Fork me on GitHub

Command Line Tools

Oak Run

This tool supports maintenance operations such as backup and compaction.

Some of the documentation of this tool is found at https://github.com/apache/jackrabbit-oak/tree/trunk/oak-run and some at http://jackrabbit.apache.org/oak/docs/query/pre-extract-text.html and some at http://jackrabbit.apache.org/oak/docs/query/oak-run-indexing.html

Compatibility

We recommend to use the same version of oak-run as the one of oak-core deployed on your system.

Backup

To back up a repository, use:

java -mx4g -jar oak-run-*.jar backup <repository> <backup>

When using the (default) Tar storage, the <repository> setting is the path to the directory that contains the segment (data*.tar) files. The <backup> option is the target directory.

Checkpoint

To list the checkpoints of a repository, use:

java -mx4g -jar oak-run-*.jar checkpoints <repository>

When using the (default) Tar storage, the <repository> setting is the path to the directory that contains the segment (data*.tar) files.

Compaction

The oak-run compact operation may be used to perform an offline compaction:

java -mx4g -jar oak-run-*.jar compact <repository>

It makes sense to find and remove the old checkpoints, using the checkpoints command described above.

Index Management

The oak-run index operation is used to manage indexes.

System properties supported by the oak-run

When performing operations on the SegmentNodeStore (eg. backup, checkpoints, compact), it's possible to tweak various system properties to get the optimal performance. These are:

  • cache - cache size for the SegmentNodeStore (default: 256),
  • compaction-progress-log - how many entries should be compacted between log messages (default: 150000),
  • update.limit - threshold for the flush of a temporary transaction to disk (default: 10000),
  • tar.memoryMapped - allow to use memory mapped files to reduce memory usage (default: false),

Example:

java -Dtar.memoryMapped=true -mx4g -jar oak-run-*.jar checkpoints <repository>

Analyzing Thread Dumps

To analyze thread dumps, filter out idle threads, and get profiling data, use the following, there the directory contains thread dump files:

java -jar oak-run-*.jar threaddump --filter --profile <directory>

MongoDB Shell with oak-mongo.js

MongoDB comes with a command line shell that can be extended with custom JavaScript. Oak provides a oak-mongo.js file with MongoMK specific extensions. You first need to download the JavaScript file and then start the MongoDB shell with the file.

wget https://s.apache.org/oak-mongo.js
mongo --shell oak-mongo.js

Now the MongoDB shell has an oak namespace with additional functions.