OEM Reports High Load Average
OEM Reports that a server load is critical and top confirms this-
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
14721 oracle 15 0 4243m 201m 196m S 1.7 0.6 0:43.83 oracle
14998 oracle 15 0 8338m 47m 42m S 1.3 0.1 0:00.44 oracle
14713 oracle 15 0 4248m 2.8g 2.8g S 0.7 8.8 0:53.50 oracle
14715 oracle 15 0 4243m 2.8g 2.8g S 0.7 9.1 0:55.10 oracle
13734 oracle 16 0 4241m 693m 690m D 0.3 2.2 0:02.96 oracle
13744 oracle 15 0 4241m 697m 693m S 0.3 2.2 0:02.97 oracle
Upon inspecting the database, you find that this isn’t due to any queries, but caused by SMON as it performs a parallel recovery after another DBA has killed a transaction that was performing less than optimally.
Captured in the alert log:
Parallel Transaction recovery caught exception 30319 Parallel Transaction recovery caught error 30319 Dead transaction 0x000a.05c.00003e84 recovered by 48 server(s)
to resolve the situation, update the fast_start_parallel_rollback parameter in the database to “low” so smon won’t spawn so many recovery processes and let the recovery complete.