README

This HOWTO contains instructions on how to build and configure a RUST validator node in TON blockchain. The instructions and scripts below were verified on Ubuntu 20.04.

Table of Contents

• Getting Started
  • 1. System Requirements
  • 2. Prerequisites
    • 2.1 Set the Environment
    • 2.2 Install Dependencies
  • 3. Deploy RUST Validator Node
  • 4. Check Node synchronization
  • 5. Configure validator multisignature wallet
  • 6. Configure DePool
  • 7. Upgrade RUST Validator Node
• Stopping, restarting the RUST Node
• Logging
  • During deployment
  • During operation
• Migrating from C++ node
• Troubleshooting
  • 1. Couldn’t connect to Docker daemon at http+docker://localhost
  • 2. thread 'main' panicked error when checking node synchronization
  • 3. Error executing command when checking node synchronization
  • 4. Cannot stop/restart/remove node container
  • 5. DePool state not updating

Getting Started

1. System Requirements

SSD/NVMe disks are obligatory.

2. Prerequisites

2.1 Set the Environment

Adjust (if needed) main.ton.dev/scripts/env.sh:

Set export DEPOOL_ENABLE=yes in env.sh for a depool validator (an elector request is sent to a depool from a validator multisignature wallet).

Set export DEPOOL_ENABLE=no in env.sh for a direct staking validator (an elector request is sent from a multisignature wallet directly to the elector).

cd main.ton.dev/scripts/
. ./env.sh 

Note: Make sure to run the script as . ./env.sh, not ./env.sh

2.2 Install Dependencies

install_deps.sh script supports Ubuntu OS only.

./install_deps.sh 

Install and configure Docker according to the official documentation.

Note: Make sure to add your user to the docker group, or run subsequent command as superuser:

sudo usermod -a -G docker $USER

3. Deploy RUST Validator Node

Do this step when the network is launched. Deploy the node:

./deploy.sh 2>&1 | tee ./deploy.log

Note: the log generated by this command will be located in the main.ton.dev/scripts/ folder and can be useful for troubleshooting.

Wait until the node is synced. Depending on network throughput this step may take significant time (up to several hours).

4. Check Node synchronization

Use the following command to check if the node is synced:

docker exec -it rnode /ton-node/tools/console -C /ton-node/configs/console.json --cmd getstats

Script output example:

tonlabs console 0.1.254
COMMIT_ID: 
BUILD_DATE: 2021-12-24 10:53:20 +0300
COMMIT_DATE: 
GIT_BRANCH: 
{
    "sync_status":    "synchronization finished",
    "masterchainblocktime":    1640343840,
    "masterchainblocknumber":    13393489,
    "timediff":    4,
    "in_current_vset_p34":    true,
    "in_next_vset_p36":    false,
    "last_applied_masterchain_block_id":    {"shard":"-1:8000000000000000","seq_no":13393489,"rh":"e74d505222bbe64617bbd42939cf01334b035990ae4b4e285e67ecbb1b537dd3","fh":"fc0be7c22310389400a7bdbd000b3737b317ce44d92e9a5ed72086c4fa404afa"},
    "processed_workchain":    0,
    "validation_stats":    {
  "-1:8000000000000000": "1 sec ago",
  "0:e800000000000000": "1 sec ago"
},
    "collation_stats":    {
  "-1:8000000000000000": "never",
  "0:e800000000000000": "6 sec ago"
},
    "tps_10":    2,
    "tps_300":    1
}

If the timediff parameter is less than 10 seconds, synchronization with masterchain is complete. "sync_status": "synchronization finished" means synchronization with workchains is complete

Note: The sync process may not start for up to one hour after node deployment, during which this command may result in error messages. If errors persist for more than an hour after deployment, review deployment log for errors and check the network status.

5. Configure validator multisignature wallet

There is a small difference between direct staking and DePool validators on this step:

• For direct staking validator it is necessary to create and deploy a validator SafeMultisig wallet in -1 chain.
• For a DePool validator it is necessary to create and deploy a validator SafeMultisig wallet in 0 chain.

You can use TONOS-CLI for this purpose. It should be configured to connect to the main.ton.dev network.

Refer to this document for the detailed wallet creation procedure, or follow the links in the short guide below:

1. All wallet custodians should create seed phrases and public keys for themselves. At least three custodians are recommended for validator wallet, one of which will be used by the validator node. All seed phrases should be kept secret by their owners and securely backed up.
2. The wallet deployer (who may or may not be one of the custodians) should gather the public keys from all custodians.
3. The wallet deployer should obtain SafeMultisig contract code from the repository.
4. The wallet deployer should generate deployment keys.
5. The wallet deployer should generate validator wallet address: in -1 chain for direct staking validator or in 0 chain for a DePool validator.
6. Any user should send at least 1 token to the generated wallet address to create it in the blockchain.
7. The wallet deployer should deploy the wallet contact to the blockchain and set all gathered public keys as its custodians. At this step the number of custodian signatures required to make transactions from the wallet is also set (>=2 recommended for validator wallets). Deploy to -1 chain for direct staking validator or to 0 chain for a DePool validator.
8. In case of direct staking, the funds for staking should be transferred to the newly created validator wallet.

Once the wallet is deployed, place 2 files on the validator node:

• /ton-node/configs/${VALIDATOR_NAME}.addr should contain validator multisignature wallet address in form X:XXX...XXX (the folder on the host is main.ton.dev/docker-compose/ton-node/configs)
• /ton-node/configs/keys/msig.keys.json should contain validator multisignature custodian's keypair (the folder on the host is main.ton.dev/docker-compose/ton-node/configs/keys/)

The node will use the wallet address and the keys provided to it to generate election requests each validation cycle.

Note: If the validator wallet requires more than 1 custodian signature to make transactions, make sure each transaction sent by the validator node is confirmed by the required amount of custodians.

6. Configure DePool

For a DePool validator it is necessary to deploy a DePool contract to workchain 0.

You can use TONOS-CLI for this purpose. It should be configured to connect to the main.ton.dev network.

Refer to this document for the detailed DePool creation procedure, or follow the links in the short guide below:

1. Obtain contract code from the repository.
2. Generate deployment keys.
3. Calculate contract addresses.
4. Send tokens to the calculated addresses.
5. Deploy contracts. Make sure to specify your validator wallet in the DePool contract at this step.
6. Configure DePool state update method.

Once DePool is successfully deployed and configured to be regularly called to update its state, you can make stakes in it. Note that validator stakes must always exceed validator assurance, otherwise DePool will not participate in elections.

Also note, that DePool and supporting contracts balance should be monitored and kept positive at all times.

Once the validator wallet and the DePool are deployed, place 3 files on the validator node:

• /ton-node/configs/${VALIDATOR_NAME}.addr should contain validator multisignature wallet address in form 0:XXX...XXX (the folder on the host is main.ton.dev/docker-compose/ton-node/configs)
• /ton-node/configs/keys/msig.keys.json should contain validator multisignature custodian's keypair (the folder on the host is main.ton.dev/docker-compose/ton-node/configs/keys/)
• /ton-node/configs/depool.addr should contain DePool address in form 0:XXX...XXX (the folder on the host is main.ton.dev/docker-compose/ton-node/configs)

The script generating validator election requests (directly through multisig wallet, or through DePool, depending on the setting selected on step 2.1) will run regularly, once the necessary addresses and keys are provided.

7. Upgrade RUST Validator Node

Note: You may need to renew your copy of main.ton.dev scripts but do not remove any working files from the previous deployment (for example, configs folder).

Adjust (specify new commit ID) main.ton.dev/scripts/env.sh:

export TON_NODE_GITHUB_REPO="https://github.com/tonlabs/ton-labs-node.git"
export TON_NODE_GITHUB_COMMIT_ID="master"
export TON_NODE_TOOLS_GITHUB_REPO="https://github.com/tonlabs/ton-labs-node-tools.git"
export TON_NODE_TOOLS_GITHUB_COMMIT_ID="master"
export TONOS_CLI_GITHUB_REPO="https://github.com/tonlabs/tonos-cli.git"
export TONOS_CLI_GITHUB_COMMIT_ID="master"

Upgrade the node:

./upgrade.sh 2>&1 | tee ./upgrade.log

Note: the log generated by this command will be located in the main.ton.dev/scripts/ folder and can be useful for troubleshooting.

Wait until the node is synced.

Stopping, restarting the RUST Node

Note: call docker-compose commands from the main.ton.dev/docker-compose/ton-node folder.

To stop the node use the following command:

docker-compose stop

To restart a stopped node use the following command:

docker-compose restart

Logging

During deployment

It is highly recommended to record the full log during node deployment:

./deploy.sh 2>&1 | tee ./deploy.log

The log is saved to the main.ton.dev/scripts/ folder next to the deployment script and can be useful for troubleshooting.

During operation

When operational, the node keeps a number of logs in the main.ton.dev/docker-compose/ton-node/logs folder.

Logs are generated with log4rs framework. For detailed documentation on it refer to https://docs.rs/log4rs/1.0.0/log4rs/.

Logging configuration is determined by the main.ton.dev/docker-compose/ton-node/configs/log_cfg.yml file. By default is contains the recommended configuration for the Rust node.

refresh_rate: 30 seconds

appenders:
  stdout:
    kind: console
    encoder:
      pattern: "{d(%s.%f)} {l} [{h({t})}] {I}: {m}{n}"

  stdout_ref:
    kind: console
    encoder:
      pattern: "{f}:{L} {l} [{h({t})}] {I}: {m}{n}"

  logfile:
    kind: file
    path: "/ton-node/logs/output.log"
    encoder:
      pattern: "{d(%s.%f)} {l} [{h({t})}] {I}: {m}{n}"

  rolling_logfile:
    kind: rolling_file
    encoder:
      pattern: "{d(%Y-%m-%d %H:%M:%S.%f)} {l} [{h({t})}] {I}: {m}{n}"
    path: /ton-node/logs/output.log
    policy:
      kind: compound
      trigger:
        kind: size
        limit: 50 gb
      roller:
        kind: fixed_window
        pattern: '/ton-node/logs/output_{}.log'
        base: 1
        count: 1

  tvm_logfile:
    kind: file
    path: "target/log/tvm.log"
    encoder:
      pattern: "{m}{n}"

root:
  level: info
  appenders:
    - rolling_logfile

loggers:
  # node messages
  ton_node:
    level: trace
  boot:
    level: trace
  sync:
    level: trace

  # adnl messages
  adnl:
    level: info

  overlay:
    level: info

  rldp:
    level: info

  dht:
    level: info

  # block messages
  ton_block:
    level: debug

  # block messages
  executor:
    level: debug

  # tvm messages
  tvm:
    level: info

  librdkafka:
    level: info

  validator:
    level: debug

  catchain:
    level: debug

  validator_session:
    level: debug

The currently configured targets are the following:

ton_node: node-related messages, except initial boot and sync, block exchange with other nodes

boot: initial boot messages, creation of trusted key block chain, loading blockchain state

sync: node synchronization - loading a certain number of most recent blocks

adnl: messages of the ADNL protocol

overlay: messages of the overlay protocol

rldp: messages of the RLDP protocol

dht: messages of the DHT protocol

ton_block: messages of the block structures library, logs are turned on in debug

executor: messages of the smart contract execution library, logs are turned on in debug

tvm: ton virtual machine messages, logs are turned on in debug

librdkafka: kafka client library messages

validator: top level consensus protocol messages

catchain: low level consensus protocol messages

validator_session: mid level consensus protocol messages

Migrating from C++ node

To migrate your validator from legacy C++ node to Rust node, complete the following steps:

1. Set up a new host for the Rust node, according to steps 1-3 of this document.
2. Wait for node to sync. Check sync according to step 4 of this document.
3. Stop the C++ node sending election requests (by default - disable scheduling of the validator script). Do not shut down the C++ validator itself, let it finish the current round.
4. Configure validator wallet and corresponding keys, optionally - DePool (copy them from C++ node files to Rust Node files). By default:
   1. Copy validator wallet address from ~/ton-keys/$(hostname -s).addr file on the C++ node to /ton-node/configs/${VALIDATOR_NAME}.addr on the Rust Node.
   2. Copy validator wallet keys from /ton-keys/msig.keys.json on the C++ node to /ton-node/configs/keys/msig.keys.json on the Rust Node.
   3. If you have a DePool set up, copy DePool address from ~/ton-keys/depool.addr on the C++ node to /ton-node/configs/depool.addr on the Rust Node. Once this is done, the Rust node validator script will start automatically with the next round.
5. DO NOT shut down the C++ host. Wait for the C++ node to complete the current round until the validator set is changed.
6. Check logs/validator.log on the Rust node, and make sure the first election request was successfully sent. There should be no errors in the log.
7. Only once the validator set changes, the C++ node is no longer a validator and the Rust node starts validating (validation_stats and collation_stats in the console output should not be empty), shut down the C++ node.

Troubleshooting

Here are some solutions to frequently encountered problems.

1. Couldn’t connect to Docker daemon at http+docker://localhost

This error occurs in two cases. Either the docker daemon isn't running, or current user doesn't have rights to access docker.

You can fix the rights issue either by running relevant commands as the superuser or adding the user to the docker group:

sudo usermod -a -G docker $USER

Make sure to restart the system or log out and back in, for the new group settings to take effect.

2. thread 'main' panicked error when checking node synchronization

The following error may occur for a short time immediately after node deployment when attempting to check synchronization:

thread 'main' panicked at 'Can't create client: Os { code: 111, kind: ConnectionRefused, message: "Connection refused" }', bin/console.rs:454:59

Currently this is expected behavior, unless it persists for more than a few minutes. If it does persist, check network status at https://ton.live/, and, if the network is up and running, review deployment logs for errors.

3. Error executing command when checking node synchronization

The following error may occur for up to an hour after node deployment when attempting to check synchronization:

Error executing command: Error receiving answer: early eof bin/console.rs:296

Currently this is expected behavior, unless it persists for more than one hour. If it does persist, check network status at https://ton.live/, and, if the network is up and running, review deployment logs for errors.

4. Cannot stop/restart/remove node container

Make sure you are running all docker-compose commands from the main.ton.dev/docker-compose/ton-node folder.

5. DePool state not updating

It's recommended to send at least two ticktocks while the elections are open. For rust node you can use the provided ticktock script, which sends 5 ticktocks after the elections open.