High availability rSeries/iSeries

Hello ImadBak 

Please check the documents - BIG-IP redundant configuration hardware and software parity requirements

BR
Aswin

Hi ImadBak ,

Welcome to the F5 DevCentral community!

Can I clarify further? Are you trying to set up HA pair between r2600 and i2600?

Cheers,

Mo.

The rSeries platform uses F5OS, which in practice is a Kubernetes-based virtualization, allowing the creation and hosting of tenants in the same way we used to with vCMP guests. The BIG-IP tenant is managed in the same way a vCMP guest is managed today on VIPRION or iSeries.

In the rSeries platform, there is no appliance redundancy; instead, redundancy is based on tenancy (similar to HA of a vCMP guest) - High Availability in rSeries

You can find more details about the rSeries appliance platform F5OS-A - rSeries Platform Introduction
You can find more details about the VELOS chassis platform F5OS-C - VELOS Platform Introduction

That said, when considering an HA cluster between an iSeries and a tenant hosted within the rSeries, it is possible but not recommended, as there are some limitations to this configuration when features depend on similar platform resources (HW or Virtual Edition). Connection mirroring and persistence may not be supported for device groups consisting of heterogeneous hardware, virtual platforms, or BIG-IP software versions, as described in the article mentioned by Aswin_mk

In the compatibility matrix for creating/hosting BIG-IP tenants on F5OS, version 16.x is not supported - Compatibility Matrix

For platform migration, I recommend using f5-journeys, which will help with UCS adjustments - f5-journeys GitHub

Hello MoFaz , 

We want to synchronize r2600/i2600 appliances between two sites , site active (A) with 2 appliances r2600 and backup site (P) with 2 appliances i2600 .

In failover , the backup takes over with the two i2600 which are synchronized .

is this possible ?

if you want to reduce cost, you can use VE for DNS part because DNS traffic is very small and DNS doesnt have to be in same node as LTM.
clients only needs to query a hostname's IP address once for multiple application requests.
and it caches that DNS response according to TTL you set (application's idle time out + 1 minute for best practice).