Customer ugprade their quay from 3.8 to 3.9, and found that after upgrade success, the postgresql 10 PV was removed and all data lost, please help to verify, we have uploaded all the logs into the SFDC case, thanks

Found this KB https://access.redhat.com/solutions/7033589

It is very important that you scale to only 1 pod. Quay 3.9 must several migrations on the schema and running multiple Quay pods will mean starting up multiple migrations as well, which will overlap each other and the migration will crash and you'll be left in an undefined state. So only 1 pod. Until the UI is available.

CU was running two replicas of the quay pod, it may create two migration job at the same time, which may cause some unexpected behavior, so there are two points that we have concern on:
1. If POSTGRES_UPGRADE_RETAIN_BACKUP must be true to run the upgrade or there maybe data lost, then we should enforce it to be true before upgrade
2. If multi quay pods may result in multi migration jobs run simultaneously, then we should correct this logic, isn't it