Hi Michael,
can you tell us all the parameters you have configured in snmp configuration?
Did the problem start after you have configured the routers in snmp
configuration or after configuring them in snmpCollect? Probably it's due to a
configuration poll with a large number of routing entries in the routers which
occures at the same time as the polling of the interfaces. If the event
reoccures in about the interval defined in "Configuration Polling Interval"
it's that. You can reduce the "number of route entries" in snmp configuration
to a value of 100 or even 50 if you have large routing tables and high cpu
usage on the routers. I never experienced any trouble with this value
configured.
Anyway, I think 10 seconds is a bit too often. Have you calculated the amount
of disk space you will need? If your /usr/OV/databases/snmpCollect is in the
same volume as /usr/OV/databases/openview you may soon get a lot of trouble
with your databases. Once they were "out of disk space" they most probably will
be corrupted, so be careful! And what do you do with all those values? xnmgraph
will get trouble displaying them after a few days, showing only mean values for
older data, so you might as well collect at a lower rate.
1 Minute may be a good choice for the data collecting interval. If I am
tracking down a problem I use xnmgraph with a 10 second interval, the rest of
time I use 1 or even 5 minute intervals.
Hope this helps
Michael Seibold
GEK Gmünder Ersatzkasse
>>> michael.hamm@tudor.lu 04.01.2001 10.51 Uhr >>>
hi,
since I run NetView to view the traffic from all interfaces
from 10 routers and the cpu usage from 10 routers,
we sometimes have some network trouble.
The cpu usage of some router grown up to 99%
and network traffic died. This happen just sometimes
for several hours.
The polling intervall was 10 seconds. Is this to short???
Any other idea???
thanx
_________________________________________________________________________
|