Many thanks to those who took the time to respond. Suprisingly no one
came up with the solution, but my first theory was correct. It appears
that when using automounter to maintain Hierarchical mounts differing
versions of NFS can cause problems. When one nfs server is NFS2 and one
is NFS3 it is advisable to force a NFS2 mount of both filesystems. I
setup the entry in the automounter table to force NFS version 2, -o
ver=2, and we have not had a problem since.
I cannot pinpoint whether this is related to differing versions of nfs,
rpc, or putting a tcp mount on top of a udp mount.
Also, for those who asked. I did verify that this was not related to
NIS+ by turning off the NIS+ entry and using flat files. Problem
persisted until the version change.
Thanks again to all who helped.
On Feb 10, 12:35, Kris Briscoe wrote:
> > Does anyone know if the following scenario will cause nfs mounts to
> > be dropped?
> > /projects/scsb4 (vers2 nfs mount)
> > /projects/scsb4/cmvcdata (vers 3 nfs mount)
> > This puts an nfs vers3 mount on top of an nfs vers2 mount.
> > Question: Would this cause intermittent drops of the nfs ver 3 mount?
> > We never lose the primary mount point, but the hierarch mount point
> > frequently.
> If you mount via automounter, you may possibly be affected by bug#1263311
> (fixed in 2.6).
> Thomas Anders <firstname.lastname@example.org>
> Hahn Meitner Institute Berlin, Germany
-- "Life brings sorrows and joys alike. It is what a man does with them-not what they do to him-that is the true test of his mettle" - Theodore Rosevelt
Kris Briscoe ( email@example.com ) Senior Unix Systems Analyst Motorola Satcom 1.888.612.8136
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:33 CDT