The original query related to a kernel panic seemingly caused by the NFS
v3 server code in a cluster running Tru64 v5.1B with all current patches.
This problem brought the entire cluster down repeatedly and was difficult
to trace as a result. The clients are Linux boxes (2.2 & 2.4 kernels).
It transpires that the crash is caused by a problem in the AdvFS code
which can be indirectly persuaded to write out-of-bounds by an NFS client.
>From the initial support query on 27th May, HP have just furnished me with
a patch and this should be included in a future patch kit release.
--
John P.
Received on Wed Jul 02 2003 - 10:36:45 NZST