# Spin up a validator {% hint style="warning" %} **CAUTION** Validators are responsible for the network stability, it is very important to be able to react at any time of the day or night in case of trouble. We strongly encourage validators to set up a monitoring and alerting system, learn more about this from our secure setup guide. {% endhint %} ### Start the AI Service #### Prerequisite Must have conda installed, follow the instruction [here](https://docs.conda.io/projects/conda/en/latest/user-guide/install/index.html). Clone the [uomi-node-ai](https://github.com/Uomi-network/uomi-node-ai/) repo ### Installation Steps #### 1. Set up Conda Environment First, open a terminal and create a new conda environment: ```bash conda create -n uomi-ai python=3.10 -y ``` If you've just installed Miniconda/Anaconda, initialize conda in your shell: ```bash # For bash users source ~/miniconda3/etc/profile.d/conda.sh # For zsh users source ~/miniconda3/etc/profile.d/conda.sh # Activate the environment conda activate uomi-ai ``` #### 2. Install PyTorch with CUDA Support Install PyTorch and related packages: ```bash conda install pytorch==2.2.0 torchvision==0.17.0 torchaudio==2.2.0 pytorch-cuda=12.1 -c pytorch -c nvidia ``` Verify the installation: ```bash python -c "import torch; print(torch.cuda.is_available())" ``` This should print `True` if CUDA is properly installed. #### 3. Install CUDA Development Tools Install CUDA toolkit and NVCC compiler: ```bash conda install -c nvidia/label/cuda-12.1.0 cuda-nvcc=12.1 cuda-toolkit=12.1 ``` Verify the installation: ```bash nvcc --version ``` #### 4. Install Additional Dependencies Install the required Python packages: ```bash # Install Optimum pip install -U "optimum>=1.20.0" # Install AutoGPTQ pip install auto-gptq --no-build-isolation # Install Transformers pip install transformers # Install specific NumPy version conda install numpy=1.24 # Install Flash Attention pip install flash-attn --no-build-isolation # Install Flask pip install flask ``` ### Troubleshooting #### Common Issues 1. **CUDA Not Found** * Verify NVIDIA drivers are installed: `nvidia-smi` * Check CUDA installation: `nvcc --version` * Ensure PyTorch CUDA is properly installed: `python -c "import torch; print(torch.version.cuda)"` 2. **Build Failures** * Ensure you have build tools installed: ```bash sudo apt-get update sudo apt-get install build-essential ``` * For Auto-GPTQ issues, try installing from source: ```bash git clone https://github.com/AutoGPTQ/AutoGPTQ.git cd AutoGPTQ rm -rf build/* python setup.py install ``` 3. **Version Conflicts** * If you encounter package conflicts, try creating a fresh environment * Consider using `pip install --no-deps` for problematic packages #### Verification To verify the complete setup, run this test script: ```python import torch import transformers import auto_gptq import flash_attn import flask print(f"PyTorch version: {torch.__version__}") print(f"CUDA available: {torch.cuda.is_available()}") print(f"CUDA version: {torch.version.cuda}") print(f"Transformers version: {transformers.__version__}") ``` ### Create the service 1. Create a new systemd service file for the AI component: ```bash sudo nano /etc/systemd/system/uomi-ai.service ``` 2. Add the following content to the service file: ```ini [Unit] Description=UOMI AI API Server After=network.target [Service] User=uomi WorkingDirectory=/home/uomi/uomi-node-ai ExecStart=/bin/bash -c "source /home/uomi/miniconda3/etc/profile.d/conda.sh && conda activate uomi-ai && python3 uomi-ai.py" Restart=always RestartSec=10 TimeoutSec=30 StartLimitIntervalSec=500 StartLimitBurst=5 [Install] WantedBy=multi-user.target ``` **Important Notes:** * The `WorkingDirectory` path (`/home/uomi/uomi-node-ai`) should be adjusted to match your actual installation directory * The `ExecStart` command assumes: * Miniconda is installed in `/home/uomi/miniconda3` * A conda environment named `uomi-ai` exists * The main Python script is named `uomi-ai.py` * Modify these paths and names according to your specific setup 3. Enable and start the AI service: ```bash sudo systemctl daemon-reload sudo systemctl enable uomi-ai.service sudo systemctl start uomi-ai.service ``` 4. Verify the service is running: ```bash sudo systemctl status uomi-ai.service ``` 5. Monitor the AI service logs: ```bash sudo journalctl -u uomi-ai.service -f ``` ## Set-up the validator Node ### Download the binary Make sure to download the uomi bin and genesis file on your machine from: ### Directory Structure Setup Create the necessary directories and set appropriate permissions: ```bash # Create directories sudo mkdir -p /var/lib/uomi sudo mkdir -p /usr/local/bin # Create user for the service sudo useradd --no-create-home --shell /usr/sbin/nologin uomi # Set permissions sudo chown -R uomi:uomi /var/lib/uomi ``` ### Binary and Genesis Setup Get available peers at Add copied peers inside the genesis.json file: ```json { "name": "Uomi", "id": "uomi", "chainType": "Live", "bootNodes": [PASTE_PEERS_ARRAY_HERE], "telemetryEndpoints": null, "protocolId": null, "properties": { "tokenDecimals": 18, "tokenSymbol": "UOMI" }, ... ``` Then install binary and genesis files: 1. Copy the Uomi binary to the correct location: ```bash sudo cp ./uomi /usr/local/bin/uomi sudo chmod +x /usr/local/bin/uomi ``` 2. Copy the genesis file: ```bash sudo cp ./genesis.json /usr/local/bin/genesis.json ``` #### Create validator account This account is the one that will validate blocks and will save output from Agents ``` /usr/local/bin/uomi key generate --scheme Sr25519 ``` You will get something like: ``` Secret phrase: tackle rebuild neither turn degree real capital armed giraffe novel resist human Network ID: substrate Secret seed: 0x5c9499827e606bcf284e8a650a96ce13ebf0484bd64a280507ff96d99da6e174 Public key (hex): 0x14ceec080a6aa464b4235dd951a85669d25e4fa659578f01a7a7dc5c95454931 Account ID: 0x14ceec080a6aa464b4235dd951a85669d25e4fa659578f01a7a7dc5c95454931 Public key (SS58): 5CXzHQ3DmYKemRAyY6NPyKsDWiChbeeKtxW3q5RWpmdUQvdR SS58 Address: 5CXzHQ3DmYKemRAyY6NPyKsDWiChbeeKtxW3q5RWpmdUQvdR ``` Generate Ed25519 key for GRANDPA using the same seed phrase: ```bash /usr/local/bin/uomi key inspect --scheme Ed25519 "YOUR_SEED_PHRASE" ``` you'll need the secret phrase and the secret seed later. {% hint style="warning" %} **CAUTION** Save your secret phrase and secret seed in a safe place (preferably offline), it's the only way to recover your account if you lose it {% endhint %} ### Systemd Service Configuration Create a systemd service file at `/etc/systemd/system/uomi.service`: ```ini iniCopy[Unit] Description=Uomi Node After=network.target StartLimitIntervalSec=0 [Service] Type=simple User=uomi Group=uomi Restart=always RestartSec=10 LimitNOFILE=65535 ExecStart=/usr/local/bin/uomi \ --validator \ --name "YOUR_NODE_NAME" \ --chain "/usr/local/bin/genesis.json" \ --base-path "/var/lib/uomi" \ --state-pruning 1000 \ --blocks-pruning 1000 \ --enable-evm-rpc \ --rpc-cors all \ --prometheus-external \ --telemetry-url "wss://telemetry.polkadot.io/submit/ 0" # Hardening ProtectSystem=strict PrivateTmp=true PrivateDevices=true NoNewPrivileges=true ReadWritePaths=/var/lib/uomi [Install] WantedBy=multi-user.target ``` ### Enable and start the service: ```bash sudo systemctl daemon-reload sudo systemctl enable uomi.service sudo systemctl start uomi.service ``` ### Add your key to your node Once your node is running, insert your keys: {% hint style="warning" %} **ATTENTION**\ Replace ***SECRET\_PHRASE*** and ***PUBLIC\_KEY*** with your data received with the first command "key generate **--scheme Sr25519**" {% endhint %} ```bash for key_type in babe imon uomi ipfs; do curl -H "Content-Type: application/json" \ -d '{"id":1, "jsonrpc":"2.0", "method":"author_insertKey", "params":["'$key_type'", "SECRET_PHRASE", "PUBLIC_KEY"]}' \ http://localhost:9944 done ``` {% hint style="warning" %} **ATTENTION**\ Replace ***SECRET\_PHRASE*** and ***PUBLIC\_KEY*** with your data received with the first command "key inspect **--scheme Ed25519**" {% endhint %} ```bash curl -H "Content-Type: application/json" \ -d '{"id":1, "jsonrpc":"2.0", "method":"author_insertKey", "params":["gran", "SECRET_PHRASE", "PUBLIC_KEY"]}' \ http://localhost:9944 ``` This data will be stored on your local node, no one can have access to them. ### Add the account Now you can import the account you created previously with SECRET\_PHRASE to your wallet extension and fund it with the tokens you want to bond #### Verify synchronization[​](https://docs.astar.network/docs/build/nodes/collator/spinup_collator#verify-synchronization) Before jumping to the next steps, you have to wait until your node is **fully synchronized**. This can take a long time depending on the chain height. Check the current synchronization: ``` journalctl -f -u uomi -n100 ``` ### Session Keys[​](https://docs.astar.network/docs/build/nodes/collator/spinup_collator#session-keys) **Author session keys**[**​**](https://docs.astar.network/docs/build/nodes/collator/spinup_collator#author-session-keys) Run the following command to author session keys: ```bash curl -H "Content-Type: application/json" -d '{"id":1, "jsonrpc":"2.0", "method": "author_rotateKeys", "params":[]}' http://localhost:9944 ``` The result will look like this (you just need to copy the result): ``` {"jsonrpc":"2.0","result":"0x600e6cea49bdeaab301e9e03215c0bcebab3cafa608fe3b8fb6b07a820386048","id":1} ``` **Register as validator** Go to the [Polkadot.js portal](https://subexp.uomi.ai) and connect to the network. Go to ***Network > Staking > Accounts*** and then select **Validator** on the right
Select the account you funded previously, the amount you want to bond and the payment method and click **next.** Now insert the key you copied before and follow the instructions, click "bond & validate" and then submit the transaction. {% hint style="info" %} **INFO**\ Onboarding takes place at `n+1` session. {% endhint %} You can now [set your identity](https://docs.uomi.ai/build/run-a-node/become-a-validator/set-your-identity). --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.uomi.ai/build/run-a-node/become-a-validator/spin-up-a-validator.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.