14:18:01 gibi dansmith: at least it is documented in the code comment above the alias https://github.com/openstack/nova/blob/dab4ec1a534d4774e61a24cdace0f3491788e4e7/nova/objects/service.py#L206 so the intention is clear for the reader
14:18:01 gibi dansmith: at least it is documented in the code comment above the alias https://github.com/openstack/nova/blob/dab4ec1a534d4774e61a24cdace0f3491788e4e7/nova/objects/service.py#L206 so the intention is clear for the reader
14:18:10 noonedeadpunk I believe there were solid reasons not to workaround in nova?
14:18:10 noonedeadpunk I believe there were solid reasons not to workaround in nova?
14:18:27 dansmith gibi: okay :)
14:18:27 dansmith gibi: okay :)
14:18:29 lyarwood noonedeadpunk: yeah it would require us to rebase the disks at launch
14:18:29 lyarwood noonedeadpunk: yeah it would require us to rebase the disks at launch
14:18:38 noonedeadpunk Just thinking if it make sense to look into this direction or not
14:18:38 noonedeadpunk Just thinking if it make sense to look into this direction or not
14:18:41 gibi dansmith: :)

Earlier   Later