Recovery

Recovery restores missing committed block data to a node from the validator's block backup. It is used when promoting a full node to sequencer after the active sequencer is lost, and the promotion target is missing committed blocks.

This complements the Sequencer Failover flow. Full nodes replicate the sequencer state asynchronously, so there is no guarantee that the node is fully up to date if the sequencer fails. These missing data can always be recovered from the validator because it has to a sign every block before it can be committed.

When to recover

Recover when both of the following hold:

• The active sequencer is lost or being replaced, and a full node is being promoted to take its place.
• The promotion target is behind the committed chain tip and cannot catch up from another in-sync source.

What recovery does

The miden-node recover command streams the validator's signed blocks into the node's local storage, starting from the node's current committed tip and stopping once it reaches the validator's chain tip. It then exits.

Recovered blocks are signed but carry no proofs — the validator backs up block data, not block proofs. These blocks must be imported or re-proven separately as part of recovery before the node resumes block production as a sequencer.

Procedure

1. Stop the sequencer, if it is not down already.

2. Stop the full node being promoted.

3. Run recovery against the validator, pointed at the promotion target's data directory:

   miden-node recover \
     --data-directory node-data \
     --validator.url http://validator:50101

   The command applies blocks up to the validator's chain tip and exits. If the node is already at the validator's chain tip, it reports that there is nothing to recover and exits successfully.

4. Commission proofs for the recovered blocks.

5. Restart the node as a sequencer. See Sequencer.