I purchased a new Dell PowerVault 124T (LTO-2) and installed it on a new quad Xeon PowerEdge. It's the only device on the Adaptec 39160 bus, but I've eliminated the SCSI bus since I'm getting nominal verify performance (1-3GB/min). I'm using BE 12.5 with SP1 and a hotfix that Dell had me install. I've installed the latest firmware on the drive and autoloader. The server has two local partitions (C:, E:) that are physical SATA II drives. I'm running ESET Nod32 3 and I believe Windows Indexing is turned on for these two drives. The drives have 0 fragmentation according to Diskeeper 2009.
I've run several tests serveral times, and here's what I've come up with on average:
1. Selecting the entire server to backup (system state, etc.) produces the slowest result. About 650MB/min. This is about 50% of the worst case spec (1.44GB/min native, no compression).
2. Selecting E drive (about 24GB) is slightly better, at 735MB/min.
3. Selecting C:\Temp where I created 80 300MB highly compressible files produced varying results. Four tests fluctuated from 1.44GB/min to 4.5GB/min.
This server isn't in production yet, so there were no activities demanding CPU (other than the backup) when these tests were run. Also, backup takes about 1-2% of CPU. The server has 4GB RAM, and I'm not getting excessive paging. For BE I'm using hardware compression, and the default block size is 64KB, buffer 64KB, buffer count=10, high water count=0. Write single block mode and Write SCSI pass-through mode are both checked (all of these values are BE's defaults for this device, which I have unchanged).
Thanks for hanging there so far. My question of course is why am I getting half the worst-case spec in a traditional backup scenario? I expect that even where I'm backing up tons of small files that aren't compressible, I should be getting 1.44GB/min as an absolute minimum. It appears as if the drive is capable, but BE or something affecting the write capability of the drive isn't configured properly?
Thanks for the help, in advance.