To find out the current estimate MTTR, we can get the value from MTTR_Advisor or v$instance_recovery.
select recovery_estimated_ios, actual_redo_blks, estimated_mttr from v$instance_recovery;
The value in estimated_mttr represent if the database instance crash now, how long it would take the database instance to recover and get back online.
We can control the number by setting up the FAST_START_MTTR_TARGET.
FAST_START_MTTR_TARGET
By setting up the FAST_START_MTTR_TARGET to the non zero value means we are going to enable the fast-start checkpointing features.
In order to achieve to control the MTTR in the certain number, oracle instance must increase the frequency of the DBWRn to keep the smaller dirty buffer cache. This would also eliminate the bulk write IO activity in the traditional interval base check point .
alter system set FAST_START_MTTR_TARGET=10 scope=both;
Once the FAST_START_MTTR_TARGET is set, the LOG_CHECKPOINT_INTERVAL
. is ignored.
Perfect guide! Especially useful for business data backup solutions
ReplyDelete