Unanswered: db2 9.7 on Linux Redirect restore partitioned database
I am restoring a big (3.6TB) partitioned database on a different machine (hardware is the same). I have backed the source database with a single system view backup , generated redirect restore scripts for each partition (4 partitions = 4 scripts) and started the restore. I first restored the catalog partition, then all the others, one by one. The database had 5 automatic storage paths defined and I am redirect-restoring it to 5 AS paths as well.
However during the non-catalog partition restore, I noticed that when the 'skeleton' database is being created (before the actual restore of data), only 4 paths are used, after some work has been done on them - fifth path is being used while other paths are idling. When the fifth has caught up with others, it then idles while other path does some more work and this goes on and on until the build faze ends. This behavior almost doubles the time needed to build the skeleton database and I was wondering why isn't the restore using all 5 AS paths at the same time when creating the 'skeleton' database. During the actual restore of data - all five paths are used at the same time. Is there any way to influence this? All paths (filesystems) are on the FC disks.