onboarding

Last updated: 2025-02-19 01:11:11.426005 File source: link on GitLab

Onboarding as a Compute Provider

Overview

As a compute provider in the NuNet network, you can offer your computer's resources (CPU, RAM, GPU, and storage) to other network participants.

This guide will walk you through the process of onboarding your resources to the network.

Prerequisites

Before onboarding, ensure:

Onboarding Process

Onboard

Use the following command to onboard your resources:

nunet actor cmd --context user /dms/node/onboarding/onboard --ram <GB> --cpu <cores> --disk <GB>

Example:

# Onboard 4GB RAM, 2 CPU cores, and 20GB disk space
nunet actor cmd --context user /dms/node/onboarding/onboard --ram 4 --cpu 2 --disk 20

If a GPU is detected on the machine, an interactive prompt will be displayed that allows choosing the GPU and the amount of VRAM to onboard from it.

To onboard a GPU without the interactive prompt, --gpus "<GPU_INDEX>:<VRAM_IN_GB>" can be used where index can be obtained from the gpu list command. In case of multiple GPUs, the pair can be expanded for each GPU separated by a comma.

Example:

# Onboard 4GB RAM, 2 CPU cores, 20GB disk space and 2GB VRAM of GPU at index 0
nunet actor cmd --context user /dms/node/onboarding/onboard --ram 4 --cpu 2 --disk 20 --gpus "0:2"

If GPU is detected on the machine but it shouldn't be onboarded, use the --no-gpu flag.

Managing Your Resources

Verify Onboarding Status

Check your onboarding status:

nunet actor cmd --context user /dms/node/onboarding/status

Checking Resource Status

Monitor your resources allocation:

# Check allocated resources
nunet actor cmd --context user /dms/node/resources/allocated

# Check onboarded resources
nunet actor cmd --context user /dms/node/resources/onboarded

Offboarding Resources

To remove the availability of your machine's resources from the network, execute:

# Offboard current resources
nunet actor cmd --context user /dms/node/onboarding/offboard

Modifying Onboarded Resources

Note: we'll facilitate this process in the future so that users do not have to reonboard again.

To modify your resource allocation:

  1. First, offboard your current resources

  2. Then onboard again with new resource values

# Offboard current resources
nunet actor cmd --context user /dms/node/onboarding/offboard

# Onboard with new values
nunet actor cmd --context user /dms/node/onboarding/onboard --ram <new_GB> --cpu <new_cores> --disk <new_GB>

Best Practices

  1. Resource Allocation

    • Don't onboard all available resources

    • Leave enough resources for system operations

    • Consider your system's stability and cooling capabilities

  2. System Maintenance

    • Regularly update your DMS instance

    • Monitor system health and performance

    • Maintain stable internet connectivity

  3. Security

    • Monitor system logs for unusual activity

    • Keep your system updated with security patches

Troubleshooting

Common issues and solutions:

  1. Onboarding fails

    • Verify system requirements are met

    • Ensure proper permissions are set

  2. Resource allocation issues

    • Verify if other allocations are still running (check allocated resources)

    • Ensure proper GPU drivers if using GPUs

    • Verify available resources

    • Check for competing processes

Last updated