Hi,
we run a Alphaserver1000 with OSF3.2 as NFS server for several different
architectures. A few weeks ago we exported our first filesystem larger
than 2GB (4GB). Meanwhile we find corrupted files on the 4GB fs. It seems
that the corruption comes from NFS writes from Suns running SunOS4.1.1.
There are fragments of mail messages in the middle of fortran and tex files.
It just looks like a blockcount wrap around above 2^31-1.
I know there is nothing in the NFS protocol itself relating to absolute
block numbers, but I'm not sure if there is some connection between
inode numbers and the file handles. And we have > 1M inodes on this filesystem.
Has anybody seen this behavior too, and is there a cure known for it.
Thanx for any help you can give me. This is really becoming a desaster!
\Magnus
Received on Tue Dec 05 1995 - 22:57:12 NZDT