Re: Distributed databases

[Øystein Grøvlen <Oy...@Sun.COM>]

>>>>> "JAC" == James A Craig/O/VCU <cr...@vcu.edu> writes:

    JAC> Hi, I'm fairly new to Derby but I was curious if its possible
    JAC> to use it in a distributed setup. I currently have a small
    JAC> cluster and want to set it up so that I have a distributed
    JAC> database on it using Derby. So my questions are:

    JAC>  

    JAC> 1) Is this possible?

It is a bit unclear to me what you mean by using it in a distributed
setup.  Do you just want to access the database from all nodes of the
cluster, or do you actually want for performance reasons a single
distributed database which runs on all nodes.  The first part is
possible and is called a network server, the latter is not.  The rest
of my answers will assume that you want to be able to access a
single-node database server from other nodes.

    JAC> 2) Is there a resource that shows me how to do this? (the
    JAC>    manuals don't seem to explain how to do this)

This is explained in the Derby Server and Administration Guide.
 
    JAC> 3) From what I've read it seems like I need to use the DB2
    JAC>    driver to do this
    JAC> (like I've said, I've found very little info on the subject
    JAC> of setting up a distributed database). Is this even remotely
    JAC> on base?

This was true for the first Derby release.  The latest release comes
with an open-source JDBC driver.

    JAC> 4) Does anyone know of any decent tutorials when it comes to
    JAC>    Derby? I haven't
    JAC> found anything that great thus far.

    JAC> 5) Can I use ij to manipulate a distributed database? How do I do this?

Yes, ij can connect to a network server if you specify an URL
containing hostname and port number.

    JAC> 6) Is there a way to setup one using ij?

There are scripts that may be used to start a network server.

-- 
Øystein

Re: Distributed databases

["Michael J. Segel" <ms...@segel.com>]

On Thursday 22 September 2005 04:07, Øystein Grøvlen wrote:
> >>>>> "JAC" == James A Craig/O/VCU <cr...@vcu.edu> writes:
>
>     JAC> Hi, I'm fairly new to Derby but I was curious if its possible
>     JAC> to use it in a distributed setup. I currently have a small
>     JAC> cluster and want to set it up so that I have a distributed
>     JAC> database on it using Derby. So my questions are:
>
>     JAC>  
>
>     JAC> 1) Is this possible?
>
> It is a bit unclear to me what you mean by using it in a distributed
> setup.  Do you just want to access the database from all nodes of the
> cluster, or do you actually want for performance reasons a single
> distributed database which runs on all nodes.  The first part is
> possible and is called a network server, the latter is not.  The rest
> of my answers will assume that you want to be able to access a
> single-node database server from other nodes.
>

I can understand your confusion. At first it looks like he wants to create an 
HADR setup using Derby. But in a later question, he used the term distributed 
meaning remote....

Assuming that he wants to do HADR and not create a parallelized engine, I 
think that there is some merit to this.


>     JAC> 5) Can I use ij to manipulate a distributed database? How do I do
> this?
>
(This is the question that I was referencing above.)

-- 
Michael Segel
Principal
MSCC
(312) 952-8175