Controllerless Networks

Aruba Version 6.4.4.11

Controller model: 7240

I have noticed that if a AP reboots while the primary LMS controller is down, the standby controller (the backup LMS IP) will not accept the AP.

Here's the log on the standby controller:

Active-mode Hello received from AP on a standby controller, ignoring

Is that normal? How could I prevent this?

Hi Bob,

Please share the topology. The outputs for the following commands will help for now:

1. show switches

2. show vrrp

3. show ha group-profile <name of profile>

what are the roles assigned to each controller in the HA profile ?

If possible, please change it to dual & check.

The topology is a N:1 Active/Standby deployment Model.

I have two buildings, each with a local controller (the primary LMS) in the active role.

The master controller is in a standy role (and configured as the secondary LMS IP) (as recommended by Aruba documentation).

The problem is, if you lose one of the local controller and some APs in that building reboot during that time, the master controller refused them with the log:

Active-mode Hello received from AP on a standby controller, ignoring

I feel like this is a weakness in the high availability of Aruba. Or am I missing something?

Hi Bob,

The general recommendation is to use role as dual wherever possible.

Do you mean the AP never comes up if Local 1 is down & AP reboots as master keeps ignoring request from the AP ?

I am summarizing the issue as per my understanding. Please correct me in case I am incorrect at any point

1. Lets assume there is one Master & 2 locals

Master 10.1.1.1

Local 1: 10.1.1.2

Local 2: 10.1.1.3

HA profile Local 1:

10.1.1.2 role active

10.1.1.1 role standby

HA profile 2:

10.1.1.3 role active

10.1.1.1 role standby

AP system profile 1:

LMS: 10.1.1.2

Bkup LMS: 10.1.1.1

AP system profile 2:

LMS: 10.1.1.3

Bkup LMS: 10.1.1.1

If we consider ap's on Local 1 & following sequence of events occurs:

1. AP's are up on Local 1

Their standby tunnels are on Master.

2. Local 1 goes down.

AP's failover to Master controller.

3. Now one of these AP's (which got failed over successfully) reboots (L1 is still down), Master ignores the Hello received from the AP.

Please let me know if the issue is summarized correctly.

What is the master discovery method being used ?

Ideally master will ignore the request from AP in case the LMS IP in AP system profile is pointing to Master rather than local.

In that case master role is configured as standby & we are trying to terminate AP's on master. So,  it will ignore the request from AP & legacy mechanism of lms/backup LMS will be initiated.

Yes this summarize the situation correctly.

So are you saying that in this case, the only difference between having the master controller (Bkup LMS) as dual instead of standby is to prevent this type of situation?

It will not cause any other difference?

Hi Bob,

The explanation for the roles is as follows:

1. Dual: A controller in this role will allow AP's to terminate on itself & can also act as standby for other AP's.

2. Standby: In this role , we can not terminate AP's directly on this controller.

Controller will only allow standby tunnels from the AP's. The standby tunnels will only become active when the other controller in the HA pair (role active or dual) will go down.

So, I would suggest you to keep the role to be dual & re-test the network.