PMP450 Compatibility Issue with Cisco Managed Switch

We had installed a PMP450 network in year 2015 in one of the refinery located in NE India.The refinery already had a wired back bone across the plant area. Due to frequent link failure of wired network which severely hamper production of the refinery, they have finally adopted our solution to overcome the downtime issue.

In initial stage, we have faced a lot of issues which arises from competibility between Cambium Networks and Cisco switches. Later on we had resolve the issues by manually configure MAC of Cambium Networks AP and SM in core  and edge switches. After observing and regorous inspection of traffic behaviour of the network, we had discovered that the Cisco switches could not able get update of radio MAC address in its MAC table due to which we had to manually gave entry in the switches . We could overcome the issue and the network is working fine till date. In the meantime, till date we could not able to make the redundant network  as automatic changeover type as spanning tree doesnot work in that particular production network which when run on test lab with same setup works smoothly. Our main struggle point in troubleshooting this case is basically arises no OEM(Cisco and Cambium Networks) took the leadership of the issue which is quite understandable . However, Cambium Networks had took initial initiative but could not succeed and the issue of STP in the network hanged there till date. The client had to keep disable the ethernet interface till his wired network goes down. Once the wired link get failed, the client need to enable the ethernet interface to bring up the link. This is the scenerio in our case which does not bring glory although it could not bring any blame for us till now. I have published this issue in our esteemed community to get any solution if anyone has face the same issue and overcome with better solution than us.

I am not clear where the issue is.

If I understand your setup, you have the Cisco setup to do STP, and the Cambium 450 network is a "standby" link, but you want it to become active when the main (wired) link stops working.  The issue is that the Cisco does not realize the wireless link is active and can make the connection.  Is that right?

Are the Cambium radios in bridge mode? Do they have an IP that resides in the Cisco MAC table?

This sounds like a bit of a complex issue, and may require more than we're able to handle via forum posts, but we can start here.

Hi Matt, it's quite complex scenario which we had faced. There were so many support cases registered in both Cambium Networks and Cisco but none could provide us a justified explanation. We can split the situation in to two stages-

Stage1:

we were unable to ping CMM,EtherWAN and AP/SM  but client could access the respective VLAN without any issue. 

Observation:-

"Being other devices including Cambium device cannot control what is learned or not learned in a Cisco or any other switch either the Cisco switch did not receive a packet with the source address of the radio or it ignored it.  It was just a guess that the MAC address table entry timed out on the Cisco switch and was cleared"     

Possible Solution:

We had overcome this by  statically added the MAC address on the Cisco switch(core) pointing to the port in which the Cambium device is connected to and it will worked.

Future Impact:

But as the solution that we had proposed was basically for redundancy if the link goes down the table entry will be incorrect and may fail. But interestingly, no such issue arise till date.

Stage2:

When both Optical Fiber link and wireless link were connected to Edge Switch for redundancy and spanning tree configured with desired parameters which is in general normal practice in wired network, then loop was occured in the network which turn down the complete network.

Observation:

Even if the wired network was in use still the wireless link had made the loop. Regarding this issue we had contact Cisco but they could not find any issue in their switch network as they are capable of stopping any loop in the network were not ready to take ownership of the STP issue which was associted with Cambium Networks device. 

Possible Solution:

As all SMs are reachable even if we disable ethernet , hence whenever any wired link at any specific switch goes down we simply enable the ethernet interface. This resolve the issue which arises from STP loop.

Future Impact:

As the solution was taken granted by client only  account of auto switch over between the links, which is not yet possible. I am still looking a justified solution from anyone due to which I have just published in our community which we could not resolve till date even with 15-20 years experience of system integration field.

A new issue has been observed in same PMP450 network for last couple of days. A new location is connected in the network with PMP450SM which carry trunk from basestation( PMP450AP) which is terminated on Cisco WS-2960X L2 switch. But the client machines connected over wire(Cat6 Cable) or wireless(ExtremeNetworks) from the L2 switch shows " Unidentified Network" error and stop communicating. The issue only resolves after rebooting the L2 switch which experiences the issue every after 15-20 minutes. During the error face by client machine, administrator can access the PMP450Sm and Cisco 2960X switch quite smoothly. Can anyone help us to understand the issue so that we can think of any plan for troubleshooting ? We have already logged one case with Cisco but I'm sure if they provide any solution for the problem. It's quite urgent to know the origin of the issue immediately.


@DEEPJYOTI BARMAN wrote:

A new issue has been observed in same PMP450 network for last couple of days. A new location is connected in the network with PMP450SM which carry trunk from basestation( PMP450AP) which is terminated on Cisco WS-2960X L2 switch. But the client machines connected over wire(Cat6 Cable) or wireless(ExtremeNetworks) from the L2 switch shows " Unidentified Network" error and stop communicating. The issue only resolves after rebooting the L2 switch which experiences the issue every after 15-20 minutes. During the error face by client machine, administrator can access the PMP450Sm and Cisco 2960X switch quite smoothly. Can anyone help us to understand the issue so that we can think of any plan for troubleshooting ? We have already logged one case with Cisco but I'm sure if they provide any solution for the problem. It's quite urgent to know the origin of the issue immediately.


Hi  DEEPJYOTI BARMAN , yout problem it's similar to this?

http://community.cambiumnetworks.com/t5/PMP-450/Data-traffic-interruption-with-Ethernet-errors/m-p/84989#M6154

@ Tuvix_IT ,Thanks for your reply.

 But the scenario here is totally unmatched that have mentioned in the link you have provided because of the following reason's

1. "shutdown" and "no shutdown" command on the port does not restore the connectivity to the connected client PC, which can only be restored by rebooting the Switch.

2. Here in our case, only L3 and L2 switches are in use, no router. We have already faced issues earlier with another notwork where router was used which later on resolved using configuring tracker on routers.

3. When we lost connectivity from switch to the connected client machines, I could able to ping the gateway or other machines in same VLAN but at differenet location(different switch) except the connected.

Waiting for more reply from our community.