Hi All,
I'm sure this is more a SUN issue rather than Compaq, but I'm tasked to
pretty
much prove it isn't a compaq problem.
We have a cluster running 4.0F which is our file server and our server for
NIS
home directories. The Sun client (also a server class) is running Solaris
2.6.
On the compaq cluster I see rpc.lockd running at round 39% of the cpu. The
sun system and compaq system are passing packets continuously accross
the network talking to each other, but we can not find the reason why this
is happening.
If we reboot the Sun system, rpc.lockd on the compaq system returns to
normal to 0% CPU, but once
it is up, it's back up to 39% immediately.
At this time, all our services (2) are running on one node, so I'm not
sure if splitting the services between
the nodes will make rpc.lockd follow over to the other system at this
time. It's hard to schedule system
changes here, but I don't think it's a compaq problem in the first place.
When zeroing out the numbers with nfsstat on the Sun system, we see under
clinet rpc, the badcalls and
timeouts climb fast under Connectionless section.
I guess I'm looking for the reason why this Sun server is causing
rpc.lockd to use a lot of my CPU time
on my Compaq cluster.
Any help would be appreciated.
thanks,
Allan
Received on Thu Feb 07 2002 - 16:56:15 NZDT