Hello-

We have an AES system deployed in GRHA with 2 nodes in different subnets but no vIP, as we understand that to have a vIP the servers need to be in the same subnet. All the documentation & SDK seems to reference managing any high availability/failover when systems are configured with 1 vIP. Is there some information on how to manage connection/reconnection/failover when using AES with 2 IP addresses.

As far as I can tell from the DMCC PG, without virtual IP Addresses, GRHA gives very limited benefits. That is:
o Any configuration made to the active is reflected in the standby
o If the standby detects an outage of the active, it will bring itself into the active state
o There is only a need for one set of licenses.

If the application detects a lost connection to the active AES, it will need to start a new session with the (previously) standby AES and re-establish registrations and monitors.

I have attached an AES Admin guide and HA white paper which may give you more information than the programmers guide.

Martin

Thanks Martin,
Should we be able to run GRHA AES with a vIP when the servers in different subnets? If there are different subnets are there 2 vIPs?

This paragraph from the PDF is not the clearest
Allows the AE Services server in each datacenter to be in different networks
or subnets where the static address of each server is used. GRHA is also
supported where both AE Services servers share a virtual IP address for each
configured network interface and each virtual IP address is in the same subnet
as both AE Services servers

I read that to mean:
o If each AES has a separate IP Address, in different subnets, then you only get the basic features that I mentioned previously.
o If Each AES and their equivalent virtual IP Addresses are in the same subnet, you get extra benefits.

Please note that AES has 3 ethernet ports. Normally, in a lab we just use one port for everything. However for a production environment, you may prefer to use all three, each would be in a separate network. The Management port is used for administration, the private port is connected to CM and the public port is connected to application servers. This is why the document mentions virtual IP Addresses (plural) in places.

Is Avaya best practice to run in the same subnet to get full GRHA functionality? We are trying to set up our lab to test AES failover and want to set up that way Avaya Professional Services would reccomend to an enterprise customer.

Sorry, I can't hep you there. Our remit is with APIs, not system design.

Martin