NFS ESTALE on RHEL5/6 (not RHEL4) Client doing mkdir/rmdir against NetApp Server
Issue
- More frequent "Stale NFS file handle" messages on RHEL5/6 (but not RHEL4) when doing some simple NFS file manipulation.
- The following script reproduces the problem, but not reliably:
while true; do rm -rf /proj/something/test_dir/bpetrini mkdir /proj/something/test_dir/bpetrini echo blah > testfile cp testfile /proj/something/test_dir/testfile done
- The following characteristics increase liklihood of reproduction:
- on a high traffic mountpoint (though the immediate parent of the directory I'm manipulating is untouched)
- from a high traffic client (a many-user interactive will repro much more often than an idle grid node)
- How do I use rpcdebug to capture verbose rpc debug logs for troubleshooting nfs problems(see diagnostics section)
Environment
- NFS client
- Red Hat Enterprise Linux (RHEL) 5.5, 5.7, 6.0
- No errors on RHEL4.9
- NFS Server
- NetApp Release 8.0.1P4 7-Mode
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
