Out of Memory Errors¶
Make sure you have enough resources allocated to
- your namespace
- your pod
- your spark application if it is a spark one
Delete a namespace Not working¶
When you want to delete a namespace, it is possible to be blocked in the state
Terminating, this issue can appear when you have some Custom resources that can not be deleted due to some other dependency.
To solve this you can follow steps in this ticket.
Drop all connections to a specific database¶
For Applications that use a database, a deployment can failed because there are data already in the database. To solve the deployment you need to drop connections and the database too.
So you need to run an exec on the Pod.
kubectl exec -it kast-default-postgresql-0 -n sql-store -c postgres -- sh
select pg_terminate_backend(pg_stat_activity.pid) from pg_stat_activity where datname='db';