Unknown access on node 2 leads to gc wait in RAC
Hi guys,
We've 2 node 11.2.0.3 RAC running on Solaris 10. We created one specific service to run on 1st node only to avoid cache fusion.
Recently, the following SQL suffers from degraded performance due to gc wait:
1sf9fyt9dcm3b update MSC_MSG_SEND set xxxxxxxxx
Top 5 Timed Foreground Events
Event Waits Time(s) Avg wait (ms) % DB time Wait Class
DB CPU 19,175 80.95
gc buffer busy acquire 798,900 4,130 5 17.44 Cluster <<<<<<<<
gc cr multi block request 865,550 2,006 2 8.47 Cluster <<<<<<<<
enq: TX - row lock contention 9,979 185 19 0.78 Application
direct path read 297,395 179 1 0.76 User I/O
SQL history from dba_hist_sqlstat:
INST TIME PLAN_HASH_VALUE EXECUTIONS ELAPSED_TIME_S ELAPSED_TIME_S_1EXEC CPU_TIME_S IOWAIT_S CLWAIT_S APWAIT_S CCWAIT_S ROWS_PROCESSED ROWS_PROCESSED_1EXEC BUFFER_GETS BUFFER_GETS_1EXEC DISK_READS DISK_READS_1EXEC DIRECT_WRITES