We're running a network related application on a cluster.
It basically times the round trip of a number of ICMP packets. When doing
this on one machine of the cluster, the round trip times are noticably
different. (about 200ms more).
There is nothing running on the servers except the core system, and this
application.
The machine which gives longer responses changes after reboots. I can only
assume that bringing one machine up first means that it's running something
extra. I can't see what it is, though, and it doesn't seem to affect the
load at all. It's very strange.
Does anyone know what could be causing it? Thanks.
Chewy
----------------------------------
Ecce Potestas Casei
Received on Mon Jan 08 2001 - 16:14:53 NZDT