Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Problem reading tar tape between two similar machines

1 view
Skip to first unread message

Bruce Pennypacker

unread,
Dec 13, 2000, 11:11:08 AM12/13/00
to
Hi all,

I have two systems that have identical tape drives on them. Here's the
basic setup:

Home PC: Pentium III 500MHz, running RedHat Linux 7.0. Contains an on-board
Symbois SCSI controller.

Work PC: Pentium II 300 MHz, running RedHat Linux 6.2. Contains an Adaptec
AHA-2940 SCSI controller.

Both PC's have an Archive Python DDS-3 tape drive. If I cat /proc/scsi/scsi
on both systems it shows an identical tape device with the same firmware,
etc. (just located on different scsi devices with different id's):

Home PC:
Host: scsi1 Channel: 00 Id: 04 Lun: 00
Vendor: ARCHIVE Model: Python 04106-XXX Rev: 715G
Type: Sequential-Access ANSI SCSI revision: 02

Work PC:
Host: scsi0 Channel: 00 Id: 06 Lun: 00
Vendor: ARCHIVE Model: Python 04106-XXX Rev: 715G
Type: Sequential-Access ANSI SCSI revision: 02

Both machines also have gnu tar 1.13.17 on them. My problem is that tapes
created on my home machine using tar aren't readable by the work machine,
and vice versa. I can create a tape on my home pc using "tar cvf /dev/st0"
and then immediately extract the files using "tar xvf /dev/st0" without any
problems. I've even gone so far as verifying the results with checksums.

If, however, I take that exact same tape to my work PC and run "tar xvf
/dev/st0" it will start to try to extract the files (I'll see the first
filename get displayed) then after a few minutes I get the following:

tar: /dev/st0: Cannot read: Input/output error
tar: /dev/st0: Cannot read: Input/output error
tar: Too many errors, quitting
tar: Error is not recoverable: exiting now

What gives? Why do these two similar systems have problems reading tapes
from one another? I've tried specifying block sizes, cleaning the drives,
etc. but nothing seems to work....

Help! (banging his head against the wall......)

-Bruce

Dan Shackelford

unread,
Dec 13, 2000, 11:58:25 AM12/13/00
to
Have you tried some obvious things to see if it might be a hardware
problem?
Clean the heads of both tape drives?
Retension?
If you do a complete reformat and make a blank tape from stem to stern
on one tape on one machine, will the other machine still be able to
use it?
If you buy a blank, formatted tape at the store, can both machines
still use it?
If these are SCSI tapes, are they both terminated correctly?

Finally, just parenthetically, earlier this year I upgraded from RH
5.2 to RH 6.2. It broke my tape commands .. MT commands no longer work
for reasons I do not understand now. They worked in 5.2, dont in 6.2.
Maybe a search on RH might help you ... in errata or something.

lih-sin the

unread,
Dec 13, 2000, 12:20:28 PM12/13/00
to

i had somewhat almost a similar problem, that at work i tar an nfs
mounted disk. the tar will try to extract the tar file, but after a few
seconds it failed. i solve the problem by using: tar cBvf /dev/st0
in making the tar file. may be you should try it.

lihsin

John Taylor

unread,
Dec 13, 2000, 10:21:15 PM12/13/00
to
In comp.os.linux.help Bruce Pennypacker <br...@pennypacker.org> wrote:
: Hi all,

: -Bruce


Depending on what you need to back up, you could try using dump or cpio
instead of tar. Another option is to try something commercial, like BRU.

John


--

John Taylor

Reply to:
john
at
giffords dot net

Michael Meissner

unread,
Dec 14, 2000, 8:59:19 AM12/14/00
to
dans...@ix.netcom.com (Dan Shackelford) writes:

> Finally, just parenthetically, earlier this year I upgraded from RH
> 5.2 to RH 6.2. It broke my tape commands .. MT commands no longer work
> for reasons I do not understand now. They worked in 5.2, dont in 6.2.
> Maybe a search on RH might help you ... in errata or something.

(disclaimer -- I'm in the GCC group, not the Linux group, so what I say is
personal opinion).

Do a:

rpm -qf /bin/mt

and if it says something like:

mt-st-0.5-?

instead of:

mt-st-0.5b-7

where the ? is some number, you have a broken mt that can't handle using the -f
option with a command that takes a numeric value (it always uses 0 for the
numeric value). I thought that the bug existed in Red Hat 6.1 and was fixed in
6.2.

--
Michael Meissner, Red Hat, Inc. (GCC group)
PMB 198, 174 Littleton Road #3, Westford, Massachusetts 01886, USA
Work: meis...@redhat.com phone: +1 978-486-9304
Non-work: meis...@spectacle-pond.org fax: +1 978-692-4482

0 new messages