This document outlines the changes that were made in ModeShape &version;. We hope you enjoy it!
&version; provides a fast, elastic, distributed hierarchical database that clients work with via the standard JCR 2.0 (JSR-283) API. ModeShape 4 is a major upgrade over 3.x and offers significant improvements in clustering, performance, query, and events. All JCR 2.0 features are supported, and ModeShape 4 has complete integration with Wildfly 8, allowing deployed applications to simply lookup and use repositories managed by ModeShape's service.
As of 4.0.0.Alpha1, ModeShape is licensed under the Apache Software License, 2.0.
This is the third alpha release of the 4.0 stream, and it includes several fixes and new features. Of note is that the event system now uses a uses our new ring buffer that is substantially faster. Of course, there's no change in the event APIs so your listener implementations will continue to work unchanged.
The first alpha release introduced a new query engine that allows clients to explicitly define the indexes used in the query system, and the second alpha release brought minor changes to the Service Provider Interface (SPI) for query providers and introduced a programmatic API and configuration modifications for defining indexes, although no complete query providers are included (see next release). Alpha2 had support for the JCR event journal feature, allowing applications to poll for changes that occurred during time ranges. This is a useful alternative to listeners that may be expensive or time-consuming.
Since this is an alpha release, not all features targeted to 4.0 are complete, and there are likely bugs with some of the newer features. Therefore, please do not put &version; into production.
However, we would like to get as much feedback as possible, so we do ask that our community do limited testing with &version; to help us identify problems. Specifically, we ask that you test the following areas:
- JDK - ModeShape now requires JDK 7. We've not yet begun testing with Java 8, but we'd be happy to hear about it if you do.
- Queries - although there are known issues (see MODE-2178 for details), we think that most queries will work, albeit a bit slower, especially on medium or large repositories. We'll increase speed by adding indexes in a forthcoming alpha release, but in the meantime please check that your queries work. If not, please file an issue.
- Clustering - ModeShape no longer has a clustering section in its configuration, since we simply piggyback on top of Infinispan's clustering setup. We've also upgraded to a newer version of JGroups.
- Journalling - Try enabling journaling and verify it works and does not affect performance. You can't yet see what's in the journals, but even some modest testing with journaling enabled will help us.
- Infinispan - We've moved to Infinispan 6.0.1.Final, which is faster and has new cache stores. Some older and poorly-performaing cache stores are no longer valid, so check out the new file-based cache stores. Also, the LevelDB cache store is supposedly very fast.
- Backup and restore - given that some older Infinispan cache stores are no longer supported, in order to test migrating 3.x repositories to 4.0 you will need to use ModeShape's backup and restore feature. If you don't regularly use that, please test it with your repository. Just be sure not to overwrite any 3.x repositories.
- Bugs - we've fixed a number of bugs reported against and fixed in 3.x; see the list below for details. All these are ready for testing.
- Configuration - comments are now allowed in our JSON configuration.
ModeShape &version; has these features:
- ModeShape uses Infinispan for all caching and storage, giving a powerful and flexible foundation for creating JCR repositories that are fast, scalable, and highly available. Infinispan offers a great deal of storage options (via cache loaders), but using Infinispan as a distributed, mulit-site, in-memory data grid provides incredible scalability and performance.
- Strongly consistent. ModeShape is atomic, consistent, isolated and durable (ACID), so writing applications is very natural. Applications can even use JTA transactions.
- Fast. ModeShape 4 is just plain seriously fast, and performance is all-around faster than earlier version.
- Larger content. ModeShape 4 can store and access the content so that a node can have hundreds of thousands (or more!) of child nodes (even with same-name-siblings) yet still be incredibly fast. Additionally, repositories can scale to many millions of nodes and be deployed across many processes.
- Simplified configuration. There is no more global configuration of the engine; instead, each repository is configured with a separate JSON file, which must conform to a ModeShape-specific JSON Schema and can be validated by ModeShape prior to use. Repository configurations can even be changed while the repository is running (some restrictions apply), making it possible to add/change/remove sequencers, authorization providers, and many other configuration options while the repository is in use.
- Elastic. Add processes to scale out, without having to have a single coordinator.
- Deploy, start, stop and undeploy repositories while the engine is running and while and other repositories are still in use.
- Sessions immediately see all changes persisted/committed by other sessions, although transient changes made by the session always take precedence.
- Monitoring API with over a dozen metrics.
- Sequencing SPI that uses the JCR API to get at the content being processed and create/update the derived content. Sequencers can also dynamically register namespaces and node types. Now it's easy to create custom sequencers.
- Connector SPI that defines how external systems are accessed and optionally updated to project the external information into the repository as regular nodes.
- Simple API for implementing custom MIME type detectors. ModeShape still has a built-in Tika-based detector that determines MIME types using the filename extensions and binary content.
- Simple API for implementing custom text extractors, which extracts from binary values searchable text used in full-text searches and queries.
- Ability to store binary values of any sizes, with a separate facility for storing these on the file system, in Infinispan caches, in relational DBMSes (via JDBC), and in MongoDB. Custom stores are also possible.
- Public API interfaces and methods that were deprecated in 2.7.0.Final (or later) have been removed. There weren't many of these; most of the ModeShape API remains the same as 2.x.
- Integration with JBoss AS 7. ModeShape runs as an integrated subsystem within AS7, and the AS7 tooling can be used to define and manage repositories independently of each other and while the server is running.
- Local and remote JDBC drivers for issuing JCR-SQL2 queries and getting database metadata via the JDBC API
- Use the RESTful API to talk to ModeShape repositories from non-Java and non-local applications
- Use the CMIS API to talk to ModeShape repositories
- Use WebDAV to mount ModeShape repositories as file servers
- Visual repository explorer web application
All of the JCR 2.0 features previously supported in 2.x are currently supported:
- RepositoryFactory access
- JNDI registration of Repository
- JAAS Authentication
- Servlet Authentication
- Custom Authentication
- Session Remapping
- Permanent Addition/Deletion
- Traversal Access
- Direct Access
- Same-Name Siblings
- Multi-Value Properties
- All Property Types Supported
- Property Type Conversion
- Create/Update/Delete Nodes
- Create/Update/Delete Properties (Through Parent Nodes)
- Moving, Copying, Cloning
- Adding/Removing Mixins
- Referential integrity enforcement
- XPath
- JCR-SQL
- JCR-SQL2
- JCR-QOM
- Full-Text Search
- System View Import/Export
- Document View Import/Export
- Inheritance Among Node Types
- Discovering available Node Types
- Discovering the Node Types of a Node
- Discovering the Definition of a Node Type
- Property Constraints
- Automatic Item Creation
- Predefined standard Node Types
- Custom Node Type Registration (CND-Based and and JCR 2.0 API Template-Based)
- Permanent Namespace Mappings
- Node Types (Built-In and User-Registered)
- Active Locks
- Observation
- Locking
- Versioning
- Shareable nodes
- Access controls
- In-memory (local, replicated, and distributed)
- Relational databases (via JDBC), including in-memory, file-based, or remote
- LevelDB
- File system
- Cassandra
- Cloud storage (e.g., Amazon's S3, Rackspace's Cloudfiles, or any other provider supported by JClouds)
- Remote Infinispan
- File system
- JDBC database
- Infinispan
- MongoDB
- Chained binary stores
ModeShape also has features that go beyond the JCR API:
- File system connector (read and write)
- Git repository connector (read-only)
- CMIS reposiotry connector (read and write, tech preview)
- JDBC metadata connector (read-only)
- Compact Node Definition (CND) Sequencer
- DDL Sequencer
- Image Sequencer
- Java Source Sequencer
- Java Class Sequencer
- MP3 Sequencer
- MS Office Sequencer
- Text Sequencers (Delimited and Fixed Width)
- XML Sequencer
- XML Schema Document (XSD) Sequencer
- Web Service Definition Lanaguage (WSDL) 1.1 Sequencer
- Zip File Sequencer (also WARs, JARs, and EARs)
- Teiid Relational Model Sequencer
- Teiid VDB Sequencer
- JNDI-Based Deployment
- Deploy as a subsystem in JBoss AS7, with access to repositories via @Resource injection
- Deploy to other containers using ModeShape's JCA adapter
- Access through two RESTful Services (the 2.x-compatible API and a new improved API)
- Access through WebDAV Service
- Access through CMIS Service
- Access through visual repository explorer web application
- Local and remote JDBC drivers for accessing ModeShape content through JDBC API and JCR-SQL2 queries
- Embedded (in Server or JEE Archive) Deployment
- JTA support, allowing Sessions to participate in XA and container-managed transactions
- OSGi-Compatible Archives
- Repository-wide backup and restoration
- Automatic MIME type detection of binary content
- Asynchronous sequencing operations, within completion notified via events
The following are the bugs, features and other issues that have been fixed in the &version; release: