opendevreview | Merged openstack/watcher-tempest-plugin master: Add metrics for RAM to gnocchi and prometheus injected data https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/949773 | 07:18 |
---|---|---|
opendevreview | David proposed openstack/watcher-tempest-plugin master: Add tests for workload_balance with injected data https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/949722 | 07:38 |
opendevreview | David proposed openstack/watcher-tempest-plugin master: Create 2 instances per host on test_execute_workload_stabilization_strategy to avoid creating empty actionplans https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950969 | 07:52 |
amoralej | dviroel, wrt https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950389 , I also proposed updating version of nova in the watcher jobs in https://review.opendev.org/c/openstack/watcher/+/950387 but i understand from the comments there to implement it in the plugin itself | 09:17 |
amoralej | and honestly, I liked the idea of decoupling the implementation of that from the jobs configuration in the different deployment tools that may be configuring tempest | 09:17 |
amoralej | although, I'm fine with updating the min_version to something more recent and in the scope of the current supportability of the plugin | 09:17 |
sean-k-mooney | amoralej: the way microversion are ment to work is you are ment to set it on every rest call to the version required by the request | 10:38 |
sean-k-mooney | the min and max version are more for deteminign if a cloud can suppprot the test then contoleing what version tehy use | 10:38 |
sean-k-mooney | this si somehting that the watcher project did incorrectly in both watcher and the tempest plugin | 10:39 |
sean-k-mooney | we shoudl not have https://docs.openstack.org/watcher/latest/configuration/watcher.html#nova_client.api_version as a config option at all. its more reasoable to have https://github.com/openstack/tempest/blob/master/tempest/config.py#L380-L397 | 10:42 |
sean-k-mooney | but they are not for selecting what microverison to use they are for selecting which test to run | 10:43 |
opendevreview | Ronelle Landy proposed openstack/watcher master: Update Workload Balance strategy documentation https://review.opendev.org/c/openstack/watcher/+/951025 | 11:56 |
dviroel | amoralej: ack, continuing our discussion in the tempest patch | 12:02 |
opendevreview | Douglas Viroel proposed openstack/watcher-specs master: Adds spec for extend compute model attributes https://review.opendev.org/c/openstack/watcher-specs/+/947282 | 12:37 |
opendevreview | David proposed openstack/watcher-tempest-plugin master: I have observe that we have a high ratio of failures on this test because the action plan do not contain any "migrate" action. https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950969 | 13:06 |
opendevreview | David proposed openstack/watcher-tempest-plugin master: Update parameters on test_execute_workload_stabilization_strategy https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950969 | 13:18 |
rlandy | reminder: watcher meeting tomorrow - pls add topics to: https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L54 | 13:52 |
rlandy | thank you | 13:52 |
amoralej | dviroel, sean-k-mooney thanks for the clarification on the review, now i understand it better. | 14:12 |
amoralej | But i still have doubts. My understanding is that the parameter https://github.com/openstack/watcher/blob/59757249bbbbb45bead28bfb2899242aea97dfc1/.zuul.yaml#L60 and/or https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950389 will not affect at all to the version of the nova api version that watcher uses | 14:13 |
amoralej | only the version that the tempest plugin uses when calling directly to nova to create instances | 14:14 |
amoralej | so, i see two different topics | 14:14 |
amoralej | tempest may use 2.74 when creating an instance, while watcher keep using 2.56 when doing api calls to nova | 14:14 |
amoralej | although, it may be a good practice to be consistent | 14:15 |
amoralej | so, getting rid of the https://docs.openstack.org/watcher/latest/configuration/watcher.html#nova_client.api_version and replacing it by a better approach as unrelated to changing the api version that the tempest plugin itself uses | 14:16 |
sean-k-mooney | amoralej: yes they are two diffent topic. related but different | 14:28 |
sean-k-mooney | amoralej: the tempest plugign shoudl ideally explcitly use 2.57 to specify the host when creating the vm | 14:29 |
amoralej | 2.74, right? | 14:29 |
sean-k-mooney | and the test shoudl be annotated with a min nova microvsion fo 2.57 unless you implemnt a fallback for the test to the older way to specify the host | 14:29 |
sean-k-mooney | am well which ever the correct one is fro --host | 14:30 |
amoralej | yeah, got it | 14:31 |
amoralej | i hope, that was i was doing with my patch | 14:31 |
amoralej | at the nova client creation time inside each test | 14:31 |
amoralej | not per api call but per client creation | 14:31 |
amoralej | wrt fallback for older way, i'd say it's not worthy, tbh | 14:32 |
amoralej | I took the idea about implementating fixin the api version from https://docs.openstack.org/tempest/16.1.0/microversion_testing.html | 14:35 |
amoralej | maybe it's wrong way to do it | 14:35 |
dviroel | amoralej: right, when you define the tempest min compute version, it only affects the resources created by tempest, but watcher still keep using its minimal. I missed that detail, that we don't change the minimal in watcher side | 14:55 |
dviroel | we will still testing watcher on 2.56 in this way, but it was confuding for me when I looked at the first time | 14:56 |
dviroel | not sure if makes sense to create the workloads with 2.74 while watcher is always using 2.56 | 14:57 |
amoralej | yes, i was missing we set specific minimal version of nova api in watcher :) | 14:58 |
amoralej | actually, in the "real world" out of openstack, the version used to create workloads it's independent of the version that watcher uses | 14:59 |
amoralej | s/out of openstack/out of tempest tests/ | 14:59 |
dviroel | i think that sean was mentioning that is still possible to select the host in 2.56, using the AZ | 14:59 |
dviroel | amoralej: ack, older clients should be able to retrieve info in the same way | 15:00 |
amoralej | i can investigate the AZ option, if you think that's a better approach | 15:02 |
opendevreview | Merged openstack/grian-ui master: Add basic docs and runserver command https://review.opendev.org/c/openstack/grian-ui/+/950472 | 15:29 |
opendevreview | Ronelle Landy proposed openstack/watcher master: Update Host Maintenance strategy documentation https://review.opendev.org/c/openstack/watcher/+/951002 | 16:02 |
opendevreview | David proposed openstack/watcher-tempest-plugin master: Update parameters on test_execute_workload_stabilization_strategy https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/950969 | 16:03 |
opendevreview | Joan Gilabert proposed openstack/watcher master: Handle missing dst_pool parameter in zone_migration https://review.opendev.org/c/openstack/watcher/+/950149 | 16:18 |
opendevreview | Ronelle Landy proposed openstack/watcher master: Update Host Maintenance strategy documentation https://review.opendev.org/c/openstack/watcher/+/951002 | 18:59 |
-opendevstatus- NOTICE: The Gerrit service on review.opendev.org is temporarily unreachable due to an ongoing issue in the hosting provider where it resides | 22:08 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!