Very Interesting.

I have a 10-drive  3584 library that came with 380 tapes about 22 months
ago. So far, two tapes have locked in drives (which seems to have been a
firmware issue corrected quite some time ago) and one of the tapes was
flawed enough that the drive would die while trying to label the tape.

No other errors in the 22 months. We haven't even clocked any usage on the
in-library cleaning tape.

I push about 1.2 TB of data into the library daily, copy about 1.0 TB for
off-site on a maximum 21-day rotation. Some of my tapes (SAP/Oracle) hit as
high as 517 GB before going full, while others (MS-Exchange) tend to fill at
101 GB (Oracle is mostly empty, the MS-Exchange db is pre-compressed).

My tapes are all 'IBM', actually OEM'd by Fuji as I understand.

Based on the above -- I'd suggest getting a couple of Fuji tapes and seeing
if you run into problems with them.

Tom Kauffman
NIBCO, Inc


-----Original Message-----
From: Jozef Zatko [mailto:zatko AT LOGIN DOT SK]
Sent: Monday, June 16, 2003 7:29 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Interesting LTO fault's symptom


Hi all,
we are using 3583 LTO library with two LTO drives for 5 or 6 months. No 
error on tapes until now.
Interesting findings. If some error occur with tapes I will check more 
detaily.

Ing. Jozef Zatko
Login a.s.
Dlha 2, Stupava
tel.: (421) (2) 60252618



"Tomáš Hrouda Ing." <throuda AT HTD DOT CZ> 
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
16.06.2003 14:14
Please respond to
"ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>


To
ADSM-L AT VM.MARIST DOT EDU
cc

Subject
Interesting LTO fault's symptom






Hi all,

during last weeks I did interesting findings at one of production LTO
3583-L18 library. There are 2 drives and both was changed past series of
media faults (one of them twice) past about 1 year of operation. We have
about 20 "historicaly touched" tapes with average 3-4 write faults. Media
faults are still repeated and my finding is all that faults were done at
70-75% of estimated capacity (set by longterm using to 105GB, we use 
client
compression) during filling the tape. It seems like all tapes were 
corrupted
nearly at the same place, of course there is some diffusion because this 
is
only estimated filling. Faults at these tapes are repeatedly occured at
these percents of max capacity.

I understand when one tape has media fault repeatedly at the same place, 
but
about 20 tapes? Could it mean that all tapes were corrupted by one bad 
drive
at the same place, or can be reason at microcode?

We are in contact with our IBM support to solve it, but I am interested if
anybody of you register similar phenomenon?

Tomas