NEW Cisco Aironet - graphing wireless AP performance - v1.1

[niobe]

Version 1.2 (20090909)

• Removed tcl script header to work better under Windows
• Script must now be called with 'tclsh'

Version 1.1

• Fixed error caused by floating point value in rate_avg (rate_avg not currently used)
• New debug mode - use "-d" option for troubleshooting
• New error behaviour - without debug mode the default behaviour is to return 0 for all values, but not return an actual error.

---------

Here are a couple of new graph templates for Cisco wireless APs:

1) Cisco Dot11 Associated Clients - graphs the number of associated clients on an access point. Includes an HRULE to indicate your max-associations if you are limiting number of associations.

2) Cisco Dot11 Client Rate Distribution - using the scripts queries the AP for the offered rate of each client, and draws an area/stack graph. This is a very useful way to see how well your APs are placed, since the access rate is directly proportional to the strength and quality of the signal.
Since it is an area/stack graph, this, in a way, duplicates the associated clients graph.

---------

Requirements/Caveats:

- You must have TCL installed on your cacti server (TCL is a perl-like scripting language but easier to use Linux users will have no problem with this, Windows users can download from ActiveState.
- Currently supports 802.11b/g rates (1Mbps - 54Mbps)
- Tested with Cisco 1210 and IOS 12.3, but should work on any Cisco AP running recent IOS.
- Tested on Cacti 0.87a/b only

Manually run the script first to see if TCL is working, then import the templates:

Code: Select all

you@cacti# tclsh ./group_dot11_client_rate.tcl <host> <community> <version>

The script returns all zeros when there is an error so it doesn't impact the cacti poller. If you see all zeros and want to know if there was an error use '-d'

Code: Select all

you@cacti# tclsh ./group_dot11_client_rate.tcl -d <host> <community> <version>

[Igloo]

I'm having some issues getting this to work, I'm running a 1200 with 12.3 IOS.

It's returning nothing but a snmpwalk returns fine.

Cheers for any help.

[mcliffe]

Does anyone have coresponding templates of this for a Cacti 8.6i installation?

[niobe]

I'm having some issues getting this to work, I'm running a 1200 with 12.3 IOS.

It's returning nothing but a snmpwalk returns fine.

Cheers for any help.

Ok, let's work this out.
- Can you try running the script manually? Use the same 3 arguments you can see in the data input method.
- What type of 1200 do you have ( 1210,1230,1240,1250 )
- What version of IOS? ( tested on 12.3(8)JE, 12.3(8)JA may not support the required SNMP MIB )

[Igloo]

I'm having some issues getting this to work, I'm running a 1200 with 12.3 IOS.

It's returning nothing but a snmpwalk returns fine.

Cheers for any help.

Ok, let's work this out.
- Can you try running the script manually? Use the same 3 arguments you can see in the data input method.
- What type of 1200 do you have ( 1210,1230,1240,1250 )
- What version of IOS? ( tested on 12.3(8)JE, 12.3(8)JA may not support the required SNMP MIB )

Running the script manualy returns 0

I'm running IOS Version 12.3(2)JA

on a cisco AIR-AP1231G-E-K9

I could try upgrading to 12.3(8)JE

If you can send the IOS?

Regards,

[niobe]

Running the script manualy returns 0

I'm running IOS Version 12.3(2)JA

on a cisco AIR-AP1231G-E-K9

I could try upgrading to 12.3(8)JE

If you can send the IOS?

Regards,

Ok, probably IOS then, I have same APs. Cisco would bust my ass if I posted the IOS, normally you can download them from cisco.com if you have a support contract of any kind (even if it's for another type of device).

[Igloo]

I know Cisco are a royal pain in the ass about theie IOS.

I'll log in and see if i can see it through another device's contract.

Cheers,

[apperrault]

when i try and run the script manually, i am getting an permission denied error. I have the script in the /cacti/scripts folder and it has the same permissions as the rest of the files in that folder (root:root)

Any thoughts as to why this is not working...

thanks much

app

[apperrault]

OK, i got a little further, but it doesn't appear to be working properly in Cacti. I can now run the script from the shell by typing

Code: Select all

tcl group_dot11_client_rate.pub.tcl 10.1.1.5 SNMPSTRING 1

and i get the following output, which appears to be correct

Code: Select all

rate_1_cnt:0 rate_2_cnt:0 rate_5_5_cnt:0 rate_6_cnt:0 rate_9_cnt:0 rate_11_cnt:0 rate_12_cnt:0 rate_18_cnt:0 rate_24_cnt:0 rate_36_cnt:0 rate_48_cnt:1 rate_54_cnt:5 rate_avg:53.00

but I am still getting

Code: Select all

10/14/2008 02:30:20 PM - CMDPHP: Poller[0] Host[238] DS[3400] WARNING: Result from CMD not valid. Partial Result: 

for these three devices.

Does anyone have any thoughts?

app

[niobe]

OK, i got a little further, but it doesn't appear to be working properly in Cacti. I can now run the script from the shell by typing

Code: Select all

tcl group_dot11_client_rate.pub.tcl 10.1.1.5 SNMPSTRING 1

and i get the following output, which appears to be correct

Code: Select all

rate_1_cnt:0 rate_2_cnt:0 rate_5_5_cnt:0 rate_6_cnt:0 rate_9_cnt:0 rate_11_cnt:0 rate_12_cnt:0 rate_18_cnt:0 rate_24_cnt:0 rate_36_cnt:0 rate_48_cnt:1 rate_54_cnt:5 rate_avg:53.00

but I am still getting

Code: Select all

10/14/2008 02:30:20 PM - CMDPHP: Poller[0] Host[238] DS[3400] WARNING: Result from CMD not valid. Partial Result: 

for these three devices.

Does anyone have any thoughts?

app

Hmm is this on windows? I don't prefix the script with 'tcl' as you might for perl or php.

The error may be for the last value 'rate_avg' which is outputting a floating point number. I thought I had fixed this on the code I posted, but will repost the slightly revised code tomorrow. The graph doesn't in fact use this value, so are your graphs working despite the error?

[apperrault]

This is a RHEL 4 box. I will let you know if the new code works when you post it. THis would be a great addition to our current monitoring.

thanks

app

[rojmab]

I found that the Data template was set to look for internal data source names with labels that had a dot11_ before them... ex: 'dot11_rate_11_cnt'
The script returns 'rate_11_cnt' named variables. I went through the data template and removed the dot11_ from each of the variables and it started to graph without issues.

[urban]

I can't use this template with cisco 1240 series.

oid: .1.3.6.1.4.1.9.9.273.1.3.1.1.1

the response from the snmpwalk command:

CISCO-DOT11-ASSOCIATION-MIB::cDot11ClientCurrentTxRateSet = No Such Instance currently exists at this OID

Is this a problem from the aironet series 1240? or a software bug?

I have version Version 12.4(10b)JA on the AP.

[niobe]

I found that the Data template was set to look for internal data source names with labels that had a dot11_ before them... ex: 'dot11_rate_11_cnt'
The script returns 'rate_11_cnt' named variables. I went through the data template and removed the dot11_ from each of the variables and it started to graph without issues.

Not how cacti works. If you look at the data input method, the output variables of the script are defined there, and do not include "dot11_".

There is a disconnect between these and the names in the data and graph templates since the DS are mapped via their index in the sql tables, not their name.

[niobe]

I can't use this template with cisco 1240 series.

I have version Version 12.4(10b)JA on the AP.

Very disappointing but it appears Cisco don't support the specific oid my graphs use in the newer APs. There is no reason for this that I can see except a marketing one. Cisco are pushing their new LWAPP model where you don't configure the AP directly.

Just to confirm here is the error I receive (with v.1.1):

Code: Select all

Error: No valid result received from snmpwalk: .1.3.6.1.4.1.9.9.273.1.3.1.1.1 = No Such Instance currently exists at this OID