[daip] help regarding tape read error

[Eric Greisen]

Hendrik Linz writes:

 > After that I tried it without skipping files, but:
 > 
 > SUN48 > UVLOD1: Task UVLOD  (release of 15OCT99) begins
 > SUN48 > UVLOD1: Reading tape drive number   1
 > SUN48 > UVLOD1: ERROR      5 READING FIRST TAPE RECORD
 > SUN48 > UVLOD1: Purports to die of UNNATURAL causes
 > SUN48 > UVLOD1: sun48        15OCT99 TST: Cpu=       0.0  Real=     488
 > 
 > PRTTP with NFILES = 0
 > 
 > SUN48 > PRTTP1: Task PRTTP  (release of 15OCT99) begins
 > SUN48 > PRTTP1: HEADER BAKF I/O ERROR =      5
 > SUN48 > PRTTP1: Read at least    0.000 Megabytes in       0 logical records
 > SUN48 > PRTTP1: Purports to die of UNNATURAL causes
 > SUN48 > PRTTP1: sun48        15OCT99 TST: Cpu=       0.0  Real=     240
 > 
 > Well, I checked the AIPS installation and the tape drive (DDS4), and this 
 > combination was able to read other AIPS tapes without problems, even 
 > DDS1 tapes (the tape in question is a DDS2). Furthermore, we use here
 > the 15OCT99 AIPS version, but the tape is written in January 1999, so there
 > should be no problems with the AIPS versions ...
 > 
 > Well, I guess the tape "just" got damaged somehow over the year which is 
 > quite sad. I just wanted to know if you can tell from the errors codes
 > what's going on and whether there are possibilities to make it work in
 > AIPS?

     Error 5 is the unix general I/O error generated normally from a
hardware fault.  I suspect that you will find records in system files
showing the fault that occurred (but with little more info than "there
was an error on device ...")  I suspect that there is nothing AIPS can
do with this tape.

     You are using a really old version of aips - there are many nice
things in later versions and it is really easy to install these days.
See the web pages:
           http://www.aoc.nrao.edu/aips
for the aipsletters telling what has changed and instructions about
downloading 31DEC04 and installing it.  I do not think this will help
with your tape problem however.

 > A more general question would be: Is it possible to read AIPS tapes at
 > the UNIX level, say, with special parameter combinations of the "dd" command?
 > (What to use for block size, density etc. ?)

FITS tapes are usually written with full blocking so the records are
28800 bytes long.  dd might be able to read this tape, but I would be
a bit surprised if it did.  Worth a try...

BTW - the NRAO archive is vastly improved making it trivial to fetch
your old data (but requiring re-calibration and a modern version of
AIPS to read the file).

Eric Greisen


 > 
 > Thank you very much in advance.
 > Best regards,
 >                  Hendrik Linz.
 > 
 >