accessing a neo4j storage from two JVMs

[Matteo Moci]

Hello everyone,
I'd like to know if a behaviour like the following can be supported by neo4j.

I have a neo4j graph, used in embedded mode,
that is accessed by two processes in two different JVMs.

The first one is a webapp that opens the graph and is able to search
in the graph's indices
(e.g. full text, exact, etc... ) - this can be done even in a read_only mode;
the second one is a batch process that once a day cleans the database
(clean indexes, remove all vertices and edges)
and writes the latest snapshot of data that should be available for
searching to the other JVM.

The thing that bothers me is that the graph is loaded two times from
the directory (one per JVM),
and I don't know if the read_only supports other JVM to load it and
modify its contents.

Thanks!
Matteo

--
Matteo Moci
http://it.linkedin.com/in/matteomoci
http://about.me/matteomoci/bio

[Michael Hunger]

Why not have just one JVM that owns the db (like the neo4j server) and
offers a remote domain-level API ?

Michael

> --
>
>

[Matteo Moci]

yes, that could work:
in this case it would offer web services to interact with the graph, right?

> --

[Michael Hunger]

whatever is suitable

- rmi
- http
- websocket
- binary stream (thrift, protobuf)
- ...

> --
>
>

[Matteo Moci]

yes, anything that works, thanks!