Skip to main content
Version: v2.0.2

Using and uploading snapshots

Overview​

All nodes on the Aztec network must download and synchronize the blockchain state before they can operate. This guide covers different sync modes, including how to use snapshots for faster synchronization and how to create your own snapshots.

Understanding sync modes​

Nodes can synchronize state in two ways:

  1. L1 sync: Queries the rollup and data availability layer for historical state directly from Layer 1
  2. Snapshot sync: Downloads pre-built state snapshots from a storage location for faster synchronization

Since Aztec uses blobs, syncing from L1 requires an archive node that stores complete blob history from Aztec's deployment. Snapshot sync is significantly faster, doesn't require archive nodes, and reduces load on L1 infrastructure, making it the recommended approach for most deployments.

Prerequisites​

Before proceeding, you should:

  • Have the Aztec node software installed
  • Understand basic node operation
  • For uploading snapshots: Have access to Google Cloud Storage with appropriate permissions

Using snapshots to sync your node​

Configuring sync mode​

Control how your node synchronizes using the --sync-mode flag:

aztec start --node --sync-mode [MODE]

Available sync modes:

  • snapshot: Downloads and uses a snapshot only if no local data exists (default behavior)
  • force-snapshot: Downloads and uses a snapshot even if local data exists, overwriting it
  • l1: Syncs directly from Layer 1 without using snapshots

Setting the snapshot source​

By default, nodes use Aztec's official snapshot storage. To specify a custom snapshot location, use the --snapshots-url flag:

aztec start --node --sync-mode snapshot --snapshots-url [BASE_URL]

The node searches for the snapshot index at:

[BASE_URL]/aztec-[L1_CHAIN_ID]-[VERSION]-[ROLLUP_ADDRESS]/index.json

Note: Currently, only Google Cloud Storage is supported for snapshots. URLs follow this format:

https://storage.googleapis.com/aztec-testnet/snapshots/

Example: Force snapshot sync with custom URL​

aztec start --node --sync-mode force-snapshot --snapshots-url https://storage.googleapis.com/my-snapshots/

This configuration forces the node to download a fresh snapshot on every startup from the specified location.

Creating and uploading snapshots​

You can create snapshots of your node's state for backup purposes or to share with other nodes. This is done by calling the nodeAdmin_startSnapshotUpload method on the node admin API.

How snapshot upload works​

When triggered, the upload process:

  1. Pauses node syncing temporarily
  2. Creates a backup of the archiver and world-state databases
  3. Uploads the backup to the specified storage location
  4. Resumes normal operation

Uploading a snapshot​

Use the admin API to trigger a snapshot upload. The upload destination must be a Google Cloud Storage URI (e.g., gs://your-bucket/snapshots/).

Example command (assumes node admin API is running on localhost:8880):

curl -XPOST http://localhost:8880 \
  -H 'Content-Type: application/json' \
  -d '{
    "method": "nodeAdmin_startSnapshotUpload",
    "params": ["gs://your-bucket/snapshots/"],
    "id": 1,
    "jsonrpc": "2.0"
  }'

Replace gs://your-bucket/snapshots/ with your Google Cloud Storage bucket path.

Scheduling regular snapshots​

For continuous backup, schedule the upload command to run at regular intervals using cron or a similar scheduler. The frequency depends on how current you need your snapshots to be.

Once uploaded, other nodes can download these snapshots by configuring their --snapshots-url to point to your storage location.

Verification​

To verify your sync configuration is working:

For snapshot downloads​

  1. Check startup logs: Look for messages indicating snapshot download progress
  2. Monitor sync time: Snapshot sync should be significantly faster than L1 sync
  3. Verify state completeness: Confirm your node has the expected block height after sync
  4. Check data directories: Ensure the archiver and world-state databases are populated

For snapshot uploads​

  1. Check API response: The upload command should return a success response
  2. Monitor logs: Watch for upload progress messages in the node logs
  3. Verify storage: Check your Google Cloud Storage bucket to confirm the snapshot files exist
  4. Validate index file: Ensure the index.json file is created at the expected path
  5. Test download: Try downloading the snapshot with another node to confirm it works

Troubleshooting​

Snapshot download fails​

Issue: Node cannot download snapshot from the specified URL.

Solutions:

  • Verify the --snapshots-url is correct and accessible
  • Check network connectivity to the storage location
  • Confirm the snapshot index file exists at the expected path
  • Review node logs for specific error messages
  • Try using Aztec's default snapshot URL to isolate custom URL issues

Snapshot upload fails​

Issue: The nodeAdmin_startSnapshotUpload command returns an error.

Solutions:

  • Verify Google Cloud Storage credentials are properly configured
  • Check that the specified bucket exists and you have write permissions
  • Confirm sufficient disk space is available for creating the backup
  • Review node logs for detailed error messages

Storage space issues​

Issue: Running out of disk space during sync or snapshot creation.

Solutions:

  • Ensure sufficient disk space (at least 2x the expected database size for snapshots)
  • Clean up old snapshots or data if running recurring uploads
  • Monitor disk usage and set up alerts
  • Consider using a larger volume or adding storage

Best practices with snapshots​

  • Use snapshot sync for production: It's faster and more efficient than L1 sync
  • Schedule regular snapshots: If running critical infrastructure, create snapshots at regular intervals
  • Test snapshot restoration: Periodically verify that your snapshots can be successfully downloaded and used
  • Monitor storage costs: Google Cloud Storage usage accumulates; implement retention policies if needed
  • Keep snapshots updated: Older snapshots take longer to sync to the current state
  • Use force-snapshot sparingly: Only use force-snapshot when you need to reset to a known state

Next Steps​