I asked:
> Does anyone have a HIGH performance tape-to-tape copy utility? I tried Sun's
> tcopy, and only got ~35 Kbytes/second copying from a dat drive ( 184
> kbytes/second) to a High density Exabyte (500 Kbytes/second). Right now, it
> takes 15 hours to copy a 1.2 Gig tape, which is just too long! Naturally, I'd
> like the copy program to produce an exact copy, including record block sizes
> and file marks. I've also tried tapecv but it wasn't any faster.
>
> I think that by doing some async. I/O, and reading/writing groups of records
> that it should be possible to get much higher performance.
bukys@cs.rochester.edu suggested taper, from
cayuga.cs.rochester.edu (192.5.53.209), which is usable locally or over a
network by using rsh.
taper rf /dev/dat0 | taper wf /dev/smt0
stanonik@nprdc.navy.mil suggested "copytape", which he uses to copy Sun
distribution tapes, which is a stress test for most tape copy programs.
A version of copytape is available from titan.rice.edu sun-source/copytape.c
If I knew the format of the tapes in advance, I could use
'dd',
"buffer" (a tape buffering utility ) from celit!hutch@UCSD.EDU
"block" (somewhere on comp.sources)
"ddd" a two process version of dd available from
ccb.ucsf.edu Location: /Comp.sources.unix/Volume15
FILE -rw-r--r-- 17915 Sep 20 1988 ddd
Glen Dudek (dudek@ksr.com) supplied me with tpcp, which does reads and writes
in sep. processes to speed up I/O, and preserves file marks and blocksizes.
I haven't had time to benchmark these tools yet, but I certainly appreciate
all of your responses!
Thanks again. I'll be happy to make any of these utilities that I now have
available to anyone who requests them.
Joe VanAndel Internet:vanandel@ncar.ucar.edu
NCAR - ATD/RSF
P.O Box 3000 Fax: 303-497-2044
Boulder, CO 80307-3000 Voice: 303-497-2071
Status: OR
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:19 CDT