[daip] Writing to exabyte tapes from AIPS, using new Redhat Linux

[Jim Ulvestad]

Hi all,

Today, I've tried to write AIPS data to exabyte drives
on three different machines--alfirk, megrez, and mizar.
I tried writing from within AIPS on my own machine 
(coyote) in two cases, and got errors that AIPS told 
me were parity errors; Eric didn't understand where
they were coming from.  Then I loaded the data onto
mizar and tried writing again, getting

mizar > FITTP2: Task FITTP  (release of 31DEC02) begins
mizar > FITTP2: Writing to tape drive number   1
mizar > FITTP2: Writing at current tape position
mizar > FITTP2: Writing beginning-of-write EOFs
mizar > FITTP2: ZTAPE: LUN 31 IO ERROR AT   1 OP OF BEGW/BEGW SYS ERR
CODE 0005
mizar > FITTP2: ZERROR: IN ZTAP2 ERRNO = 5 (Input/output error)
mizar > FITTP2: ERROR     3 BEGINNING WRITES WITH EOFS
mizar > FITTP2: Purports to die of UNNATURAL causes
mizar > FITTP2: mizar        31DEC02 TST: Cpu=       0.0  Real=       0

In all cases, the tapes mount okay, and I've tried killing the
TPMON processes and re-starting them, with no improvement.

Out of sheer frustration, I moved the data to OHSUMI, the old
SGI machine, and tried writing to tape there.  Though I'm
amazed that explorer and ohsumi still run, this seems to be
working fine.  So my conclusion is that there is some 
interaction between AIPS and the new redhat is preventing
(at least) exabyte drives from being usable.

Can someone please look into this before a visiting user
runs into the problem in the last night of trying to
back up his/her data?

Thanks,

Jim