db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Francois Orsini <francois.ors...@gmail.com>
Subject Re: Using Derby with Microsoft Clustering Service
Date Mon, 22 Jun 2009 22:44:47 GMT
When you create a database with Derby, you can specify an absolute database
location path as part of the JDBC connection URL; so you can place a
database on a particular drive, whether it is a shared one or not.

What you cannot do is place the database transaction journal log file on a
separate drive.

As far as running Derby in some embedded or network client-server
configuration, I think you have to take into consideration how much CPU /
DISK resources would be required by the database engine.

If you run Derby embedded, it is very likely that the Derby engine might
compete with the application server that embeds it. Again, it all depends on
the activity coming from the application. Of course on an SMP (multi-cpu /
core machine), it might be less.

Running Derby as a standalone server allows the database processing to be
isolated on a particular node but a client-server topology introduces
potential network latency due to data exchanged between the client and the
server....It is a pretty traditional topology of course.

As far as running in a cluster, bear in mind and this is important that
multiple Derby database engine instances CANNOT share a database (on-disk).
This is because Derby does not have a shared-disk cluster mode. In other
words, it means that each cluster node where a Derby instance is running can
only manage its own database whether this last one in on a shared drive or
not. Of course, multiple client users can still access a single Derby engine

More information about Derby database connection URL and how to specify a
database path can be found here:

Hope this helps a bit


On Mon, Jun 22, 2009 at 2:23 PM, forumer <gmamla@gmail.com> wrote:

> What is the recommended configuration for running Derby in Microsoft
> Clustering Service to improve its availability? Would it be to run one
> Derby
> Server on each node with database on shared drive? If so, how can one
> specify so data is stored on a designated drive?
> Thanks
> --
> View this message in context:
> http://www.nabble.com/Using-Derby-with-Microsoft-Clustering-Service-tp24155721p24155721.html
> Sent from the Apache Derby Users mailing list archive at Nabble.com.

View raw message