AIX

I have a HACMP 6.1 configured in a active/passive. I have 1 NIC with 3 IP address on (Boot, Persistent and Service ) . All address are routable.
One of the application on the HA cluster is also using Boot Ip to send application data.

Question : Since all the traffic is passing thru the same NIC , Does it matter if the Boot Ip address is also used by application ?

I understand that Boot Ip is strictly for HA purpose and is used for heart beat messages, but it would be helpful if someone can answer this question ?

TIA

The bootaddress does not move to another node or even another adapter by HA means. So if your application that is using the boot address is part of a Resource Group successful failover/takeover cannot take place (without a lot of scripting). Don't do that.

It might also be helpful if you submerge yourself in the PowerHA documentation to read about the basic concepts

Boot Ip is not part of the resource group. Application uses Java RMI and picks up the first address.

Since BootIp is routable in this case, application has no problem to connect to the external resources over this IP.

In case of failover, I believe application will still be fine , as it will pick up boot ip of the failover node (local node) in that case.

I am not sure about what will be happen if the application send a large data over this IP . Will this cause heart beat skips . ?

The point of HACMP is that users can rely on a certain IP address delivering a certain service regardless of some single system being operative or not. Suppose the following situation: for the user a certain database is always available at IP 1.2.3.4, regardless of machineA being up (and answering under this address) or machineB being up (and answering under this address). Both machines will be aware of the state of the other and HACMP is a set of rules how the application, its data and it related IP addresses are transferred from one machine to the other in case of failure of one of these machines.

If your application picks up any IP address at random this process is not working any more and you might check if you really need HACMP in this case.

Shockneck is absolutely correct: you may want to get some general knowledge about the concepts and workings of HACMP (for instance by downloading and reading some Redbook) before you attempt to configure a cluster.

I hope this helps.

bakunin