>Number:         25460
>Category:       bin
>Synopsis:       sendmail goes bonkers with "Stale NFS file handle" log messages
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue May 04 16:27:00 UTC 2004
>Closed-Date:
>Last-Modified:
>Originator:     Ed Ravin
>Release:        1.5-20020917
>Organization:
Public Access Networks
>Environment:
NetBSD mailproc1.panix.com 1.5.4_ALPHA NetBSD 1.5.4_ALPHA (PANIX-APPLIANCE) #0:
Wed Mar 10 18:16:31 EST 2004     release-1.5-20020917/src/sys/arch/i386/compile/PANIX-APPLIANCE i386

>Description:
We run sendmail with its mail queue directory on NFS.  Every now and
then, we get gazillions of messages like this:


May  4 12:15:41 logsources@mailproc1/166.84.1.87 sendmail[17544]: i44EReh17448: collect: premature EOM: Stale NFS file handle

May  4 12:15:41 logsources@mailproc1/166.84.1.87 sendmail[17544]: i44EReh17448: collect: premature EOM: Stale NFS file handle

May  4 12:15:41 logsources@mailproc1/166.84.1.87 sendmail[17544]: i44EReh17448: collect: premature EOM: Stale NFS file handle

May  4 12:15:41 logsources@mailproc1/166.84.1.87 sendmail[17544]: i44EReh17448: collect: premature EOM: Stale NFS file handle

The process can't be killed except with -9.  It continues to soak CPU and produce those log messages (filling up our disk at an alarming rate) until we kill it.
>How-To-Repeat:
happens at least once every 2-3 weeks in our environment.  I'll be happy to get a core dump or other diagnostics if you think you need it.
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: