Print

[spacitron]

Hi all, I'm new to JPPF. Great piece of kit by the way!

Anyway, I've been going through the documentation in the last couple of days but I still have a question on configuring nodes and servers. I'm trying to create a small test application using Amazon EC2. So ideally I'd like to instantiate EC2 instances when I need them, perform whatever computation I need and finally terminate them. But obviously I won't know the location of each EC2 instance until it has been created. So I'd have to be able to configure servers and nodes at runtime from the client.

How can I do this? Can I just use JPPFConfiguration.getProperties().setProperties(); from within the client to set properties on the servers/nodes?

[lolo]

Hello,

A good way to handle this would be to use the JPPF configuration source plugin.
You could use it with 2 implementations of the JPPFConfiguration.ConfigurationSource interface:
- one implementation for the server: it would determine its IP address, read the rest of its configuration from an accessible datasource (like a file in an S3 bucket, a database, etc.), then write the nodes configuration to the same kind of datasource
- one implementation for the nodes: it would read the configuration from where the server implementation wrote it

Sincerely,
-Laurent

[spacitron]

----EDIT----

Ok forget that, how do I pass the IP address from the server to the JPPFConfiguration class?