Database upgrade fails during tserver restart

Environment:

This issue affects all released versions of YugabyteDB.

Issue:

Database upgrade is stuck with the following error message reported in the tserver logs:

The yb-tserver logs are at /home/yugabyte/tserver/logs/

2022-01-25 17:56:49.017 UTC [23881] FATAL: pre-existing shared memory block (key 5433001, ID 0) is still in use
2022-01-25 17:56:49.017 UTC [23881] HINT: Terminate any old server processes associated with data directory "/data/pg_data".

Resolution:

This issue can be resolved by restarting or rebooting the physical node (i.e. instance from a cloud provider or any infrastructure)

Root Cause:

This issue occurs if the upgrading of the database fails while restart ingthe tserver. This is a known issue  It is being tracked here