[memoryerror]

4) Removed getCacheModeInternal().

[memoryerror]

By making the cache have <K,V>, does that mean for every different pair of K and V we need to create a different cache?

[memoryerror]

Okay, I found the docs. I've ported our code. Now we need to deploy it into production, which is using Jboss 4.3.CP02 (and not easily changed). Will we be able to do that, or are the versions too different to be compatible?

[manik]

"jshowalter" wrote:
Okay, I found the docs. I've ported our code. Now we need to deploy it into production, which is using Jboss 4.3.CP02 (and not easily changed). Will we be able to do that, or are the versions too different to be compatible?

Depends. You cannot upgrade the versions of JBC that the app server uses for its own clustering code, but if you isolate class loaders, you can package newer JBC versions in your app and use it. Read this doc for details:

http://www.jboss.org/community/docs/DOC-10254

[memoryerror]

Interesting--that would allow us to decouple our use of the cache from our dependencies in the data center. However, it looks like that would require introducing our own config file for our cache, and double cache configuration on each node in the cluster. Or am I reading the docs incorrectly?

[manik]

Yes, it would be completely independent of the cache setup that the app server uses. Different IP and port would be recommended if you can do this, especially for your multicast address.

[mjrother]

Since putForExternalRead is not in JBC 1.4.1....is using Option Local still the recommended approach?

What are the timing issues with that approach?