OS : Oracle Linux 6.5
RDBMS version : 11.2.0.4
In our 2-node RAC DB, we have configured the redo log size as 2GB . We see around 7 to 8 switches on peak hours and 3 to 5 switches on off-peak hours. Recently we have see quite a high number "Checkpoint not complete" errors in our alert log.
We don't have any latency on our storage array. Currently , we have 4 redo log groups per thread/instance.
Should I increase the Online Redo Log size or should I increase the number of Online Redo Logs to fix the 'Checkpoint not complete' error ?