10-27-2010 12:40 AM - edited 03-06-2019 01:45 PM
Hi!
I have setup the network as shown in the attached topology.
The two Barracuda Load Balancers have been deployed in bridge mode meaning that they are each configured with only one physical IP address.
They have been configured to operate in an Active-Standby mode whereby the WAN port of each Load Balancer is connected to Switch A and the LAN port of each Load balancer is connected to Switch B. The Load balancers communicate and monitor each other by sharing a secret cluster password.
In this setup all traffic going to/coming from the servers need to pass through the Load Balancers.The network on both sides of the Load Balancers (WAN and LAN ports) have the same IP addressing which is 172.22.92.0/24.
The setup works fine but at some times, a loop occurs in the network and this brings the network completely down where no device can be reached.
This usually occurs when plugging in the second Load Balancer or when modifying the configuration of the Active Load Balancer.
Please find attached the topology, configurations and errors generated.
I have used spanning-tree portfast on Switch A ports to lessen the time it takes for the Backup Load Balancer to discover that the Active Load Balancer is no longer operating correctly.
Regards,
Alvin
10-27-2010 04:02 AM
Hello,
Could you please check to which device do MAC addresses 0017.5401.03e0 and 0017.5401.192e belong to?
Could you please post the output of "show spanning-tree vlan 22" command?
Regards.
10-27-2010 09:16 AM
Alvin
I can't read visio files so this may not be relevant but just to check.
You have the Barracuda's in bridge mode using the same IP subnet. Are you using 2 vlans with the same IP subnet ie. you are not using the same vlan on both "sides" ?
Jon
11-03-2010 06:07 AM
11-03-2010 06:28 AM
The Barracuda's are most likely not participating in STP and will probably not forward STP BPDU's.
Also, I expect them to communicate with eachother (for redundancy) using some multicast-based protocol.
Because this is forwarded by the switches you are seeing the source addresses loop over their ingress / egress ports.
This is most likely an unsupported topology. Please check the manual for the Barracuda's to verify this.
You might be able to select a different mechanism for the redundancy checking.
regards,
Leo
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: