I want to get clarification and make sure I understand. keySet as well as MapReduce do not get keys from the loader. They will only get keys that are in memory. Is this correct? If this is correct, is there a issue an Jira where I can track this? I want to make sure I understand because this would have a significant impact on our design. We are looking to use different caches as indexes for faster look up operations. And if this limitation does currently exist, then I do not think we want to use a persistent store for those caches.
Is it okay to log an enhancement in JIRA to enable using the distributed execution and map-reduce on any cache?
Yes please, thank you!
Done. Please see ISPN-2192.
I'm actually wrong: the Map/Reduce integration with CacheLoader was adde din 5.2[1] and is available in the last Alpha. Sorry for the confusion!