Tuning Recovery - Expert Oracle RAC Performance Diagnostics and Tuning - page 326

From the preceding output, the current FAST_START_MTTR_TARGET value is set much higher than the default value.

3.

Check if the advisory has updated the OPTIMAL_REDOLOG_FILE SIZE column. Executing the
following query against the V$INSTANCE_RECOVERY view will help understand the current
value for the redo log file size:

SELECT actual_redo_blks ARB,
target_redo_blks TRB,
target_mttr TMTTR,
estimated_mttr EMTTR,
optimal_logfile_size OLS,
ckpt_block_writes CBW
FROM v$instance_recovery;

ARB TRB TMTTR EMTTR OLS CBW
---------- ---------- ---------- ---------- ---------- ----------
2089 7309 15 2 367 1040
Based on the current activity of the database, the suggested MTTR value is 15, and the recommended size for the
redo log file size is 367. Making changes to the database would help reduce the actual redo blocks to be recovered.
These measurements and tuning of the redo log file size are only guidelines to help improve/optimize instance/crash
recovery times. Instance recovery times can vary depending on if the instance crash was due to node failure or
instance failure (without node failure). In the case of a node failure, additional activity, for example, reconfiguring
the group membership of the cluster, has to be completed before instance recovery begins. Once the clusterware
completes the cluster-level reconfiguration, LMON process takes over to complete its activity related to the cluster level.
Oracle records this in the LMON trace file of the instance that performs recovery. (The output following is from a Oracle
Database version 12.1.0 instance recovery and contains more details compared to the prior versions of Oracle).

*** 2014-04-02 22:23:22.077
kjxgmrcfg: Reconfiguration started, type 1
CGS/IMR TIMEOUTS:
CSS recovery timeout = 31 sec (Total CSS waittime = 65)
IMR Reconfig timeout = 75 sec
CGS rcfg timeout = 85 sec
kjxgmcs: Setting state to 4 0.

*** 2014-04-02 22:23:22.239
Name Service frozen
kjxgmcs: Setting state to 4 1.
kjxgrdecidever: No old version members in the cluster
kjxgrssvote: reconfig bitmap chksum 0x19a68269 cnt 1 master 1 ret 0
kjxgrpropmsg: SSMEMI: inst 1 - no disk vote
kjxgrpropmsg: SSVOTE: Master indicates no Disk Voting
2014-04-02 22:23:22.253464 : kjxgrDiskVote: nonblocking method is chosen
kjxgrDiskVote: Only one inst in the cluster - no disk vote
2014-04-02 22:23:22.468535 : kjxgrDiskVote: Obtained RR update lock for sequence 5, RR seq 4
*** 2014-04-02 22:23:22.589
2014-04-02 22:23:22.589726 : kjxgrDiskVote: derive membership from CSS (no disk votes)
2014-04-02 22:23:22.589788 : proposed membership: 1
2014-04-02 22:23:22.833840 : kjxgrDiskVote: new membership is updated by inst 1, seq 6
2014-04-02 22:23:22.833913 : kjxgrDiskVote: bitmap: 1