Blockchain snapshots

Download blockchains state in order to set up a validator node or RPC. See nearcore for more information on node requirements and usage.

Also, visit https://near-nodes.io for comprehensive details.

The instructions below utilize logic from this FastNEAR repository: https://github.com/fastnear/static

Mainnet

RPC Mainnet Snapshot

Note: This is likely the preferred approach for syncing, as opposed to downloading an archival snapshot, which is significantly larger and more special-purpose.

Make sure you have rclone installed. Install it by running:

rclone installation

sudo -v ; curl https://rclone.org/install.sh | sudo bash

Before running the snapshot download script, you can set the following environment variables:

CHAIN_ID to either mainnet or testnet. (default: mainnet)
THREADS to the number of threads you want to use for downloading. Use 128 for 10Gbps, and 16 for 1Gbps (default: 128).
TPSLIMIT to the maximum number of HTTP new actions per second. (default: 4096)
BWLIMIT to the maximum bandwidth to use for download in case you want to limit it. (default: 10G)
DATA_PATH to the path where you want to download the snapshot (default: ~/.near/data)
BLOCK to the block height of the snapshot you want to download. If not set, it will download the latest snapshot.

Run this command to download the RPC Mainnet snapshot:

We will set the following environment variables:

DATA_PATH=~/.near/data - the standard nearcore path
CHAIN_ID=mainnet - to explicitly specify the mainnet data

RPC Mainnet Snapshot » ~/.near/data

curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/fastnear/static/refs/heads/main/down_rclone.sh | DATA_PATH=~/.near/data CHAIN_ID=mainnet bash

Archival Mainnet snapshot

Time and storage intensive.

Be prepared for a large download and the inherent time constraints involved.

The snapshot size is ~60Tb and contains more than 1M files.

Work in progress

The archival snapshots are work in progress. We are updating the docs and the scripts to make it easier to download. One concern is the snapshot BLOCK has to be the same between hot and cold data runs. Also you have to run script twice to download hot and cold data.

Make sure you have rclone installed. Install it by running:

rclone installation

sudo -v ; curl https://rclone.org/install.sh | sudo bash

Before running the download script, you can set the following environment variables:

CHAIN_ID to either mainnet or testnet. (default: mainnet)
THREADS to the number of threads you want to use for downloading. Use 128 for 10Gbps, and 16 for 1Gbps (default: 128).
TPSLIMIT to the maximum number of HTTP new actions per second. (default: 4096)
DATA_TYPE to either hot-data or cold-data (default: cold-data)
BWLIMIT to the maximum bandwidth to use for download in case you want to limit it. (default: 10G)
DATA_PATH to the path where you want to download the snapshot (default: /mnt/nvme/data/$DATA_TYPE)
BLOCK to the block height of the snapshot you want to download. If not set, it will download the latest snapshot.

By default the script assumes the paths for the data:

Hot data (has to be on NVME): /mnt/nvme/data/hot-data
Cold data (can be on HDDs): /mnt/nvme/data/cold-data

Run the following commands to download the Archival Mainnet snapshot:

Download the latest snapshot block height:

Latest archival mainnet snapshot block

LATEST=$(curl -s "https://snapshot.neardata.xyz/mainnet/archival/latest.txt")
echo "Latest snapshot block: $LATEST"

Download the HOT data from the snapshot. It has to be placed on NVME.

We will set the following environment variables:

DATA_TYPE=hot-data - downloads the Hot data
DATA_PATH=~/.near/data - the standard nearcore path
CHAIN_ID=mainnet - to explicitly specify the mainnet data
BLOCK=$LATEST - specify the snapshot block

Archival Mainnet Snapshot (hot-data) » ~/.near/data

curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/fastnear/static/refs/heads/main/down_rclone_archival.sh | DATA_TYPE=hot-data DATA_PATH=~/.near/data CHAIN_ID=mainnet BLOCK=$LATEST bash

Download the COLD data from the snapshot. It can be placed on HDDs.

We will set the following environment variables:

DATA_TYPE=cold-data - downloads the Hot data
DATA_PATH=/mnt/hdds/cold-data - the path where to place cold data. Note: the nearcore config should point cold data store to the same path.
CHAIN_ID=mainnet - to explicitly specify the mainnet data
BLOCK=$LATEST - specify the snapshot block

Archival Mainnet Snapshot (cold-data) » /mnt/hdds/cold-data

curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/fastnear/static/refs/heads/main/down_rclone_archival.sh | DATA_TYPE=cold-data DATA_PATH=/mnt/hdds/cold-data CHAIN_ID=mainnet BLOCK=$LATEST bash

Testnet

RPC Testnet Snapshot

Note: This is likely the preferred approach for syncing, as opposed to downloading an archival snapshot, which is significantly larger and more special-purpose.

Make sure you have rclone installed. Install it by running:

rclone installation

sudo -v ; curl https://rclone.org/install.sh | sudo bash

Before running the snapshot download script, you can set the following environment variables:

CHAIN_ID to either mainnet or testnet. (default: mainnet)
THREADS to the number of threads you want to use for downloading. Use 128 for 10Gbps, and 16 for 1Gbps (default: 128).
TPSLIMIT to the maximum number of HTTP new actions per second. (default: 4096)
BWLIMIT to the maximum bandwidth to use for download in case you want to limit it. (default: 10G)
DATA_PATH to the path where you want to download the snapshot (default: ~/.near/data)
BLOCK to the block height of the snapshot you want to download. If not set, it will download the latest snapshot.

Run this command to download the RPC Testnet snapshot:

We will set the following environment variables:

DATA_PATH=~/.near/data - the standard nearcore path
CHAIN_ID=testnet - to explicitly specify the testnet data

RPC Testnet Snapshot » ~/.near/data

curl --proto '=https' --tlsv1.2 -sSf https://raw.githubusercontent.com/fastnear/static/refs/heads/main/down_rclone.sh | DATA_PATH=~/.near/data CHAIN_ID=testnet bash