tape drive problems using fbackup

[Sven Boden]

Has anyone ever got the following error using fbackup to backup to a DAT tape
and found out what really was wrong:

...
fbackup(3011): WRITE ERROR while writing a checkpoint EOF, at media record XXXX
fbackup(3102): attempting to make this volume salvagable
fbackup(3123): could not read the previous checkpoint record
fbackup(3105): writing 2 EOFs and rewinding the tape
fbackup(3106): please mount a good tape


For a longer explanation and how far I got with my problem:
The tape drive I'm getting the error on is an external "HP SureStore
Dat24e drive" using DDS3 tapes. The drive is attached to an
HP D250/10.20 (+-10Gb of data is put on tape).

When running a nightly fbackup it will stop at some file issuing the
error message above, usually after 1Gb but always at a random file.
When the error started occuring nothing was changed to the system
in the weeks before (no patches, no hardware changes).

When using cpio instead of fbackup I get a general write error also
at seemingly random places.

What I did so far:
* replace the used tapes with new ones => same effect
* clean the tape-drive with new cleaning tapes => same effect
* attach the drive to another HP machine using HP-UX 11 => same effect
* The internal DDS2 tape drive still works correctly.

* I attached the tape drive to a Windows machine and ran
TapeAssure2 (something on the install CD of the tape), it
came back with the message that all electronics, and all other
tests were ok.
* Using "HP Colorado II" backup under Windows the backup just
hangs after a while or comes back with the error message:
"media may have been removed, please check ..."
* I tried using HP Instant Support but could not get it to run:
- When I fill in Belgium as "Country" in contacts it changes it
to Croatia on using submit;
- It does something strange that makes the corporate
firewall go nuts (even while providing the correct proxy, id and
password);
- When using "offline mode" (trying to run without using Internet)
I get the error: "Offline error: you have to be ONLINE...",
after which you can't get to any of the menu's anymore to
change the settings.
- I can't run it at home as I don't have an NT Windows machine,
but a Windows ME.

Questions:
- Is there anybody who had the same problem and solved it?
- Is there anybody out there who ever got "Instant Support" running?

[Thierry Poels_1]

hi Sven,
hmmz it skips Belgium? :(
the cause of the error is most of the times a bad tape. But after your perfect investigations (new tape, other server, other software) I can only say what you're probably already fearing: probably a bad tape drive.
If you have a support contract, log a call at HP, let them know your finding, and they will certainly replace the drive.
good luck,
Thierry.

[boley janowski]

i know that the surestore DAT8/24i have a jumper switch box on the bottom that is configured for specific platforms, im not sure about the externals, but if so you might mention this to surestore when you call them for support if you have it, there are some info under the surestore page about the tape drives if not. anyway sence these jumpers set the platform compatability this might be something to check.

good luck

[A. Clay Stephenson]

I agree it's probably a marginal tape drive but let's not overlook the obvious. Sven, are you absolutely certain that you are terminated
on (and only on) both ends of the SCSI bus?
Is your cable length (including internals) anywhere near the maximum allowed lengths? You do have terminator power somewhere on the bus enabled? Any one of these conditions could cause exactly the flaky behavior you observe.
I once saw an external DAT drive like this that someone had borrowed a terminator. It did just like yours does and I was surprised by how well it actually worked with only 1 terminator on the bus. (It would usually failed after about 1-2GB.) It you haven't swapped terminators, before a bought a new drive I would invest in a terminator.
My 2 cents worth, Clay.