Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: (ASCEND) mrtg and p50's
Stephen Balbach wondered:
> Does anyone have an example mrtg.cfg for monitoring bandwith on a P50
> In particular looking for a way to have 1 graph for the combined 2B
I do not monitor individual P50s, but I know how I would do
this for a P130, and assume that the same approach would
work for a P50.
Trick one is that Pipelines have a annoying habit of playing
3-card monte with your active channels. Therefore, your only
choice in this area is to sum up wan0 through wan10 inclusive,
firm in the knowledge that you will get the data totals, since
channels not presently in use will not increment. While this
is brute force, it get around the shell game.
Trick two is that (at least on the P130) 5.0Ap1 does not show
the wan0 thru wan10 interface entries. Release 5.1A does.
Soon-to-hit-the-street releases are rumored to have SNMP
support that might allow me to change my opinion of Pipeline
SNMP support from "nearly none" to "minimal". (nudge, nudge).
Trick three is to add the wans together. With mrtg, this
is done like this:
Target[PipeBW]: 1st_port:community@PipeLineIP
+ 2nd_port:community@PipeLineIP
+ 3rd_port:community@PipeLineIP
+ 4th_port:community@PipeLineIP
...and so on until...
+ 11th_port:community@PipeLineIP
where:
community = SNMP community string
PipeLineIP = The IP Address of the Pipeline
xth_port = the interface number of wan0 through wan10
Trick four, if you are reading the P50 over the measured
channel, is to subtract from each "read" the bandwidth used
by all the SNMP Gets required to do this. This would be something
like "- 300" on the very end of the lines above, if a series of
reads done on a quiet channel gives you a baseline delta reading
of 300 bytes (I have no idea what the actual figure would be).
This is optional, but it depends on what you want to measure, and
how accurate you want to be.
A ship at dock is certainly safer, but that is not what ships are for.
james fischer jfischer@supercollider.com
++ Ascend Users Mailing List ++
To unsubscribe: send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd: <http://www.nealis.net/ascend/faq>
Follow-Ups: