search

azimuth-cloud / azimuth-caas-operator

K8s operator to create ansible based clusters using K8s CRDs
Apache License 2.0
1 stars 2 forks source link

Deployed appliances don't always get scheduled deletion #40

Closed jjo93sa closed 1 year ago

jjo93sa commented 1 year ago

Deployed an appliance. Appliance monitoring wasn't available in time, so the Ansible task failed:

 poplar_infra : Wait for Monitoring to become available

Nonetheless, in the Azimuth portal, I can open both the monitoring and the web console, so the appliance is deployed, and seems to be functional. However, when I check the cron schedule, there is nothing present:

ubuntu@azimuth-cl2-seed:~$ kubectl get cronjob -A
No resources found

Ideally, the scheduled deletion would be configured early, such that it can clean-up appliances that fail to fully deploy.

JohnGarbutt commented 1 year ago

Just seen this, at the moment auto delete time is calculated from successful create.

... maybe we want all failed creates to be deleted after 1 hour anyway? Not sure.

JohnGarbutt commented 1 year ago

I have gone for creating the auto delete job as soon as we kick off create now, so its always auto deleted, even if never fully created.

Please test that out, and let me know if it works for you.

jjo93sa commented 1 year ago

Tested and works fine. An appliance failed during early stages of deployment & the cronjob was created.