NVIDIA / NVIDIA/Personal-AI-Router

[Feature]: Multiple ollama instances on single host support

Open
#77 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Go
Stars
1.4k
Forks
250
Avg merge
23h 27m
Merged PRs (30d)
1

Description

Area

Routing and scheduling

User problem

I have a system with two GPUs from different vendors, one being a nvidia GPU. I've had to setup two ollama instances locally on different ports to get this to work with different settings. I'd like to be able to expose both ollama instances to the cluster

Desired outcome

Both ollama instances show up for a node (or it appears as two nodes)

Alternatives considered

No response

Compatibility and security implications

No response

Validation approach

Launch two ollama instances on different ports, ideally with specific GPUs enabled for each

Confirmations
  • I searched existing issues for duplicates.
  • I agree to follow the Code of Conduct.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the routing and scheduling entry points that discover and expose Ollama instances. Trace how a node's Ollama endpoint and configuration are represented, then determine whether two instances should be exposed as separate nodes or as endpoints on one node. Validate by launching two Ollama instances on different ports with separate GPU settings and confirming both appear in the cluster.

Written by the indexing model from the issue text.

Assessment

Tech stack
go, ollama
Domain
ai, backend
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
52/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.