Data Domain: Understanding Initialization and Resynchronization | MTree Replication

Initialization

Initialization is performed when creating a new MTree replication context.

Process

• • Initialization is performed using the full destination MTree path.
  • When initialization begins, the system creates a snapshot on the source to capture the MTree at a specific moment.
  • Ingest to the source MTree continues without interruption.
  • The initialization process copies all existing data and snapshots from source to destination.

Behavior

• • Initialization depends on snapshots to determine what must be transferred.
  • There is no impact to ingest during initialization.
  • If an attempt is made to re‑initialize a previously initialized context, an error occurs because the destination MTree is read‑only.
    In this case, a resync must be used instead.

Resync

Resync is used when an existing MTree replication context must be re‑established.

Requirements

• • A resync can only be performed after the MTree replication context has been initialized.
  • Attempting a resync without prior initialization results in an error.
  • Resync uses the full path of the destination MTree.

Process

• • Resync identifies the most recent common synchronization point between source and destination.
  • Data is transferred from that point forward to re‑establish synchronization.
  • As with initialization, resync uses snapshots to determine changed data.

Behavior

• • There is no ingest impact during resync due to reliance on snapshots.
  • Resync is used whenever:
    • The replication context has been broken.
    • The context had previously been initialized.
    • Replication must be restarted without re‑initializing (or changing) the destination MTree.
    • To reinstate an existing replication context to working order