For using the „parallel“ parameter during an import (impdp) on a Oracle RAC Cluster you need to prepare your environment.
The „parallel“ parameter works correctly when you do the following:
– mount point were the export dump resides must be available on ALL cluster members
– create a Service on the database for the impdp job
srvctl add service -s impdp_service -d xdb1 -pdb xpdb1 -preferred xdb11,xdb12 -available xdb13
srvctl start service -s impdp_service -d xdb1
– Check that the service is running
srvctl status service -s impdp_service -d xdb1
Now you are ready to use the impdp „parallel“ parameter
Here an example with „cluster=y parallel=6“
impdp system@xpdb1 directory=dump dumpfile=full_%u.dmp schemas=DB1 cluster=y parallel=6 service_name=impdp_service status=180 logfile=imp_xpdb1.log METRICS=Y logtime=all
impdp Log Parameter which are really helpful for analyzing are:
Extract from the Logfile
You see that there are detailed informations about the worker process for example W-1 = Worker 1
W-1 Completed by worker 1 757 TABLE objects in 38 seconds
W-1 Completed by worker 2 764 TABLE objects in 37 seconds
W-1 Completed by worker 3 765 TABLE objects in 48 seconds
W-1 Completed by worker 4 765 TABLE objects in 53 seconds
W-1 Completed by worker 5 766 TABLE objects in 34 seconds
W-1 Completed by worker 6 765 TABLE objects in 44 seconds
W-5 Processing object type DATABASE_EXPORT/SCHEMA/TABLE/TABLE_DATA
Worker 5 is processing TABLE_DATA
For analyzing the impdp process you get so detailed informations try the next time.
Depending on your hardware you can also use different integer values for the „parallel“ parameter but a large number will not help in every situation.
Have fun with impdp on your RAC Cluster….