Parallelism backup in RMAN Orcale 10g RAC

Question

Hi All,
I want to know how parallelism works for RMAN RAC
Can you forward me to some useful links where I can get this information.
One way of doing it is:
CONFIGURE DEVICE TYPE DISK PARALLELISM 2;
CONFIGURE CHANNEL 1 DEVICE TYPE DISK connect 'SYS/rac@node1';
CONFIGURE CHANNEL 2 DEVICE TYPE DISK connect 'SYS/rac@node2';
But in 10.2.0.2 onwards it says that we dont have to specify nodes. Oracle will automatically spread it accross node. is that true.. but doesnt seesm to be working?
node1> rman target /
run {
CONFIGURE RETENTION POLICY TO REDUNDANCY 1;
CONFIGURE BACKUP OPTIMIZATION ON;
CONFIGURE CHANNEL DEVICE TYPE DISK MAXPIECESIZE 8G;
CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO'$RMANOUT/%F';
CONFIGURE DEVICE TYPE DISK BACKUP TYPE TO COMPRESSED BACKUPSET PARALLELISM 2;
CONFIGURE DEVICE TYPE DISK BACKUP TYPE TO COMPRESSED BACKUPSET;
CONFIGURE DEVICE TYPE DISK PARALLELISM 2;
allocate channel c1 device type disk ;
allocate channel c2 device type disk ;
backup incremental level 0
database format '$RMANOUT/%d_%s_%T_%c_DB0' tag="DB_LEV0";
backup current controlfile for standby format
'$RMANOUT/%d_%s_%T_CONTROL_STDBY' tag "CNTRL_STDBY";
sql "ALTER SYSTEM ARCHIVE LOG CURRENT";
backup archivelog from time='sysdate-5/24'
format '$RMANOUT/%d_%s_%T_%c_ARCH0' tag "ARCH_LEV0" ;
backup current controlfile format '$RMANOUT/%d_%s_%T_CONTROL' tag "CNTRLFL" ;
backup spfile format '$RMANOUT/%d_%s_%T_SPFILE' tag "SPFILE" ;
will the above script run parallel sessions in RMAN.
I queried gv$session I cant see 2 sessions, I can only see 1.

Answer 1

if you want to enable parallelism you have (so to say) three ways related to channel allocation:
1. in rman settings directly
configure default device type to disk;
configure device type disk parallelism 2;
configure channel c1 device type disk connect 'aaa/bbb@node1'
configure channel c2 device type disk connect 'aaa/bbb@node2'
2. or directly in rman script, something like:
RUN
allocate channel c1 device type disk connect 'XXX/YYY@node1';
allocate channel c1 device type disk connect 'XXX/YYY@node2';
backup database plus archived log delete input;
3. or you can use automatic workload balancing with setting:
configure device type <disk or sbt> parallelism <number of channels>;
and after that issue backup command.
parallelism you may monitor by checking gv$session_longops
also rman have option for 'node affinity awareness' which means that rman during backup determines which nodes in cluster have faster access to certain datafiles than to other datafiles.
pay attention that the granule of parallelization for backupsets is the backupset...inappropriate setting of the parameter "filesperset" can lead backup to be serialized...specially if you use tapes...