Hello John Schneider

We have had a few problems with cleaning tapes and some that did not
seem to be connected to cleaning tapes.

For now we have been using the stk 9714 built in cleaning support.
We were looking at the wnt adsm software support, but the .30 level
was withdrawn and the .40 level had problems which may be fixed now.
It is also interesting to see if the operators notice when the 9714
reports on it tiny little screen that the cleaning tape has expired.

What is most disturbing is when the a tape is written with no errors
reported then we have a problem reading parts of the tape at a
later time.

I have been trying to learn more about how all this works. But unlike the
old days went one had access to the source and the authors of the source
most things are black boxes today.

From the dlt doc it appears that the cleaning light does not mean that the
dlt drives needs cleaning. But that the drive has a severe read or write
error that might be cleared by the use of a cleaning tape. And a cleaning
tape is just a dlt tape that has not been burnished. If two cleaning
attempts fail then call the CE.

Since an dlt7000 uses around two hours to fill a tape. What happens if the
cleaning indicator comes on at the start of filling operation for an dlt7000.
Will adsm stop using the tape and dismount it. Mount the cleaning tape
and then remount the tape checking that it can read what it had written.
etc.? (They really have some hard work in this area)

As part of this process we have put an request to ibm (via ibmlink) to
look at the TapeAlert standard referenced on the www.dlttape.com web
site. Veritas reports that they use this to enable as-needed cleaning
instead of timed cleaning. It appears that this extension to the scsi
sense protocols also allows for reporting on media and drive problems
at a threshold before they become fatal errors.

IBM has accepted a requirement for TapeAlert. In their words:

ADSM development and marketing are looking at what enhancements will
 be necessary to fully implement this.

I do not know if this is a magic bullet or just another peg in the wall.
But hopefully between all the vendors things will get better.
If anyone from IBM has suggestions, I (and I assume others) are open to
try them out.

Thanks len

-------------------------------------------------------------------------
Leonard Boyle                               snolen AT vm.sas DOT com
Leonard Boyle                               snolen AT vm.sas DOT com
SAS Institute Inc.                          ussas4hs@ibmmail
Room RB448                                  len.boyle AT sas DOT com
1 SAS Campus Drive                        (919) 677-8000 ext 6241
Cary NC 27513