twice now, with the 1103 snap and with the 1123 snap, while under
   heavy load, my SS2 has decided to eat its buffer cache or page pool
   and overwrite the front end of the disk.

   The symptoms start with the loss of shared libraries, or commands
   not working well, and then when I go to reboot, massive (cylinders &&
   cylinders) of stuff blown through the disk.

   I'm doing mfs for tmp, I do have swap. No warning messages. 16 meg
   system (limited) and this has happened when I've been trying to rebuild
   things.

   Anyone seen same?



YES....  I saw something very very simular with a NetBSD/i386 system
recently [1.3_ALPHA as of 11/22].  Lots of paging and the like, X locked
up, system locked up, reboot, and then TONS and TONS of problems with all
of the filesystems on disks which also have a swap partition on it.

Filesystems which should not have been written to in a long time had
trouble, which was very odd.  The /var partition was in such a terrible
condition, that it had to be newfs'ed and loaded from tape.




Brad Spencer - brad@anduin.eldar.org   http://anduin.eldar.org