Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ucbvax!SLACTWGM.BITNET!ESMP09
From: ESMP09@SLACTWGM.BITNET (Ed Miller SLAC x3291 or [415]854-1055)
Newsgroups: comp.os.vms
Subject: Re: Repairing disks with corrupted index files
Message-ID: <8801132356.AA02036@ucbvax.Berkeley.EDU>
Date: 12 Jan 88 19:54:00 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 46

>  We have an eagle disk with a SI 9900 controller which has had it's index file
> corrupted 4 times in the last 4 months. Has anybody else experienced these
> problems. Is it a software or hardware problem.?
>  The error message I get when I try and mount the disk is
> BITMAPERR- I/O error on storage bitmap; volume locked
>  Now the system error message manual implies you can repair this by
> using the verify utility but the catch 22 is that you have to mount
> the volume files-11 ... and the mount won't let you do this.
>  Is there a way of repairing a disk with a bad index file ?

We had a similar problem several months ago with disks on an
SI9900 controller.  The problem happened twice within a few
days (once on a 9751 disk, once on a 9798 disk), but has not
recurred since.  There was no evidence that it was software related
(we'd been running VMS 4.5 for months before, and still are running
it).  There was no obvious connection to hardware changes, but there may
have been some upgrades of CPU (not disk) interfaces in the controller
since the two occurences.

Our problem took the following form:  when we tried to MOUNT a
disk, MOUNT complained that it was the member of a shadow set, and
proceeded to mount it, but with a software writelock.  (We don't
use shadow volumes, so that was the first puzzle.)  It turns out
that the indication that a disk is a member of a shadow set is
stored in the first block of BITMAP.SYS--when we dumped that file
it was obvious that it had been overwritten with irrelevant data--
not only the first block, but the first few blocks.

For our situation, the fix was easy:

        MOUNT/OVERRIDE=SHADOW
        ANAL/DISK/REPAIR

(There was a lot of repair to be done, since the first few blocks of
BITMAP.SYS needed to be reconstructed, but there were no damage that
could not be repaired.)

If your problem is similar, you might be able to make the same
kind of fix with

        MOUNT/OVERRIDE=LOCK
        ANAL/DISK/REPAIR

                                Ed Miller
                                ESMP09@SLACTWGM.BITNET
                                Stanford Linear Accelerator Center