Version Cleaner module

Edition

Incubator (services)

Git

Git

Latest

1.6.2

Compatible with Magnolia 6.2, 6.1, 5.7.

The Version Cleaner module provides commands to remove old or duplicated versions from your version store. Which versions are selected for removal can be configured through voters.

Analyzing and removing unneeded versions can take a lot of time, the version cleanup command can be configured to examine a set number of versions (actually nodes in the version workspace) and then stop. The next run of the version cleanup command will resume from the node last analyzed.

The version cleaner can be configured to select which versions are deleted or retained.

This module is at the INCUBATOR level.

Installing with Maven

Maven is the easiest way to install the module. Add the following to your bundle:

<dependency>
  <groupId>info.magnolia.services</groupId>
  <artifactId>versioncleaner</artifactId>
  <version>1.6.2</version>
</dependency>

Usage

The module provides a set of commands for removing selected versions from the version history.

Which versions are selected for deletion are configurable through voters. You can define complex conditions to identify if a version should be deleted.

Cleanup Commands

The cleanup commands can be found in the version cleaner catalog: /module/versioncleaner/commands/cleanup.

Command Description

cleanVersions

Remove unneeded versions according to its configuration.

cleanAllVersions

Remove all versions found except root and leaf versions.

Unlike cleanVersions, cleanAllVersions does not require possibly complicated configuration.

cleanNodeVersions

Clean versions from a workspace and path.

cleanVersions and cleanAllVersions work directly on the JCR version workspace. They clean versions from all workspaces with versions.

cleanNodeVersions however cleans versions from a specific workspace and optional path. Unlike cleanVersions and cleanAllVersions, cleanNodeVersions will not resume its cleaning run after reaching its visitLimit; it doesn’t save its progress while cleaning versions. cleanNodeVersions does not archive versions before deleting them.

The cleanVersions and cleanAllVersions commands can be configured with the following parameters when invoked:

Parameter Description

noop

true or false (the default)

Does a "dry run", going through the selected versions but without deleting or modifying them.

archive

true or false (the default)

If true, will save an export of each version to be deleted to the specified archiveDir.

Using this option may generate many files. Be sure you have sufficient space to store the archive versions.

archiveDir

Directory for storing archived versions and cleaning run history.

Defaults to: ${magnolia.cleanup.dir} if specified or ${magnolia.repositories.home} if not.

Magnolia must have read and write permissions for the directory.

visitLimit

The maximum number of version nodes that will be checked during a run.

Defaults to 1000 if not specified.

If set to a number less than 0, all version nodes will be checked.
These parameters can also be set as properties in the command configuration, at /module/versioncleaner/commands/cleanup/cleanVersions or /module/versioncleaner/commands/cleanup/cleanAllVersions.

Diagnostic Commands

The diagnostic commands can be found in the version cleaner catalog /module/versioncleaner/commands/diagnostic.

Command Description

findBadVersions

Locate corrupted versions in the version workspace.

A version is corrupt if it:

  • has no root version

  • has multiple root versions

  • the root version identified by name differs from the root version returned by the JCR API

Configuring Cleanup

The versions selected for removal by the cleanVersions, cleanAllVersions and cleanNodeVersions commands are controlled by Voters. The voters determine if a versions should be deleted (the result of the voting is true) or retained (the result of the voting is false).

You can combine the results of different voters with logical operations (and, or, xor…​).

Each command can be configured with the pruneVersions property:

configuring cleanup

Version Voters

You can use the following voters to identify versions to be deleted or retained:

EligibleVersionVoter

EligibleVersionVoter returns true if the version is not a root version (the first version of a node) or a leaf version (the last version of a node).

Class: info.magnolia.services.cleanup.commands.voters.EligibleVersionVoter

NoModificationVoter

NoModificationVoter returns true if the version has the same modification date as the previous version or false if the modification dates are different.

Class: info.magnolia.services.cleanup.commands.voters.NoModificationVoter

OldVersionVoter

OldVersionVoter returns true if the version is older than the maxAge threshold or false otherwise.

Class: info.magnolia.services.cleanup.commands.voters.OldVersionVoter

Table 1. Properties
Property Description

maxAge

optional

The age threshold in milliseconds.

Defaults to 90 days (i.e. 7776000000)

UnneededVersionVoter

UnneededVersionVoter returns true if there are at least versionsThreshold newer versions and false otherwise. Use UnneededVersionVoter to limit the total number of versions retained.

Class: info.magnolia.services.cleanup.commands.voters.UnneededVersionVoter

Table 2. Properties
Property Description

versionsThreshold

optional

The number of versions to be retained.

Defaults to 5.

Changelog

Version Notes

1.6.2

Corrected ordering of version workspace traversal

1.6.1

Corrected naming for pruning configuration for info.magnolia.services.cleanup.commands.NodeVersionCleanupCommand (pruneVersionspruneVersion).

1.6

  • Removed obsolete pruneHistory configuration.

  • Corrected resumption of pruning from history.

  • Added cleanNodeVersions command (info.magnolia.services.cleanup.commands.NodeVersionCleanupCommand).

Feedback