Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 27 Aug 2002 22:24:59 -0400
From:      "David E. Cross" <crossd@cs.rpi.edu>
To:        freebsd-hackers@freebsd.org
Subject:   (hopefully) easy NFS question
Message-ID:  <200208280224.g7S2OxO36242@monica.cs.rpi.edu>

next in thread | raw e-mail | index | archive | help
I have an NFS server that recently began to display the following behaviour
(technically the brhaviour is displayed on the clients):
(4.6.2-RELEASE)

mount_nfs -3 -T server:/path /mnt (UDP doesn't exhibit this)
dd bs=64k if=/dev/zero of=/mnt/somefile

I now get about 7 times/second on the client:
NFS server server:/path is not responding, still trying
NFS server server:/path is OK.

NFS server is a Dual P2/400, 512MB ECC cache, exact details can be found
in my microuptime() posts last month (which have been solved with a
new powersupply)  They aren't included here for the sake of brevity, and
I hope they don't matter.

A tcpdump on the client shows seemingly normal request/reponses from the 
server and _lots_ of NFS Null requests from client-->server.  I am guessing
this is the client trying to see if the server is back?  This is on the order
of many times a second:

21:31:09.275806 client.0 > server.nfs: 1448 null (DF)
21:31:09.275900 server.nfsd > client.1015: . ack 91673 win 30280 <nop,nop,timestamp 15510694 88743> (DF)
21:31:09.275993 client.0 > server.nfs: 1448 null (DF)
21:31:09.276008 client.0 > server.nfs: 1448 null (DF)
21:31:09.276025 client.0 > server.nfs: 1448 null (DF)
21:31:09.276137 server.nfsd > client.1015: . ack 94569 win 27384 <nop,nop,timestamp 15510694 88743> (DF)
21:31:09.276409 server.nfs > client.1710096066: reply ok 164 (DF)
21:31:09.276451 client.0 > server.nfs: 1448 null (DF)
21:31:09.276466 client.0 > server.nfs: 1448 null (DF)

Ethernet card in question is a 3c950C in the server and a FXP 100B in the
client (client is also 4.6.2).  I tested another machine with a 3c950C and
the same client with no errors.  I would say hardware but this is too 
predictable with no dataloss, and this is seemingly initiated on the client.

Any ideas on where to look?
-- 
David Cross                               | email: crossd@cs.rpi.edu 
Lab Director                              | Rm: 308 Lally Hall
Rensselaer Polytechnic Institute,         | Ph: 518.276.2860            
Department of Computer Science            | Fax: 518.276.4033
I speak only for myself.                  | WinNT:Linux::Linux:FreeBSD

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200208280224.g7S2OxO36242>