05-10-2012 06:09 PM
We have just received a couple of 240H's to implement, and I have some questions around HA & active-active clustering in conjunction with two Cisco 3750's.
Each SRX has one link to a different Cisco 3750. The idea is that the design would allow for one SRX and one switch to fail simultaneously, whilst still operating.
My question is, what sort of arrangement should I use for the redundant ethernet interfaces on the Juniper? From what I've read we'll require two - reth0 & reth1, but I'm not sure what VLANs I should be trunking on each interface (cable pair, essentially). It seems ineffective to have the same VLANs trunked on each pair of cables.
Would one reth suffice and provide for active-active clustering with sufficient failover?
05-15-2012 03:17 AM
Its not compulsary that you configure reth0 and reth1
You can configure only one reth interface and create subinterfaces on it. (Just like any other eth interface)
I hope this helps...
Please Mark My Solution Accepted if you think it helped!
05-15-2012 05:03 AM
first, there is no real active-active clustering on the srx 240. (per documentation.)
(because you can only have redundancy-group 0, which holds the routing-engine, active on one node.)
you can achieve semi-active-active clustering if you put two reth-interfaces in seperate redundancy-groups, and have each group active on one node.
also make sure you read this thread:
tbehrens wrote a very nice post on active/active clustering, focusing on the srx.
You can also find me on Freenode IRC in #juniper, my handle is "cy"