RMAN clone/restore problem

We opened a call with Symantec and sent them full verbose logs from the client and the media server.

It turns out our problem is related to ET3016768. 

We ran a script on the appliance that shows the nbMSDPpoCounts totals.

Our offending images were all in one policy, but with different schedules.

The counts for this policy were many thousands higher than any other policy. See below

The issue is how the appliance "paths" the images, and thousands of backups under the same "path" means the appliance has to traverse the paths to find the images required.

Count   Client/Policy

    ------- ------------------------------------------

      65907 clienta-b/CLIENTA-B-B_ORA_BKUP

       5269 clientb-b/CLIENTB-B_BKUP

       4060 clientc-b/CLIENTC-B_ORA_BKUP

       4019 clientd-b/CLIENTD-B_BKUP

    . . . SNIP . . .

Symantec recommend using more policies to spread the load on the "paths".

We have rehydrated lots of long-term retention backups off the appliance to tape, and reduced the above counts significantly. We have also changed our SLP's to do the same thing moving forward.

The above actions have fixed our optimised duplication / rehydration delays, and timeouts on RMAN restores.

There is currently no fix for this, but may be fixed in 7.6.1.

Thanks for your help.