Restore multiplexing of the controlfile

Scenarios:

• restore of the multiplexing of the controlfile after a failure
• additional controlfile multiplexing

Configuration:

• Two diskgroups (1) +DBNAME_DATADG1 for db files and redo logs (Oacle-managed files) (2) +DBNAME_RECODG1 for the flash recovery area
• The controlfile is mirrored/multiplexed in two copies one per diskgroup
• The role of the two copies is not symmetrical: the first one is the primary copy, a loss of of the primary copy will quickly cause a crash of the (cluster) DB
• The loss of a non-primary copy of the controlfile will not crash the DB immediately, but will require (cluster) downtime to fix it

How to restore a failed copy of the controlfile:

• stop all instances
• create pfile='/tmp/pfileDB' from spfile (use this to find the name of the failed controlfile copy)
• start one instance in nomount
• with rman (Example):

set DBID=....
 restore controlfile to '+DBNAME_RECODG1/DBANME/controlfile/current.258.614023949' 
 from '+DBNAME_DATADG1/DBNAME/controlfile/current.256.612033453';

• startup the DB

If you have to add a copy of the controlfile:

• typical scenario, after a controlfile copy failure the DB was reopened with only 1 copy, a maintenance is scheduled to restore full redundancy
• same as above but (Example:)
  • restore controlfile to '+DBNAME_DATADG1' from '+DBNAME_RECODG1/DBANME/controlfile/current.258.614023949'
  • use ASMCMD to find the name of the controlfile copy generated above and edit the spfile accordingly