02-04-2010 07:24 AM
itself as down. It is, in fact, not down but associated with a different
controller. (We only have two - Master and Local setup.)
Last Friday, our Master rebooted. (Still looking into why.) When
this happened, the AP's on that controller jumped to the Backup Controller
(as they should have).
This one Group was working along just fine until one of the AP's
rebooted and went back to the master controller. It then started to report
itself as Down (by the Backup controller). The Master has it as Up and
running while the Local Controller says it's down.
I then rebooted the entire group (one AP at a time) and they all
came backup associated with the Master (as they should). The problem is
they are still reported as Down on the Local Controller. It's as if the
backup controller's database did not get cleared(?).
The two controllers seem to be syncing just fine:
xxx.xxx.xxx.xxx aruba-master Kendade.KMR master 22.214.171.124 up
UPDATE SUCCESSFUL 0
xxx.xxx.xxx.xxx aruba-local CHP MR local 126.96.36.199 up
UPDATE SUCCESSFUL 21
I believe that the AP's in this group are working
correctly... they are just reporting as down.
Here is one of those AP's listed from the ap group from both
prospect-1e prospect 65 10.23.238.24 Up 20h:12m:46s
prospect-1e prospect 65 10.23.238.24 Down
How do I get these two controllers back in sync? (Not that this
matters but - I use snmp calls to monitor both controllers for down ap's.
So I've been getting email every hour that they are down, when they really
are not.) I've been searching through the Command line ref pdf looking for
some way to fix this but have not come up with anything yet.
Any thoughts on this? (Am I missing something obvious?)
02-04-2010 11:03 AM
02-05-2010 06:56 AM
Yes, the Master is the primary LMS and the local is the backup LMS.
The AP-Group points at the Master's IP (Loopback) for the "LMS IP".
The the Local (Backups Loopback) for the the "Backup LMS IP".
02-05-2010 07:16 AM
The Master and local AP status is NOT synched. The master should have a unified view of the network, and that is where you should look for status. The local controller only has status for APs that have terminated on its own controller. Status is never sent from the master to the local.
Aruba Customer Engineering
Looking for an Answer? Search the Community Knowledge Base Here: Community Knowledge Base
02-05-2010 07:26 AM
I would also like to suggest an architectural change. If you only have two controllers and all of the APs you have now or plan on having in the future will fit on one controller, why not make both controllers masters, setup a VRRP group and have the APs only point to the VIP. That way, failover time is reduced. This assumes you have a layer two connection between the controllers. If not, I would suggest making the local controller the primary LMS and the master the backup. As Colin said, the master always has a good view of the network and your monitoring host would only need to poll that single controller. The master is also responsible for a bunch of calculations for the APs, so it has a bit more of a load than a local controller would normally. In most networks, I try not to terminate APs on the master controller (unless it is very small and doesn't justify a dedicated controller).
02-08-2010 07:42 AM
I would also like to suggest an architectural change. If you only have two controllers and all of the APs you have now or plan on having in the future will fit on one controller, why not make both controllers masters, setup a VRRP group and have the APs only point to the VIP.....
We have thought of this but we are not done adding AP's yet. We will end up with more than will fit on one controller.
02-08-2010 11:33 AM
You may want to touch base with your SE to figure out a scenario that would work the best for what you have to work with.
02-09-2010 07:22 AM
I'm sure there are reasons you have chosen your current architecture. ARM and WMS calculations are sent to the master controller (assuming Airwave is not being used with WMS offload), which are the calculations to which Olin spoke. This is additional overhead, so the ideal situation is that your master controller would be the least loaded with access points.
To get back to the post topic, you can clear the erroneous "Down" messages from the local controller using the following command on the master controller.
Ryan Holland, ACDX #1 ACMX #1
The Ohio State University
10-10-2013 06:44 AM
I cant read the command you posted, I'm having this problem, our master is seeing the APs of a local controller as DOWN, but in the local, they are up and working in some level....except the captive portal.