Hi,
I have a single BE server running all latest patches except 2014 SP1.
I have two SAS based tape autoloaders connected to a dual port SAS card.
One is LTO6 (Quantum) and the other is LTO5 (DELL).
The LTO5 drive runs well. I recently needed a new drive and in our case a LTO6 autoloader was only a little bit more expensive than an LTO5 so I purchased this instead (Future proofing so that when the LTO5 dies we will buy another LTO6).
However the LTO6 drive performance is worse than LTO5. I am measuring this by performing the same backup job to each tape drive (Literally one after the other - I am aware that variations occur due to the different times I am backing up - but this is a persistent pattern no matter when I test).
Initially I attempted to use LTO5 tapes in the LTO6 unit (Because I do not want to be in a situation where the LTO6 drive fails and be unable to restore the data).
I am aware that using the LTO5 tape in the LTO6 drive means I am limited to LTO5 throughput but I am not even getting that.
After speaking to Quantum they suggested trying an LTO6 tape as a test. I have done this and performance is no better in BE. We than ran ITDT tools and the results show the drive is capable of pushing 160 mbps non-compressed and various higher numbers with compression, so it is not an issue with the unit.
Things I have tried....
- Tried amending various settings within BE (Such as block/buffer size etc..)
- Disconnecting the DELL unit (So the Quantum is only device connected to the SAS card)
- As mentioned LTO5 and LTO6 media
- Latest firmware/drivers on SAS card. server, autoloader, tape drives
- Devices show correctly in Device Manager (Autoloader - Unknown device, Tape drives using latest Symantec drivers via TapeInst.exe)
I am backing up VMware VM's using AVVI and SAN Transport (With no GRT). I believe it is highly unlikely that shoe-shining etc. is occuring as throughput using LTO5 is around 8-13GB a min with a verify of 10-16GB a minute.
I have opened several cases with Symantec about this and even though I explain the situation over and over again they are more obsessed with how I am doing the backup!!!
Despite me telling them it does not matter what VM's I select etc. (And have sent job logs and the ITDT output to them). They even suggested backing up a different server (A physical one) to see if the problem is with AVVI. I have tried to tell them that my physical server jobs do not have the throughput to be able to see the problem but they do not understand and eventually I lose the will to live and give up with them!
I can and have literally shown them BE performing the exact same backup job to each drive and the LTO6 always being considerably slower.
Does anyone have any ideas?