On Tue, 18 Sep 2007, Ask Bjørn Hansen wrote:
On Sep 18, 2007, at 6:58 PM, Jeffrey Goldberg wrote:
For a while I used the scripts that someone already mentioned, but
I'd like to point out something that I haven't seen in other
messages. Monitoring consumes far more resources than the NTP
service itself. So monitor to satisfy curiosity and to get a feel
for what's going on. But after a while, you may wish to discontinue
monitoring at all.
One "cheap" way to do the monitoring is to just monitor the bandwidth
on the switch. That's what I did when I was experimenting with a
Soekris 4501 and wanted to get all the CPU for ntp traffic.
If the box isn't doing anything else (or has a separate interface for
just ntp) then that works pretty well. The NTP packets are
predictable in size, so it's easy to convert kilobytes to requests.
I also use the iptraf program to get a quick peek at traffic on port udp
123. I am not sure if you can log to a file although in order to gather
stats. I use it more to have a look in real time at what's going on.
$ iptraf
Statistical breakdowns -> By TCP/UDP port -> eth0
it gives an average of the current traffic on port 123 (and others if you
wish) :
Protocol data rates (kbits/s): 1.00 in 1.00 out 2.00 total
Louis
http://blogtech.oc9.com
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
