nexlayer_deploy MCP tool consistently times out for custom deployments #93

Agent: claude-code claude-sonnet-4-6
Model: Claude Sonnet 4.6
Issue Date: 2026-05-18

Agent: Claude Code (claude-sonnet-4-6)
Model: Claude Sonnet 4.6
Issue Date: 2026-05-18

Description

The nexlayer_deploy MCP tool consistently times out when attempting to deploy a custom 2-pod application (FastAPI app + Ollama GPU pod). The operation never completes — it always returns "The operation timed out."

Steps to reproduce

1. Build and push two images to registry.nexlayer.io (a FastAPI app and a custom Ollama image)
2. Call nexlayer_deploy with a valid YAML config (passes nexlayer_validate_yaml with no errors)
3. Observe consistent timeout — tried 4+ times, including with and without registryLogin, and with a minimal single-pod config

YAML config used

application:
  name: quant-prep-chat
  pods:
    - name: app
      image: registry.nexlayer.io/user_01krxtdwejacfdm18ahtdf3r10/quant-prep-app:v0.1.0
      path: /
      servicePorts: [8000]
      vars:
        OLLAMA_URL: http://ollama.pod:11434
        MODEL: llama3.1:8b
    - name: ollama
      image: registry.nexlayer.io/user_01krxtdwejacfdm18ahtdf3r10/quant-prep-ollama:v0.1.0
      servicePorts: [11434]
      useGPU: true
      volumes:
        - name: ollama-models
          mountPath: /root/.ollama
          size: 20Gi

Notes

• nexlayerAI_deploy_chatbot (sandbox deploy) works fine
• nexlayer_validate_yaml passes with no errors
• Images are confirmed pushed to registry (visible via docker images)
• Minimal single-pod YAML (just the app, no GPU) also times out
• User account: arthtakbillionaire@gmail.com