tape read errors: how to read past? recover?

M

monkeyboy

I have some data/backup tapes written with TAR, on a DLT1 (similar to DLT7000) drive, on an ahc0 (AHA2940W) controller.

Some of these tapes now throw errors when I try to read them. Is there any decent way to ignore those errors and read whatever good data must exist beyond the error point?

Right now, once the drive throws the error, the tape driver simply gives up and does not allow any further I/O. So commands such as:
dd if=/dev/sa0 conv=noerror,sync

don't work.

There must be some way to recover most of the contents of the tape, no?

Thanks...
 
Sounds like you are having some kind of drive hardware or SCSI termination problems. Do you see any errors in the message log?
 
Oh tons of error messages, but remember that at least 60-70% of the tapes are readable without any errors, all 40GB of each. So I kinda of doubt its an SCSI termination error. The problem is going to be somewhere between the signal from those 30% of the tapes that throw errors and the drive itself. Oh and another tidbit, I have 3-4 of these drives, and they all behave similarly, though some drives seem to succeed in reading the marginal tapes better than others.

Here are some error messages (but my main question is how to get a drive to read PAST an error and not just quit):

dlt1errs.gif
 
This looks like the same problem I was having with my DLT drive/library and it was a SCSI cable problem. I was able to write to the tapes, but would get read errors when I tried to restore files.

This problem came months after I had setup the library and it took me a while to believe that it was a cable problem. I had a spare cable and kept moving it down the chain till I was able to read the tapes all the way thru.

Are all the drive in one enclosure or do you have multiple enclosures with cables running between them?

Sorry I didn't think about this when I did my first post.
 
all drives (DLT and a couple of SCSI hard drives) in one enclosure, in one system.

another DLT drive, just by itself in the same box as the CPU and IDE drives.

a 3rd setup is like the 1st.

All exhibit this essentially same behavior, which makes me think it is probably a tape aging/printthrough issue.

but I would have thought that via the right software, it should be possible to read past an error and try to recover the rest of the data...
 
I think that the reason you can't get past it is do to the fact that you are having hardware error (SCSI) and not tape read errors.
 
You must log in or register to reply here.
Back
Top