Redundancy best practices

Hi,

 

I am in the process of implementing several 4000s clusters in our datacenters and have some questions regarding redundancy and route domains.

 

I have attached a png of our setup, yes... made in MS Paint... what is currently available for me :)

 

 

The clusters will be setup in a 'Sync-Failover' mode in active-standby. Running version is currently 11.3 hf5

 

Partitions will be used to segment our customers, where a route domain will be created and set to that partition's default route domain. The customers will all reside on different vlans.

 

 

Our F5 partner told us that the Management IP's should be unique, for us meaning we have to use public IPs to avoid overlapping IPs. Is that correct? I see no reference for the management routing in RD0

 

But I DO see that our ConfigSync selfip's are represented in RD0 in other partitions which inherit RD0 from /Commons.

 

Does this mean that it is acually the ConfigSync/Mirroring interfaces that should be unique to avoid a potential IP overlap?

 

 

Furthermore, under System >> HA >> Device Connectivity >> Network Failover - I see the device's own management IP listed at Failover Unicast.

 

Should only the local ip be listed? or both nodes to function correctly - In our specific setup I think I would add a new selfip on a vlan on the bonded interface, to test the connectivity where the customers connectivity will flow.

 

Thank you for any comments.

 

 

Best regards,

 

Henrik Noerr