QUESTION:
This is an NFS troubleshooting question that manifests itelf through an application.
How can I find/resolve NFS incompatibilites between SCO/RHL NFS client/server.
Recommendations for tools/information that will not take a week to learn. The O’Reilly NFS/NIS book didn’t help much because the protocol analysis section was to dense and assumed protocol experience beyond my current abilities.
PROBLEM DESCRIPTION:
I’m using SCO UnixWare 7.1.0 (a requirement)and an older version of MKS_RCS where the RCS archives reside on an NFS filesystem.
R/W file I/O NFS SCO 7.1.0 <-> SCO 7.1.0 works okay. RCS operations over same work okay.
R/W file I/O NFS SCO 7.1.0 <-> RHL 7.2 works okay.
RCS operations over same fails withRCS error “Can’t obtain lock record”.
The SCO clients appears to be defaulting to NFS V2 and the transport is UDP. Traffic is on the same ethernet segment and the same IP subnet. There are no known networking problems between the two hosts.
DETAILS:
The reported error is probably misleading becuase I suspect RCS is getting an unexpected error return from lower levels and maps it as best it can.
I encounter no other problems reading and writing small or large files across the mixed platform NFS configuration. I am only using NFS, not any type of “RCS server”. The RCS application executes on each users host reading/writing across NFS. I have triple-checked UID/GID’s on both sides, mount and export options (block size, r/w, not using anon UID).
I not a full-time unix or network admin and this has me stumped.
