iC2I: What Pools Rely on This Server?
Prerequisites: This article assumes you have read the first article in the series, that you have access to a Java compiler, and that you have built the iControl API on your development machine – or ...
Published Apr 11, 2008
Version 1.0
Hey ard,
You are absolutely correct. The reasons for advertising a server on two IP addresses is to make it appear as two end-points to the network. Thus, the BIG-IP treats them as two distinct endpoints because it is a networking device.
Because of this relationship of IP to endpoint, BIG-IP stores the definition of a "node" as an IP address with a "screen name" a user-friendly name that can be displayed.
There are some ways you could get around this - use a common naming convention for multiple IPs on a single machine and then use get_list() on the node API to get a list of all nodes and search through for the common names (like WEB1IP1, WEB1IP2 for screen names and then search the list of all nodes for anything starting with WEB1). Then this code could be run once per "node" to come up with a complete list.
This might be a great iC2I article in the near future, it builds on this topic and is useful for anyone using virtualization or higher end servers. Even those using lower end multi-homed servers for redundancy.
Don.