Tuning Recovery - Expert Oracle RAC Performance Diagnostics and Tuning - page 317

8. At this stage, buffer space for recovery is allocated, and the resources that were identified

in the previous reading of the redo logs are claimed as recovery resources. This is done to
avoid other instances from accessing those resources.

9. A new master node for the cluster is created (assigning a new master node is only

performed if the failed node was the previous master node in the cluster). All GCS shadow
processes are now traversed (redistributed between all of the remaining instances), GCS is
removed from a frozen state, and this completes the reconfiguration process.

■ Note gCS shadows to be traversed could be estimated from V$GES_ENQUEUE . For example,

SELECT COUNT(*) FROM v$ges_enqueue WHERE resource_name1 LIKE '%BL%';

would help determine the blocks to be traversed.

The following extract is from the alert log file of an Oracle Database version 11.2.0.2 recovering instance; it
displays the steps that Oracle has to perform during instance recovery:

Thu Apr 10 09:00:29 2014
Reconfiguration started (old inc 4, new inc 6)
List of instances:
1 (myinst: 1)
Global Resource Directory frozen
* dead instance detected - domain 0 invalid = TRUE
Communication channels reestablished
Master broadcasted resource hash value bitmaps
Non-local Process blocks cleaned out
Thu Apr 10 09:00:29 2014
LMS 0: 1 GCS shadows cancelled, 1 closed, 0 Xw survived
Set master node info
Submitted all remote-enqueue requests
Dwn-cvts replayed, VALBLKs dubious
All grantable enqueues granted
Post SMON to start 1st pass IR
Thu Apr 10 09:00:29 2014
Instance recovery: looking for dead threads
Beginning instance recovery of 1 threads
Submitted all GCS remote-cache requests
Post SMON to start 1st pass IR
Fix write in gcs resources
Reconfiguration complete
parallel recovery started with 2 processes
Started redo scan
Thu Apr 10 09:00:50 2014
Completed redo scan
read 34017 KB redo, 8416 data blocks need recovery
Thu Apr 10 09:01:25 2014
Started redo application at
Thread 2: logseq 4, block 418