VMware Blockchain nodes have a limited local persistent storage capacity. The local persistent storage cannot save the complete blockchain history because it requires massive storage resources to save the history of high-volume clearing and settlement transactions.
To avoid performance problems due to a lack of storage resources, it is essential to securely back up and delete stale key values and free up storage space on the nodes. Deleting stale key values from the VMware Blockchain nodes is called pruning.
The Daml execution engine on the Replica nodes determines which key values are stale. The execution engine tags the stale key values as part of executing a transaction. These stale key values remain in the local persistent storage until pruning is initiated. The default pruning window is two weeks. Any key marked as stale and older than 2 weeks, is pruned. This time frame can be manually reduced or increased.
Pruning requires the blockchain to stop processing requests. Therefore, the blockchain should be initiated during the maintenance window. The pruning process stops when the last prunable block agreed by the consensus mechanism is reached.
Data pruning is irreversible. After the data is pruned from the Replica nodes, the data is no longer available to the Client nodes. As a best practice, you must back up the Replica and Client nodes before initiating a pruning process.
See the VMware Blockchain Back Up and Restore chapter in the Using and Managing VMware Blockchain guide to back up your data.