fsck not fixing root filesys?

[rich]

We have a Sun 3/160 running SunOS 3.4 for which fsck reports an INCORRECT
BLOCK COUNT I=1772 (4 should be 2), and this is on the root filesystem.
Thus, after fsck has fixed the problem, it says to "REBOOT UNIX" (to avoid
writing the old in-memory copy of the superblock info on top of the fixed
disk version, I believe).  However, after rebooting (L1-A on the console
keyboard), the problem still exists.  The only way to get around the
problem is to boot single user, then exit to multi-user.  Since this
machine normally reboots itself once per week (early Monday morning, from
crontab), it went into an infinite loop this morning (fsck "fixes" the
problem, reboots, etc.).  Note that the system message log does not report
any hard disk failures, and the stand/diag program can read the entire
disk without error.

Does anyone have any idea what is happening, or have any suggestions for fixing
this problem?  I have never seen this behavior before.

Rich Baughman		  rich at cfi.com OR ima!cfisun!rich
Price Waterhouse          Waltham, MA  617-899-6500