-
Task
-
Resolution: Done
-
Undefined
-
None
-
None
There are more attentions on the gating tests now, we're also trying our best to improve the stabilities of our tests. I've observed many failures are acutally caused by the environment, hence suggest improve the health check before the component tests. Below are suggestions from console team:
1. Ensue the default bootable volumes (eg: DS.png) are available.
2. Ensure the VM can be created on the cluster from templates and instanceType. (issues on webhook_timeout.png, upload the vm-yaml-for-instancetype.txt which can be used to create VM from CLI)
3. Ensure the artifactory server is up, download images from the server is successful.
4. Ensure CDI uploadproxy works, no network issues. (there are network errors while upload images on UI sometimes)
Attach the screenshot from the official test job.
- is cloned by
-
-
- Closed
-
[CNV-52727] Health check for console tests
gouyang1@redhat.com so this part of the setup script is not executed?
https://gitlab.cee.redhat.com/cnv-qe/kubevirt-ui/-/blob/main/.setup?ref_type=heads#L59-72
since the re-importing is handled either by the deployment or by the health check
I don't think that the deployment job or the health check is doing the re-importing if you change the default storage class. If UI setup changes the default SC - it also needs to do the rest of the steps to re-import the images.
(I do think it would save us a lot of time if our deployment jobs will set OCS as default by default 
)
Guohua Ouyang
added a comment - rhn-support-dbasunag Thank you for the quick action on this, feel free to close this one as you already send the patch to fix the bootable volume check. I cloned this ticket to https://issues.redhat.com/browse/CNV-53372, let's continue there.
Guohua Ouyang
added a comment - rhn-support-dbasunag Thank you for the quick action on this, feel free to close this one as you already send the patch to fix the bootable volume check. I cloned this ticket to https://issues.redhat.com/browse/CNV-53372 , let's continue there. Guohua Ouyang
added a comment - Thanks jpeimer@redhat.com for your attention, since the re-importing is handled either by the deployment or by the health check, the console tests acutally didn't call the checkup locally anymore. We don't want to have a duplicate process which adds extra execution time.
Guohua Ouyang
added a comment - Thanks jpeimer@redhat.com for your attention, since the re-importing is handled either by the deployment or by the health check, the console tests acutally didn't call the checkup locally anymore. We don't want to have a duplicate process which adds extra execution time. gouyang1@redhat.com please create a new ticket for the new requirements. I already created a patch for the boot volume check, the moment that PR gets merged, this ticket would be automatically closed.
rhn-support-dbasunag I would prefer to keep it totally isolated from each other, avoid these dependencies between totally different jobs. These cases need to be validated in the Cluster Health Check which is executed before and after each and every job. Makes sense?
Guohua Ouyang
added a comment - rhn-support-dbasunag The console tests use datasources imported by default once the cluster is deployed, I attached the screenshot on 4.18, does it make sense to you to ensure these datasources are ready before the tests?
Guohua Ouyang
added a comment - rhn-support-dbasunag The console tests use datasources imported by default once the cluster is deployed, I attached the screenshot on 4.18, does it make sense to you to ensure these datasources are ready before the tests? gouyang1@redhat.com I did, I am trying to understand what would be the expected behavior here. I would need to know which data sources the console tests relies on.
[cloud-user@ocp-psi-executor-xl ~]$ oc get datasource -A NAMESPACE NAME AGE openshift-virtualization-os-images centos-stream10 5d17h openshift-virtualization-os-images centos-stream8 5d17h openshift-virtualization-os-images centos-stream9 5d17h openshift-virtualization-os-images centos7 5d17h openshift-virtualization-os-images fedora 5d17h openshift-virtualization-os-images rhel10-beta 5d17h openshift-virtualization-os-images rhel7 5d17h openshift-virtualization-os-images rhel8 5d17h openshift-virtualization-os-images rhel9 5d17h openshift-virtualization-os-images win10 5d17h openshift-virtualization-os-images win11 5d17h openshift-virtualization-os-images win2k16 5d17h openshift-virtualization-os-images win2k19 5d17h openshift-virtualization-os-images win2k22 5d17h openshift-virtualization-os-images win2k25 5d17h [cloud-user@ocp-psi-executor-xl ~]$
Do you want me to ensure all these exists or there is specific ones that console test looks for and I can simply look for those? New ones can be added anytime, which would be normal, and old one can be removed as well (e.g. centos 7, 8 etc). If I fail because those are missing, Devops lanes would terminate unnecessarily. I personally feel I should only check this is not empty and console tests should perform cluster sanity to ensure the one they need exists, before beginning their lane.
And one thing to consider: virt, storage and IUO lane changes settings to boot sources, so it might be a good idea to not to schedule console around those lanes. cc: dkeler@redhat.com /lbednar@redhat.com
Guohua Ouyang
added a comment - rhn-support-dbasunag Did you see the screenshot in this ticket, we at least need to address this problem, hold test execution if the datasources are missing.
Guohua Ouyang
added a comment - rhn-support-dbasunag Did you see the screenshot in this ticket, we at least need to address this problem, hold test execution if the datasources are missing. gouyang1@redhat.com /mschatzm@redhat.com I would need to know exactly what you would need. I am assuming you are talking about cluster health check tests that are run prior to any lane executions.
The issue has been closed because the pull request at https://github.com/RedHatQE/cnv-tests/pull/2310 has been merged.