Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2021-05-19
10:37:45 sean-k-mooney gibi: one casue fo this in the past was that during evac in the ci we were previously just stoping the nova compute agent not the neutron l2 agent so wehn we did the port update it would respond
10:37:58 sean-k-mooney e.g. the souce agent woudl say yep its already wired
10:37:58 sean-k-mooney e.g. the souce agent woudl say yep its already wired
10:38:10 gibi sean-k-mooney: I confirmed that the q-agt is dead on the source host during this run
10:38:10 gibi sean-k-mooney: I confirmed that the q-agt is dead on the source host during this run
10:38:12 sean-k-mooney i think we fixed that in our job
10:38:12 sean-k-mooney i think we fixed that in our job
10:38:21 sean-k-mooney ok good
10:38:21 gibi yes, it is fixed in the test
10:38:21 sean-k-mooney ok good
10:38:21 gibi yes, it is fixed in the test
10:38:35 sean-k-mooney am did we also ensure the job is not swapted to ovn
10:38:35 sean-k-mooney am did we also ensure the job is not swapted to ovn
10:38:38 sean-k-mooney i belvie we did
10:38:38 sean-k-mooney i belvie we did
10:38:50 gibi checking...
10:38:50 gibi checking...
10:41:14 sean-k-mooney it looks like ml2/ovs
10:41:14 sean-k-mooney it looks like ml2/ovs
10:41:16 gibi we only held back nova-next on ovs https://review.opendev.org/c/openstack/nova/+/776944 but nova-live-migration moved to OVN as far as I see
10:41:16 gibi we only held back nova-next on ovs https://review.opendev.org/c/openstack/nova/+/776944 but nova-live-migration moved to OVN as far as I see
10:41:47 sean-k-mooney really the build you linked to has teh screen-q-agt.txt
10:41:47 sean-k-mooney really the build you linked to has teh screen-q-agt.txt
10:43:49 sean-k-mooney i guess it has not run since the default change maybe?
10:43:49 sean-k-mooney i guess it has not run since the default change maybe?
10:43:59 gibi hm it run last week
10:43:59 gibi hm it run last week
10:44:09 gibi let me check a more recent run...
10:44:09 sean-k-mooney this was from friday yes
10:44:09 gibi let me check a more recent run...
10:44:09 sean-k-mooney this was from friday yes
10:44:47 gibi hm a todays run also has q-agt.txt
10:44:47 gibi hm a todays run also has q-agt.txt
10:45:13 gibi but I don't see in the zuul config where we set the Q_AGENT option to be openvswitch for this job
10:45:13 gibi but I don't see in the zuul config where we set the Q_AGENT option to be openvswitch for this job
10:45:14 sean-k-mooney yep also looking at it
10:45:14 sean-k-mooney yep also looking at it
10:45:30 sean-k-mooney is this still zullv2?
10:45:30 sean-k-mooney is this still zullv2?
10:46:18 sean-k-mooney no? https://github.com/openstack/nova/blob/master/.zuul.yaml#L53-L87
10:46:18 sean-k-mooney no? https://github.com/openstack/nova/blob/master/.zuul.yaml#L53-L87
10:46:24 sean-k-mooney that looks like it v3
10:46:24 sean-k-mooney that looks like it v3
10:46:58 sean-k-mooney ok well that is a different mistery to look at after
10:46:58 sean-k-mooney ok well that is a different mistery to look at after
10:48:03 sean-k-mooney maybe it has something to do with neutron-trunk: true but that seams odd if it does
10:48:03 sean-k-mooney maybe it has something to do with neutron-trunk: true but that seams odd if it does
10:48:24 gibi yeah, that is the only odd thing in the job config
10:48:24 gibi yeah, that is the only odd thing in the job config
10:48:54 sean-k-mooney do you think the previos live_migation_evetns patch would fix evacuate?
10:48:54 sean-k-mooney do you think the previos live_migation_evetns patch would fix evacuate?
10:49:22 sean-k-mooney we were planning to backport that to train as the race exits basically since netuuron has been a thing
10:49:22 sean-k-mooney we were planning to backport that to train as the race exits basically since netuuron has been a thing
10:50:12 gibi I don't think it fixes the evacuate as we see the error on master _after_ the live migration event patch merged
10:50:12 gibi I don't think it fixes the evacuate as we see the error on master _after_ the live migration event patch merged
10:50:23 gibi but the logic behind the failure can be the same
10:50:23 gibi but the logic behind the failure can be the same
10:50:38 sean-k-mooney gibi: well its disabled by default
10:50:38 sean-k-mooney gibi: well its disabled by default
10:51:12 sean-k-mooney because it need the nova patch to be present to enable it
10:51:12 sean-k-mooney because it need the nova patch to be present to enable it
10:51:19 gibi also it only changes _get_neutron_events_for_live_migration() which is fairly live migration specific
10:51:19 gibi also it only changes _get_neutron_events_for_live_migration() which is fairly live migration specific
10:51:55 gibi ohh you talk about the neutron fixes
10:51:55 gibi ohh you talk about the neutron fixes
10:52:04 sean-k-mooney yes
10:52:11 gibi we can try that
10:52:11 gibi we can try that
10:52:25 gibi as you said we want that to default to true anyhow
10:52:25 gibi as you said we want that to default to true anyhow
10:52:26 sean-k-mooney the nova patch was just so that the neutron fix would not break live migraiton
10:52:27 sean-k-mooney the nova patch was just so that the neutron fix would not break live migraiton
10:52:51 sean-k-mooney and because we needed the nova fix the neutron one defaulted to false for last cycle
10:52:51 sean-k-mooney and because we needed the nova fix the neutron one defaulted to false for last cycle
10:53:14 sean-k-mooney but now it can go to ture since the nova patch is present and backported
10:53:14 sean-k-mooney but now it can go to ture since the nova patch is present and backported
10:54:07 gibi lets switch it to true on master
10:54:07 gibi lets switch it to true on master
10:54:25 sean-k-mooney would you like me to submit a patch to change the default and note the evacuation bug you filed as a related bug? or will i leave that to you
10:54:25 sean-k-mooney would you like me to submit a patch to change the default and note the evacuation bug you filed as a related bug? or will i leave that to you
10:54:58 gibi you have a bit more context on that neutron flag so if you have time please propose the switch
10:54:58 gibi you have a bit more context on that neutron flag so if you have time please propose the switch
10:55:39 sean-k-mooney ok ill go do that shortly i have a docs meeting at the top of the hour so ill likely do it after that
10:55:39 sean-k-mooney ok ill go do that shortly i have a docs meeting at the top of the hour so ill likely do it after that
10:59:01 gibi sean-k-mooney: thanks
11:00:18 gibi sean-k-mooney: it seems the devstack default OVN change has been reverted https://review.opendev.org/c/openstack/devstack/+/791104
11:00:18 gibi sean-k-mooney: it seems the devstack default OVN change has been reverted https://review.opendev.org/c/openstack/devstack/+/791104
11:01:03 sean-k-mooney oh ok that would explain why we are seeing ml2/ovs
11:01:03 sean-k-mooney oh ok that would explain why we are seeing ml2/ovs
11:01:10 sean-k-mooney things exploded?
11:01:10 sean-k-mooney things exploded?
11:01:25 gibi based on the revert commit message yes there are extra things to fix
11:01:25 gibi based on the revert commit message yes there are extra things to fix
11:01:47 sean-k-mooney im really happy we did not make the default change during feature freeze then :)
11:01:48 sean-k-mooney im really happy we did not make the default change during feature freeze then :)
11:02:35 gibi yes :)
11:02:35 gibi yes :)
11:02:40 gibi good judgement
11:02:40 gibi good judgement
11:23:29 bauzas * bauzas goes afk for the afternoon if you look at me (getting a new electric car ;) 

Earlier   Later