Apache Derby Releases 10.3 Beta and Gains Experimental Hot Standby Replication
Apache Derby is a pure-Java, Open Source, relational database derived from IBM's Cloudscape. In the past couple weeks it has made two significant advances. First, the project made a beta available of the upcoming 10.3 release. The 10.3 release includes security enhancements, language-based ordering, other features, bug fixes, and performance enhancements. Second, Egil Sørensen, a student at the Norwegian University of Science and Technology, submitted his thesis work on replication and availability, including a prototype implementation.
- DBA Powers - operations such as shutting down, encrypting, and upgrading the database are restricted to the database owner, rather than being allowed by any authenticated user.
- Secure Server - the Network Server now runs with a Java security manager by default.
- Language Based Ordering - locale-sensitive collation is now allowed rather than simply sorting by Unicode character codes.
Minor features include:
- ALTER TABLE now supports DROP and RENAME of columns.
- SSL/TLS communication between client and server is now supported.
- It is now possible to enable client-side tracing without changing the application.
- Support for import/export of tables with clob, blob and other binary type columns has been added.
- JDBC methods for auto-generated keys have been implemented.
- It is now possible to create a new empty table based upon a sub-query.
Egil Sørensen has submitted his MS thesis work to add hot standby replication functionality to Derby.
By implementing a hot standby scheme in Apache Derby several features are added. The contents of the database is replicated at run time to another site providing online runtime backup. As the hot standby takes over on faults availability is added in that a client can connect to the hot standby after a crash. Thus the crash is masked from the clients. In addition to this, online upgrades of software and hardware can be done by taking down one database at the time. Then when the upgrade is completed the upgraded server is synchronized and back online with no downtime.
A fully functional prototype of the Apache Derby hot standby scheme has been created in this project using logical logs, fail-fast takeovers and logical catchups after an internal up-to-crash recovery and reconnection.
The prototype was submitted as a patch against the 10.2.2.0 release (i.e. the latest production release), but it has already been ported to the trunk. There is another issue requesting that replication functionality be added. As a result the patch may receive consideration for 10.4.
Still interesting, specially hot standby (called 'Clustering' in H2).
Re: Catch Up
Re: Catch Up
Maybe the problem is the default settings? The default transaction isolation level of H2 is SERIALIZABLE. For Derby, it is READ COMMITTED. H2 supports that as well. To try it out, use the following database URL: jdbc:h2:~/test;LOCK_MODE=3.
Re: Catch Up