I've seen some discussion of using a separate JVM for security. Does it
also provide a solution to the problem of JVM migration in a
continuously operating River-based system?
Suppose the objective is to migrate from 1.5 to 1.6.
The first step would be to get each client and server into a state in
which it can recognize that a proxy needs 1.6, and spawn a 1.6 JVM, with
1.6 versions of River jars, to run it. That can be done on a gradual,
machine-by-machine basis.
The second step is to replace each program installation with its JVM 1.6
version, again on a machine-by-machine basis. During this phase, there
will be times when a 1.5 application needs to run a proxy that needs
1.6, and solves the problem by spawning a 1.6 JVM to run it.
At the end of the second step the system is fully 1.6, and the 1.5 JVM
and libraries can be uninstalled.
If a service is implemented on multiple servers, and the servers are
updated at different times, the service remains continuously available.
Does this make any sense?
Patricia
