Skip to content
Percona Software for MongoDB Documentation

For help, click the link below to get free database assistance or contact our experts for personalized support.

Get help from Percona

Incremental physical backups

Version added: 2.0.3

Considerations

  • ⚠ Incremental backups made with Percona Backup for MongoDB prior to 2.1.0 are incompatible for restore with Percona Backup for MongoDB 2.1.0. This is because of the changed set of metadata files that are now stored in backups. These files are absent in backups made with previous PBM versions but are required for the restore with PBM 2.1.0.

    We recommend to make a new incremental base backup and start the incremental backup chain from it after the upgrade to Percona Backup for MongoDB 2.1.0

  • Incremental backup implementation is based on the $backupCursor aggregation stage that is available only in Percona Server for MongoDB. Therefore, you must be running Percona Server for MongoDB in your deployment to use incremental physical backups.

  • Incremental backups are supported for Percona Server for MongoDB starting with the following versions: 4.2.24-24 , 4.4.18 , 5.0.14-12 , 6.0.3-2 and higher.
  • Due to WiredTger restrictions in Log-Structured Merge (LSM) trees behavior when the $backupCursor is opened, incremental backups are not available if the LSM tree is configured in the database.

Owners of large datasets may need to back up data frequently. Making full physical backups every time is costly in terms of storage space. Incremental physical backups come in handy in this scenario, enabling you to optimize backup strategy and reduce storage costs.

During incremental backups, Percona Backup for MongoDB saves only the data that was changed after the previous backup was taken. This results in faster backup / restore performance. Since incremental backups are smaller in size compared to full backups, you also save on costs for their storage and transfer in case of cloud deployments.

graph LR
  A[Full physical ] --> B([Increment 1 ]);
  B --> C([Increment 2 ]);
  C --> |.....| D([Increment n ]);

Implementation specifics

  1. Percona Backup for MongoDB tracks the backup history only on the node where the base incremental backup was taken. This means that subsequent incremental backups must always be run on that very node. To make this happen, Percona Backup for MongoDB tries to schedule backups on that same node.

    If the node with the base incremental backup is down or unavailable, you must start the incremental backup chain anew on another node.

  2. Adjusting node priorities for backup interferes with the default behaviour described above. If you listed only a subset of nodes for the priority, the remaining nodes receive the default priority 1 and it may happen that the incremental backup is taken not from the node where the base backup was taken.

    To avoid this, set the highest priority for the node where you wish to incremental backups from. This instructs Percona Backup for MongoDB to start the backup from the same node and will preserve the backup chain.

    For sharded cluster, define the highest priority node for every replica set.

Make a backup Restore a backup

Last update: June 13, 2025
Created: June 13, 2025