AirWave and Network Management

Reply
New Contributor
Posts: 3
Registered: ‎05-22-2009

Monitoring Aruba APs with cacti

Has anyone had luck in setting up cacti to monitor AP stats through the Aruba controller? I am fairly sure I could monitor each AP individually but that would require a lot of setup and add to the traffic load. I know the Controller monitors this already but it would be convenient to add this to the already existing cacti monitoring system. There seem to be two tables that between them carry all the information I would need to poll wlsxWlanAPTable contains the mapping of APs to IP address, name, loc etc and wlsxWlanAPStatsTable contains all the traffic and user load info. I am assuming that these tables have a one to one correspondence so that the first entry in the APTable corresponds to the first entry in the APStats table. So far I have found no way to "hook" these two tables together to produce an output that has all the info I need and before getting into more esoteric solutions I thought I would see if someone else has already done this.

Specs:
Aruba 2400 OS Version 3.3.2.11 and 40 type 61 APs.
Cacti Version 0.8.6i-3.5 running on Debian Linux 2.6.2-etchnhalf.1, 40G ide hard drive, 2G memory.

TIA
Mark
Occasional Contributor II
Posts: 100
Registered: ‎11-07-2008

Re: Monitoring Aruba APs with cacti

I actually gave up trying to monitor the Arubas in Cacti. I now just use MRTG for APs on the backup controller - in case of a controller or network failure, and users connected to a controller and which SSID they're on. And Splunk for all syslogs which allow me to keep extremely detailed information on users like what time they connected, which access point, where they moved from and to, power for APs, rogue detection, firewall activity... etc. I'm only monitoring a couple thousand users a day so the database logs for Splunk are very low, 50Mb/day max.
Occasional Contributor II
Posts: 13
Registered: ‎09-22-2009

I use cricket and nagios

I use cricket and nagios in conjunction to monitor my APs. Nagios does alarming, and writes performance data into a log file. Cricket picks up that log file and graphs the data. You can probably do the same thing in cacti with nagios2cacti.

http://exchange.nagios.org/directory/Addons/Graphing-and-Trending/nagios2cacti/details

The trick is you need to use an external program that caches the controller and mib location of the individual aps. aps can move around between controllers, or get replaced, and you want your monitoring to automatically follow them.

The first thing you do is query the switchListSwitchRole table in the master controller, to get the ips of all the local controllers.

Next grab the wlsxWlanAPTable.wlsxWlanAPEntry.wlanAPIpAddress column from each controller. The column index (last 6 octets of the oid) will be the mac address of the ap, and the value will be the ipaddress. Then pull the wlsxWlanAPTable.wlsxWlanAPEntry.wlanAPName column, and match up the mac addresses. You now have the name, mac, ip, and controller for each AP. Save that in a cache file (I use a perl Storable).

Once you have the cache, you can get the details on any given AP by looking up it's name in your cache and finding the right controller, and the ap's mac. Then issue a series of snmp gets (not walks) against the ap table sort of like:


for col = 2 to 31
snmpget wlsxWlanAPTable.$col.$mac


That will give you that row out of the table with the minimum number of snmp hits on the controller. If the row doesn't exist, the ap was either moved to a different controller, or removed, so regenerate the cache (use a hold-down timer).

Next get the data from the wlsxWlanRadioTable with something like this:


for radio = 1 to wlsxWlanAPTable.wlanAPNumRadios
for col = 2 to 14
snmpget wlsxWlanRadioTable.$col.$mac.$radio


Now you have most of the interesting data for the ap. You can apply thresholds, alarms, etc to it, and graph it.

I have code to do this for nagios, which also writes performance data that you can graph with various tools. Let me know if you're interested.
Occasional Contributor II
Posts: 13
Registered: ‎09-22-2009

Re: Monitoring Aruba APs with cacti



I have code to do this for nagios, which also writes performance data that you can graph with various tools. Let me know if you're interested.




I got a couple queries about this, which I accidently let sit for way too long. So I'm just going to post this, regardless of how embarassingly ugly the code looks.

Depending on options, this will check a controller, or query the controllers about an individual AP.

http://copland.udel.edu/~doke/aruba/check_aruba
Search Airheads
Showing results for 
Search instead for 
Did you mean: