1 Reply Latest reply on Aug 5, 2013 12:04 PM by mircea.markus

    Lack of useful examples.

    alex_ro_bv

      Hello guys.

       

      I am currently working on a project where we tried a framework for storing on disk some data and recover after a sudden failure of the server but it wasn't very useful.

      I searched a little bit on the internet and infinispan looks promising. But the lack of examples and missing documentation is not an advantage.

       

      Basically what we need would be a cache

           -     which can flush on disk in some conditions (more than 5k objects are created in memory, so decides with a strategy to flush on disk).

           -     also I would need to be able to reload all the flushed data back on restart of the cache (am aware that in memory might not have been sync). If all the data from memory is stored on disk aswell via some async process, that would be great (so all the objects in memory have a mirror on disk).

           -     transactional. Basically short lived transactions.

      Optional:

           -     standalone( maybe the server which I already played with)

           -     shareable (between multiple nods)

           -     clustering (mirror)

           -     self discovering

       

      We are using amazon ec2, so the above optional queries would make a lot of sense.

       

      I understand that clustering would mean a communication via hot rod, or something else, and that would exclude transactional part.

       

      I was searching for this type of information in order to build a proof example, but was unlucky. I could not find any solution for retrieving data from disk.

       

      I think I do not speak just for myself when asking for this type of example, it could prove very useful for a lot of devs for the future.

       

      Thank, Alex.