Backup and Restore Strategy

Regular Backup and Restore

Bloomreach Experience Manager stores all data in the repository database. To prevent data loss, back up the entire repository database daily.

Creating a Backup

All content and configuration in Bloomreach Experience Manager is persisted in the database. Creating a database backup is sufficient to preserve all data. You do not need to stop the system during backup. Use your standard database backup procedure.

Restoring a Backup

To restore a backup, follow these steps:

1. Stop all nodes in the cluster.
2. Restore the database from the backup.
3. Delete the index folder on every cluster node before restarting. This ensures the index is rebuilt to match the restored data.
4. Start each cluster node. The index will rebuild automatically during startup.

The index location is typically configured in context.xml using the repository-directory parameter.

For example:

<Parameter name="repository-directory"
           value="${catalina.base}/../repository"
           override="false"/>
copy

With this configuration, the index folder is located at:

${catalina.base}/../repository/workspaces/default/index
copy

Backup and Restore Including the Lucene Index

Rebuilding the Lucene index after a database restore can require several hours if the repository contains a large volume of content (such as hundreds of thousands of documents). During index rebuilding, the entire cluster remains offline, and both the site and CMS are unavailable.

If downtime during index rebuilding is unacceptable, you can back up the Lucene index along with the database. This process is more complex and should only be used when strict uptime requirements apply. In most cases, the added complexity does not justify the time saved.

Creating a Backup with the Lucene Index

1. Stop one cluster node.
2. Record the JOURNAL_ID for that node from the REPOSITORY_LOCAL_REVISIONS table in the database.
3. Back up the Lucene index folder from the stopped node.
4. Back up the database.
5. Restart the node.

Restoring a Backup with the Lucene Index

1. Stop all cluster nodes.
2. Restore the database from the backup.
3. Restore the Lucene index folder from the backup to every cluster node.
4. Set the JOURNAL_ID in the REPOSITORY_LOCAL_REVISIONS table to the value recorded during backup for all cluster nodes.
5. Start all cluster nodes.