Hello,
I have a problem with a number of systems connected via NFS
through a FDDI ring. The systems are Alpha Server 1000's running
DU3.2. The symptoms of the problem are that commands executed
on one of the clients against a disk mounted off the server are taking
a very long time to complete, however there are no NFS errors being
displayed, the command eventually completes OK. Having looked at
the network traffic I can see that the traffic from the server to the
client is arriving rather slowly a packet every second or so. While this
is taking place, the disk from which the data is being read is very busy
indeed. If I repeat this command upon the server itself, the command
completes in the normal expected time, and the disk experiences only
a very short period of activity. (The command in question is "ls").
NFS server running 31 nfsd's, the clients are running 15 nfsiod's. The
disk in question is a 32 GByte RAID system connection through a
Fast/Wide SCSI controller. The cleint machine has >300 MBytes
memory, the NFS server has 128 MBytes memory. The MTU on the
FDDI interfaces has been set to 1500 (so that packets get though
a switch device connected to a network hub).
What is going on? Where do I start looking? As far as I am aware
there have been no changes since well before christmas. The only
recient problem we have experienced is this "date" thing - could this
have caused this problem?
As normal - I'll summarise any responses.
Cheers, Jeff.
-------------------------------------------------------------------------------
Jeff Penfold, Technical Support Specialist, Image and Workflow.
Commercial Union Assurance, 431 Godstone Rd, Whyteleafe, Surrey, CR3 0YQ. UK.
Phone: +44 (0)171 283 7500 x 24795, Email: Jeff.Penfold_at_comunion.demon.co.uk
-------------------------------------------------------------------------------
My opinions, my thoughts, my words. Not necessarily my employers.
-------------------------------------------------------------------------------
Received on Thu Jan 02 1997 - 11:25:33 NZDT