No MAC addresses on Cisco with 2.5-2

[TheWitness]

There is a wireless scanning function. However it may not be finished.i

[NBD]

I have tried that one to. It displays the correct number of userports up but not any mac-addresses. I am not after a wireless function, what i want is that it displays the correct number och user ports up. Userports up should never be a bigger value than user ports or the number of physical interfaces. I use MacTrack to make reports on interface usage of the total number of ports in my network.

For now it seems like the value user ports up and mac-addresses are exactly the same value with the get_IOS_dot1dTpFdbEntry_ports function. There can´t be a one to one relationship between mac´s and userports up. Let say that you have a unmanaged layer-2 switch connected to a port in the Cisco 3560 then it´s normal to have several mac´s on a port.

I´m using the latest version from SVN btw.

[TheWitness]

For the wireless device you can have several more macs than ports.

TheWitness

[TheWitness]

You may have to exclude the port ranges that the ap's are connected to if they are not detected as trunk or link ports.

[NBD]

It seems like we missunderstand eachother. And i don´t see how i can explain the problem any better.

But your suggestions will not help me

You may have to exclude the port ranges that the ap's are connected to if they are not detected as trunk or link ports.

Well.. let me put it this way.. i have a few hundred 3560´s and the AP´s are not allways connected to the same interface.

For the wireless device you can have several more macs than ports.

This is not the problem and have never been. If i use the Cisco function to scan a device it displays the number of macs correctly BUT not the number of "User ports Up" , this value are allways exactly the same as the number of macs.

I don´t see how you can have more interfaces up than the total amout of physical interfaces in a device, it doesn´t make sense.

[TheWitness]

I get it now. This is all about "vocabulary" (likely a bad design on my part). There was always assumed to be a 1-1 port to user port mapping. So "User Ports Up" should be "Active Macs". In the Generic Switch Port example, it actually reflects the ifUp count. That might simply be the scanning function too.

I'll have to research it maybe add another column. Since the ifInterfaces table is scanned, we clearly know the number of up Interfaces.

TheWitness

[NBD]

Great.. i was getting rather frustrated when beeing missunderstod all the time

I don´t know anyting about programing but can´t you just copy the the interfaces up function in the get_generic_switch_ports scanner? That one clearly works ok.

[TheWitness]

I'll do something, but you have to open a bug here first. Otherwise, I forget: http://bugs.cacti.net

[georgek]

NBD, whether you have opened Issue? For me a similar problem.

[Mehlano]

Same problem here..
...we have 2950, 2960 and 3750 Cisco Switches. (Using latest MAC-track 2.9)

With the generic port scanner all "user ports up" and "active macs" seems to be ok.
With the IOS template both counts are the same.

For example even bad when you have a VM ESX environment with many virtual clients...

TheWitness, will there be a fix for this issue? Cause when I see the bug tracking ticket it's on "feedback status".



In any case - thank you so much for this fantastic tool and your good work on it! Keep rockin'!

[Mehlano]

In addition...

[Mehlano]

bump

[TheWitness]

To do the VMWare stuff, we have to perform some fundamental design changes to MacTrack since the servers physical port become trunk ports. So, I've given it some thought, just don't know when I can get the time to make the change. I have several other projects and not a network to test on as I do mainly HPC stuff these days.

If you would like to come up with a patch, I will review/consider/commit it though.

TheWitness