Skip to main content

Telos EVM Docker

Requirements

  • git
  • docker
  • python3

Quickstart

## Need installed python version >= python 3.9

git clone https://github.com/telosnetwork/telos-evm-docker.git -b v1.5.0-rc8
cd telos-evm-docker

# Two posibilities:
# 1) Recommended: Install poetry python package manager dependecy to /usr/local (REQUIRES SUDO)
sudo ./install.sh

# 2) Install dependency to another $DIRECTORY that is already on $PATH
./install.sh $DIRECTORY

source ./activate.sh
## from now on we are working in the virtual environment.
## just run 'deactivate' when you want to deactivate virtual environment.

tevmc init testnet
cd testnet

## Modify tevmc.json
## For mainnet, signer_account, signer_permission, signer_key
## Containers use ports of host machine so you need to check the ports if used already.
tevmc build

## Launch and store logs, also detach from session
nohup tevmc up 2>&1 > tevmc.log &

## Wait until synced and evm-rpc activated.
## To stop:
tevmc down

Installation

tevmc is a python installable package, and its currently available on github.

  1. git clone git@github.com:telosnetwork/telos-evm-docker.git
  2. cd telos-evm-docker
  3. git checkout v1.5.0-rc8
  4. sudo ./install.sh

To load the virtual environment (and make tevmc command line available):

source ./activate.sh

tevmc can be used as a command line utiliy to manually manage Telos EVM nodes, but is usable as a library.

Once the package installation finishes, you should be able to invoke:

  • tevmc --help

Initialization

tevmc uses a specific self-contained directory structure for each individual node, to create a new node directory from template use the init sub command.

  • tevmc init {node name}

There are three templates available:

  • local: Local development testnet
  • testnet: Standard testnet node
  • mainnet: Standard mainnet node

If your node name contains any of this names the corresponding template will be selected. Examples:

  • tevmc init local-dev-1 # will use local template
  • tevmc init us-testnet-04 # will use testnet template
  • tevmc init rpc.mainnet-07 # will use mainnet template

Configuration

Inside the directory created by the init command you'll see a docker directory and the unified config file tevmc.json.

From the newly created node directory run:

tevmc build

The first time you build the images tevmc will populate all the different config files in the docker directory based of the values in the unified config file. After that it will only re-create the files from templates if it detects changes on tevmc.json.

Bootstrap

With the docker images built we can launch the node:

tevmc up # launch as foreground process Ctrl+C to exit

nohup tevmc up 2>&1 > tevmc.log & # launch as background process

You should eventually see control point reached! in the logs, at that point the full node stack should be up and ready to serve requests.

Where is the data?

The idea is to create self contained node directories, which can be easily moved with the stack down.

Each sub-directory inside main docker path generally contains a combination of this directories:

  • build docker build context
  • config editable configuration files, mounted to were they are needed inside the respective container.
  • data permanent storage kept between runs
  • logs log file location

Full directory structure

node/
├── docker/
│   ├── elasticsearch/
│   ├── leap/
│   ├── telosevm-translator/
│   ├── telos-evm-rpc/
│   ├── kibana/
│   └── redis/
├── logs/
└── tevmc.json

Landing

To bring the stack down run:

tevmc down

This will perform graceful nodeos stop and bring down all containers.

FAQ

While in the process of launching the nodeos service, it crashes with an error like this one:

warn  2023-08-20T19:22:30.845 nodeos    chain_plugin.cpp:1085         plugin_initialize    ] 13 N5boost10wrapexceptINSt3__112sy
stem_errorEEE: Database dirty flag set
rethrow Database dirty flag set:
    {"what":"Database dirty flag set"}
    nodeos  chain_plugin.cpp:1085 plugin_initialize

error 2023-08-20T19:22:30.845 nodeos    main.cpp:157                  main                 ] database dirty flag set (likely du
e to unclean shutdown): replay required

Usually means nodeos data is corrupted due to an ungraceful stop. Run tevmc repair from the node directory with the stack down to repair data.

Can I start indexing from a custom snapshot?

Yes set the snapshot config on nodeos section and also start_block on evm-rpc section.

Stack is correctly launched but when running transactions I get an error like this one on hyperion-api logs:

2022-02-10T01:15:37: In raw, tx is: f86c80....
2022-02-10T01:15:37: transaction declares authority 
'{"actor":"rpc.evm","permission":"rpc"}', but does not have signatures for it.
2022-02-10T01:15:37: RPCERROR: 2022-02-10T01:15:37.913Z -
{"error":{"code":3,"message":"transaction declares authority...

Setup correct signing account credentials on rpc service configuration.

Is kubernetes supported?

No