Thanks for the replies all. I've entered a support ticket in again to investigate to see what we can do.
The other issues we see spawn from these slow tape copies, the queue for the Auxillary copies grow so large, they will never effectively finish, which then fills up our disk storage as the Full jobs are retained until they are copied off to tape.
Answering previous replies with the information below.
For context, we're running all current drivers and HBA configuration, along with the same media with Hardware Compression for the Auxillary jobs. In relation to the block size configuration, we're set with the option of "Use Media Type Setting" for the aux copy block sizes.
The read source for the libraries are 12GB SAS disk arrays locally attached to each media agent, which is then connected over 8GB fiber lines to the tape library. Our media agents showing the worst performance all have 7200RPM read intensive SAS drives, we have our newer builds of media agents utilizing read intensive heavy I\O SSD disks, and are not seeing the problem as bad with those, but that is to be expected with SSD vs HDD performance.
The interesting part with the disk performance in Commvault with the HDD disks, we are seeing as poor as 70GB/hr to 200GB/hr for Aux jobs for our media agents with the HDD disks. Where with the same model of disks in BackupExec for it's tape copies, we average 400-500GB/HR. (keeping in mind, the backup exec servers are on similar hardware, just older all around. Same 8GB fiber to the tape library, we were expecting to see the same speeds with Commvault here...)
We're on V11 SP 16, we stay current with Hotfixes and are usually n-1 from major service packs. We actually get better performance with Multiplexing OFF as opposed to turned on.
I will continue to follow up and post any helpful results or findings.
Thanks for the replies all.