Hp Procurve Switch

[piznak]

I've got two switches I'm monitoring with cacti, an hp procurve 2524
and a 4148 which is a 4 module chassi with 2-24 port 10/100 modules
and a 6 port gigabit module.

I've retrieved both switches ok and added them as smtp devices, and also
added graphs of the ports I want to monitor.

My question is about the setup of the actual data source and graphs. It doesnt appear that cacti is getting the real flow of traffice through the switches. As a test I send some 600 meg isos between servers on the ports that cacti is collecting info about. The flow never seems to go above a few hundred k bytes. Just wondering if I need to do anything special for the data conversion etc.

If I disable cacti and run mrtg against the same switches/ports, I see the traffic surge to 30-75 megabytes etc like it should. Note, I don't run mrtg and cacti at the same time.

TIA

Piz

[raX]

How long do these 600 MB ISO's take to copy? Remember that cacti gives a five minute average (for daily graphs), so if the file copy took less than that interval, the spike would be much less. Of course, you could be used to seeing bits/second in MRTG, and cacti defaults to bytes/second. You can change cacti's behavior by using the "Turn bits into bytes" CDEF.

-Ian

[Wade]

Is there any way to convert all graphs to bits from bytes after they have been made? BTW, great tool! I have it monitoring 4 HP Procurves myself. Along with a bunch of Cobalt RaQ's. I love cacti and can't wait for the next version.

[poperszky]

Look in the scripts section of the forum, there is a Perl script for changing the graphs, use the version at the bottom. It changes the labels as well.

Terry

[Wade]

Thanks, I didn't see that last time I looked.