Peter,

I think there are a lot of different issues here.

Bad batches of tapes has been a problem in the industry for many years.

Unfortunately manufacturers do not practice what they preach when it comes
to tape handling.

There is absolutely no way that you can tell if a tape has been damaged
before you even use it.

Next time you order some tapes have a look how they are delivered to your
door.

Are they shipped from the plant to the distributors in air-conditioned
containers? Are they shipped to you in bubble wrap?

The manufacturers say not to store them horizontally but nearly every tape
drive installed is horizontal, as are many storage slots in robotic LSMs.

Most modern tapes subsystems will work around tape defects during the write
phase, so the only risk is tape damage between the write and the subsequent
read.

I have worked on mainframe systems with over 1 million tapes writing 700
tapes per hour and rarely have I seen data loss.

On the other hand I have personally seen lots of data lost on Windows and
UNIX platforms but this has always been because of the poor design of the
backup tools.

Think about it, if you were an IT guy who can not restore a tape because
you realized that the backup had not run for 6 months what are you going to
say other than "this tape has an IO error".

I presume that Bank of America's DR test failed. Was it because the tapes
had IO errors, or was it because their offsite lost a few hundred tapes?

Gerard

List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance