Hi Duncan,
    I've got some similar requirements.  What I was going to attempt (looks good; but I have no idea about practice) was ha-jdbc.  A jdbc driver that allows you to write to several identical dbs simultaneously.  There's overhead in insert/update/delete statements but it might be worth a shot.


Raymond Kroeker

On Feb 7, 2008 7:04 AM, Jørgen Løland <Jorgen.Loland@sun.com> wrote:
Duncan Groenewald wrote:
> Hi, does anyone know if it is possible to get some kind of log shipping
> working with Derby such that it is possible to create a second database
> instance that runs and periodically loads transaction log dumps from the
> primary database ?
> Is it even possible to load a transaction log while a database is running ?
> Basically I want to have a second copy of the database running in read
> only mode but still reasonably up to date, say every 15 minutes.
> Thanks
> Duncan Groenewald
> mobile: +61406291205
> email: dagroenewald@optusnet.com.au

Hi Duncan,

There is currently no such feature in Derby, but the next release will
include asynchronous replication based on log shipping from a master
version to a slave version of the database. However, the first version
of replication will not allow read operations on the slave database.

Take a look at the functional specification for replication on this
page: https://issues.apache.org/jira/browse/DERBY-2872

If what is described there suits your needs, you'll find this
functionality in the next release of Derby (ETA in a few months). If a
read-only second copy of the database is what you really need, I think
you will have to hack something together using scripts outside Derby for
now. I think I can guide you through that, but it *will* be pretty messy
and I have not tried it myself. Please let me know :)

Hope this helps
Jørgen Løland

Raymond Kroeker