We're using fully patched (via Live Update) Backup Exec 2014 on fully patched Windows Server 2008 R2.
Backup Exec is installed on a physical server, and is integrated into our VMWare (Vsphere 5.0) infrastructure using SAN transport, and has R/W to all LUNs presented on the SAN.
We have 4 datastores, all VMFS 5, the largest being about 5.6TB.
We use Sophos AV, which has exclusions for the B2D folders and the catalog.
I'm trying to back up a file server running Windows Server 2008 R2 with GRT. Historically, this server has backed up (to disk) at over 3000MB/min for a full backup. (It has the VMWare VSS writer removed by the way).
I have performed a datastore migration to consolidate all of the VMDK files for this host onto one datastore.
What I'm seeing is that the job can start running at over 5000MB/min, the first couple of 100's of GB come down to the backup server rapidly, but consistently while backing up the 2nd VMDK file , about 340GB into the job (which is of a size of 1.8TB) the job transfer rate starts to drop. By the time it gets to about 500GB it is at a crawl, claiming a rate of about 800MB/min. It's now 10.5 hours into the job and has transferred 537GB. All VMDK files are in the same VMFS folder.
I've set the job off numerous times over the last few days, and it does this every time consistently.
The SAN has no errors in its log - it is an HP Storageworks MSA 2324i G2 with 2 x MSA70 expansion units.
CBT is enabled on all VMDK files - I've checked that.
All other backups are running at the correct speed.
Any help appreciated.
Thank you.