Wireless Access

Hello to everyone, 

 

I have 2 controllers (as we said Controller A and Controller B) clustered as well as running VRRP. Also we are running LACP on ports 0/0/0 – 1 on both controllers.

 

-        I have configured the virtual AP as you can see (192.168.109.30 is the VIP from VRRP):

 

This way, data from clients will flow anyway, any time if any controller is UP.

 

-        I have configured to use ADP for discovering the Master controller and download the image:

So far every thing is OK.

 

Well, If I shut down Controller A the AP losts the conection and stop working, so, it means, it is not able to go on working with out Controller A. At the same time, Controller B turns to be Master for cluster and VRRP so, it is OK.

 

The AP reboots itself again and again till Controller A is connected again.

 

Is it the right way to work??

 

If you need more information do not hesitate contact me.

 

Thank you in advanced

In the AP Group (configuration> wireless> AP Configuration> Edit), expand AP and then AP system profile.  In the AP system profile, put the VRRP IP in the LMS-IP parameter.  That tells what ip address access points in that AP group point to.

 

When an access point comes up, it finds a controller.  After that it is redirected to the ip address in the LMS-IP field.  You want that to be the VRRP ip address..

 

 

It has been already configured. The images I attached was taken from the LMS IP field you said.

Do you have the controllers configured as master-local or master-backup master?  Do they have the same configuration?  Do they have the same VLANs created on both controllers?

 

As I can see on the config Master - Stand by, so I guess they are Master - Backup.

 

Master has several VLANS that the backup does not have but they have not been enabled so far.

 

Can you please share how do you have configure your VRRP ?

 

Make sure that your VRRP is not in a shutdown state too

 

show vrrp <vrrp id>

 

Also does the VRRP VLAN exist on both controller and is it up/up ?

 

 

 

192.168.109.30 is VRRP IP

The upper part is taken from the cluster (wich Master is Aruba_7210_1), the other part is taken from backup controller (Aruba7210_Bull_2)

 

Are any of these 2 controllers acting as Master (Not in VRRP) or do you have another controller for that and these are the locals ?

 

Have you console into the AP to see what are they doing when this happens or what controller they are trying reach?

 

 

 

---

@vfabian wrote:

 

Are any of these 2 controllers acting as Master (Not in VRRP) or do you have another controller for that and these are the locals ?

 

Have you console into the AP to see what are they doing when this happens or what controller they are trying reach?

 

 

 

---

The first one is the master for the cluster. When it is down, the second one takes over.

 

What AOS version are you using ?

Both controllers are running 6.2.0.0 version

 

Edit:
When Controller A goes on again, the same cuts on continuous ping happens

Okay.

 

If you have control plane security enabled, please disable it.

 

Configuration> Network> Controller> Control Plane Security> Disable.

---

@cjoseph wrote:

Okay.

 

If you have control plane security enabled, please disable it.

 

Configuration> Network> Controller> Control Plane Security> Disable.

---

Hi cjoseph, 

 

as you told me I have disabled Control Plane Security and it work perfectly now. No ping lost on a continuous ping even disabling link on Controller A.

 

Thank you very much to you and the other for the fast support. You are great.

Hi, Colin,

Please advice if Control place security is enable on both controllers than wat is the reason to rebooting of AP ?

If the AP has control plane security it could take longer to fail over.  I do not have access to his setup to determine why the access points are rebooting.

 

1.  Start a constant ping to 192.168.109.30 (ping 192.168.109.30 -t)

2.  Start the failover process by unplugging the cable to the Master controller

3.  See if you can still ping 192.168.109.30

4.  See if the access points end up on that controller by doing a "show ap database" on the "new" Master Controller.

 

 

---

@cjoseph wrote:

1.  Start a constant ping to 192.168.109.30 (ping 192.168.109.30 -t)

2.  Start the failover process by unplugging the cable to the Master controller

3.  See if you can still ping 192.168.109.30

4.  See if the access points end up on that controller by doing a "show ap database" on the "new" Master Controller.

 

 

---

That proof has been done. 

I have reloaded both contrllers and now this is the scenario:

1) AP connects with Controller A to download the image.

2) AP reloads woth the image

3) Constant ping from my computer to the AP IP.

4) I shut down the Controller A links so that AP goes to Controller B.

5) After 20 seconds the AP reloads and works.

6) After 1 minute aprox the AP reboots again and work fine indefinitely (now, it is connectdo to Controller B)

7) The continuous ping has 2 cuts at point 2) and point 5)

 

For sure, that cuts means no data traffic for clients so redundancy is not working

---

@JoseMiguel wrote:

---

@cjoseph wrote:

1.  Start a constant ping to 192.168.109.30 (ping 192.168.109.30 -t)

2.  Start the failover process by unplugging the cable to the Master controller

3.  See if you can still ping 192.168.109.30

4.  See if the access points end up on that controller by doing a "show ap database" on the "new" Master Controller.

 

 

---

That proof has been done. 

I have reloaded both contrllers and now this is the scenario:

1) AP connects with Controller A to download the image.

2) AP reloads woth the image

3) Constant ping from my computer to the AP IP.

4) I shut down the Controller A links so that AP goes to Controller B.

5) After 20 seconds the AP reloads and works.

6) After 1 minute aprox the AP reboots again and work fine indefinitely (now, it is connectdo to Controller B)

7) The continuous ping has 2 cuts at point 2) and point 5)

 

For sure, that cuts means no data traffic for clients so redundancy is not working

---

Why are the controllers not the same image?  That is very important.