Hi Nilesh,

I'm not sure how the JDBC connection string you mentioned would have worked for you.

As far as I'm aware, this is not supported by Derby - you should have gotten an ERROR 08001: No suitable driver exception.

Maybe you are using a 3rd party product to achieve auto connection failover at the client level? sequoia does support a fairly similar syntax, where you can have more than 1 controller for a virtual database within a Sequoia realm, such as:

The only failover mechanism supported by Derby during replication is a manual one where the slave can become accessible as a regular database endpoint (using the 'failover' connection property), see:



On Wed, Sep 16, 2009 at 3:28 AM, Nilesh Ranade <nranade@gemstone.com> wrote:


I am using the later derby 10.5.1 version. I read about the replication & failover mechanisms in the product manuals.

I had a few queries regarding the network client usage.


  1. I tried multiple host/port pairs in the connection string. My connection string looks like the following

jdbc:derby://host1:1534/sample1, host2:1534/sample2; create=true;


            Derby did accept such a connection string and established connection with the first entry. What I would like to know is in case node “host-1” goes down will the network client failover to the “host-2” database. I was looking for the auto failover capability rather than the “warm standby” mentioned in the manuals.


  1. In the above connection string representation, if “host-1:1534/sample1” is non rechable at the time of client connection. Will the client move to the 2nd entry and try to establish connection with “host2:1534/sample2”
  2. If point 2 can be accomplished can I simulate the auto failover using reconnect timeout ie to say that once host-1 goes down the client will try to establish connection again and connect with the second entry ( “host2/sample2) as the first entry is unreachable.



Nilesh Ranade