Yes, it is a very real concern. But you can figure out when you have
gone too far. First, just do the math. Use ovtopodump -l (I think that's
the option) to figure out how many interfaces you have, and do a
sanity check. Then, do a reality check: use 'netmon -a 3' to have netmon
dump a little status report to /usr/OV/log/netmon.trace. It will have three
parts,
and those three parts can be generated separately with options 11,12,and
13. I look at the second part, and check the numbers down the left
side. Those seem (this is not documented, so I am guessing) to indicate
the time until the next scheduled status check. If it is negative, netmon
is behind. It is not unusual to be behind if you recently stop/started
netmon, but if it never catches up, you have caught your tail. You
can safely erase the netmon.trace file between runs of netmon -a 3.
A new polling cycle will not start before the last one finishes, since the
'next' scheduled poll is not established until this one finishes for each
interface in turn. But the netmon -a 12 will tell you if it missed the time
for
the next poll by showing a negative number (of seconds, I believe).
To check status of snmp polling, use netmon -a 4 (the parts are 15, 16,
17).
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
========================================================================================
Right now my Status Polling under SNMP Configuration is set to 5minutes.
I'm thinking about shortening this. How will I know if I 'catch my tail?'
That is, a new Polling Cycle starts before the old one is compelte? Is
this
something to be concerned about?
Steve (living in Y2K hell) Stamper
Grand Rapids, MI
|