Hello,
In fact, we are addressing this problem in two ways:
- up to JPPF v4.0, there is a
dedicated sample/demo which illustrates how a failover strategy can be implemented for the nodes via an
initiialization hook.
- in the upcoming JPPF v4.1, we make it a fully fledged JPPF feature, as described in this feature request:
JPPF-212 Add a hook for specifying the server to which a node should connect.
I would also like to add 2 small remarks about your stated understanding of how the configuration of the connection works:
- instead of "driver2.priority", you should now use the new form "driver2.jppf.priority". This results from an effort in harmonizing the properties names in JPPF. Currently, the code still checks for "driver2.priority" when the new form is not specified, to ensure backward compatibility with older configuration files, but this will not always be the case.
- the strategy based on the priority hierarchy is no longer part of the JPPFClient implementation. As a consequence, to use it as a connection failover strategy, you need to wrap it in another built-in class, as detailed in the
the ClientWithFailover wrapper class section of the documentation.
I hope this clarifies.
Sincerely,
-Laurent
