Double-booting system behavior

Derby attempts to prevent two instances of Derby from booting the same database by using a file called db.lck inside the database directory (see "The Database Directory").

On all platforms running with a JDK of 1.4 or higher, Derby can successfully prevent a second instance of Derby from booting the database and thus prevents corruption.

On some platforms running with a JDK lower than 1.4, Derby may prevent a second instance of Derby from booting the database (previous to JDK 1.4 the ability to do this was OS dependent).

If this is the case, you will see an SQLException like the following:

ERROR XJ040: Failed to start database 'sample', see the next exception for details.
ERROR XSDB6: Another instance of Derby might have already booted
the databaseC:\databases\sample. 

The error is also written to the error log.

If you are running a JVM prior to 1.4, Derby issues a warning message on some platforms if an instance of Derby attempts to boot a database that already has a running instance of Derby attached to it. However, it does not prevent the second instance from booting, and thus potentially corrupting, the database. (You can change this behavior with the property derby.database.forceDatabaseLock.)

If a warning message has been issued, corruption might already have occurred. Corruption can occur even if one of the two booting systems has "readonly" access to the database.

The warning message looks like this:

(instance 80000000-00d2-3265-de92-000a0a0a0200) is
attempting to boot the database /export/home/sky/wombat
even though Derby
(instance 80000000-00d2-3265-8abf-000a0a0a0200) might still be active.
Only one instance of Derby
should boot a database at a time. Severe and non-recoverable corruption can
result and might have already occurred.

The warning is also written to the error log.

If you see this warning, you should close the connection and exit the JVM, minimizing the risk of a corruption. Close all instances of Derby, then restart one instance of Derby and shut down the database properly so that the db.lck file can be removed. The warning message continues to appear until a proper shutdown of the Derby system can delete the db.lck file.

When developing applications, you might want to configure Derby to append to the log. Doing so will help you detect when you have inadvertently started more than one instance of Derby in the same system. For example, when the derby.infolog.append property is set to true for a system, booting two instances of Derby in the same system produces the following in the log:

Sat Aug 14 09:42:51 PDT 2005:
Booting Derby version Apache Derby - - (29612): 

instance 80000000-00d2-1c87-7586-000a0a0b1300 on database at
directory C:\tutorial_system\sample
Sat Aug 14 09:42:59 PDT 2005:
Booting Derby version Apache Derby - - (29612):
instance 80000000-00d2-1c87-9143-000a0a0b1300 on database at
directory C:\tutorial_system\HelloWorldDB

Derby allows you to boot databases that are not in the system directory. While this might seem more convenient, check that you do not boot the same database with two JVMs. If you need to access a single database from more than one JVM, you will need to put a server solution in place. You can allow multiple JVMs that need to access that database to connect to the server. The Derby Network Server is provided as a server solution. See the Derby Server and Administration Guide for more information on the Network Server.

Related concepts
One Derby instance for each Java Virtual Machine
Booting databases
Defining the system directory
The error log
Recommended practices
Related tasks
Shutting down the system