Xref: utzoo comp.protocols.nfs:1857 comp.unix.admin:1117
Newsgroups: comp.protocols.nfs,comp.unix.admin
Path: utzoo!utgpu!watserv1!watmath!att!pacbell.com!tandem!wizard.DSG.Tandem.COM!scott
From: scott@DSG.Tandem.COM (Scott Hazen Mueller)
Subject: rpc.mountd dying
Message-ID: <1991Mar1.173518.13401@tandem.com>
Followup-To: comp.protocols.nfs
Sender: news@tandem.com
Nntp-Posting-Host: work.dsg.tandem.com
Organization: Tandem Computers VLSI, Cupertino CA
Date: Fri, 1 Mar 91 17:35:18 GMT

About once every three weeks, the mountd process on our main NFS file server
dies; more specifically, inetd kills it off, and logs the following message:

inetd[16462]: rpc/udp server failing (looping), service terminated

Has anyone else seen this?  I'm trying to determine if it's a generic mountd or
inetd problem, or if it's specific to the implementation that we're running.
If I had to hazard a guess, I'd say that I'm seeing a burst of requests to
mountd, and that inetd is mistaking a 'normal' momentary overload as a failure
in the mountd process.  Since we run automounters on the clients, it is quite
possible that we would see such request bursts.

-- 
Scott Hazen Mueller             +1 408 285 5762  scott@dsg.tandem.com
Tandem Computers VLSI, Unix System Administrator (apple|sun)!tandem!wizard!scott