Background:
Our machine is maxed out at 100% CPU with the Netview task nvserverd using 50%
of the 2 way J50 and a product called Service Centre using the other 50% during
primetime (8am-5pm). The machine appears to have enough memory (1GB) and the
disk I/O appears light.
Problem:
Quite often it is reported by our OPS staff that they do not see critical
events in the control desk.. This is not consistent as most of the time it
seems to work??
We have a trap setup for "Netview6000" under event IBM_NVNDWN_EV that runs a
script that simply pings the router 5 times before doing a ovxbeep to the
operators terminal. It appears that script is getting executed for the routers
that they say are not showing up in the control desk because we have a log from
this script of all routers that go down..This kind of proves that Netview knew
about the router down condition..
Does anyone know why the control desk would not have some of these critical
events? Perhaps with the machine at 100% cpu could it affect these events???
Thanks,
Rob
|