A “Waiting for Freeze” stall occurs when Rdb or DBMS process fails and the RDMS/DBMS monitor creates a recovery process to roll-back the failed transaction. The freeze lock is a database-wide lock that is necessary to prevent other users from accessing these records until the recovery process can successfully roll-back the failed transaction. There are two ways to minimize the impact of freeze locks: reduce their frequency or reduce their duration. To reduce their frequency, you should avoid killing processes with STOP/ID or the use of idle process killers that perform a DELPRC (delete process) system service call; try using a FORCEX system service call first.