>-----Original Message-----
>From: Peter Ford [mailto:PFord AT STENTOR DOT COM]
>>As for reliability. That turns out to be a very mixed bag.>
><...snip...>
>>For instance,
>>we have a site with a large fiber channel and LTO
>>configuration. No end to
>>the problems so far and they are very serious problems. Is
>>this a result of
>>the tape technology? I doubt it, but one never knows, do one?\
From: Kelly J. Lipp [mailto:lipp AT STORSOL DOT COM]
>I would be very curious to hear what type of reliability
>problems you have seen with LTO. I have posted here before,
>but we have been experiencing an incredibly high number of
>read errors with our 3584 LTO library. We regularly see
>errors when trying to restore data from tapes. We have been
>auditing volumes recently and have seen errors on a tape
>during one audit, and then audit again, with no errors.
>There is no discernable pattern to these errors (across
>multiple tapes and multiple drives). Due to the nature of
>the data we are backing up, the data does not change often
>(and therefore the tapes are generally written to once, and
>the data stays there), so over-used tapes should not be an issue.
>
>Anything that you could share with the list, or me directly,
>would be greatly appreciated.
A problem we've had several times up here in the Great Dry North this
winter has been an environmental one. The 3583 library is fairly
vulnerable to a lack of humidity. While the docs say that 20% is the
minimum required for proper operation, we've found that 40% is really
the minimum needed, particularly in server rooms that are not really
equipped as server rooms; i.e., carpet on the floor, no raised floor,
lots of foot traffic, etc. If you can scoot your feet around, touch the
outside of the library cabinet, and get **zapped**, you've got a
problem. (Your server room should be at 40% in any event; tape 'floats'
best across tape heads at that humidity.)
IBM found a workaround for the lack of humidity. At two sites I've been
to, they've taken one of those yellow-and-green grounding straps they
use to ground mainframe boxes, and attached the library's outside panel
to a decent ground. One customer went from multiple, daily, severe
problems to no problems at all in one day. (Many thanks go to Bryan
Hanson, IBM tape Top Gun, for the fix.)
The problem is that the 3583 is a complex machine that combines many
moving mechanical parts and electronics in a relatively small metal box.
When you send a charge through the box, its relatively small surface
area allows a substantial charge through the box, rather than
dissipating it across its surface. Larger libraries (like the 3584) can
dissipate the charge faster and are therefore less vulnerable.
Don't trust those Wal-Mart temperature/humidity meters. If you're having
3583 problems that can't seem to get fixed, and your environment looks
like the one I've described above, get a good meter and check your
server room.
--
Mark Stapleton (mark.stapleton AT berbee DOT com)
|
