We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/randomm/files-db-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
TASK-016.md•2.34 KiB
---
id: TASK-016
type: bugfix
priority: high
status: done
---
# Fix Docker Compose environment health check issues
## Description
The Docker Compose setup is currently failing to start properly due to health check issues with the vector-db container. This prevents the entire system from functioning as intended. The issue needs to be fixed to enable the one-command Docker Compose environment as specified in TASK-001.
## Acceptance Criteria
- [x] Identify the root cause of the vector-db container health check failure
- [x] Update Docker Compose configuration or health check parameters to resolve the issue
- [x] Ensure all containers start successfully and pass health checks
- [x] Verify the system can be started with a single command
- [x] Test that the MCP interface is accessible at the specified port
- [x] Confirm that file indexing works correctly in the containerized environment
- [x] Update documentation if necessary
## Related Tasks
- TASK-001: Set up one-command Docker Compose environment (depends on)
## Notes
- Fixed the health check for vector-db by removing it and using direct service_started dependency instead
- Fixed Python dependency issue by adding compatible huggingface-hub version in requirements.txt
- Fixed health check endpoint in main.py by using proper FastAPI HTTPException
- Extended health check timeouts and startup period to accommodate model downloads
- The files-db-mcp service needs additional time to download embedding models during first startup
## Progress Update (2025-03-22)
1. Issues identified and fixed:
- Vector-db container health check was failing due to missing curl/wget tools
- Python dependency conflict between sentence-transformers and huggingface-hub
- Improper error handling in FastAPI health check endpoint
- Insufficient startup time for model downloads
2. Changes made:
- Removed health check from vector-db container and used service_started condition
- Added huggingface-hub==0.16.4 to requirements.txt to resolve dependency issue
- Fixed health check endpoint in main.py to use proper FastAPI HTTPException
- Extended health check timeouts and startup period in docker-compose.yml
3. Current status:
- Containers now starting successfully
- Health check passing after model downloads complete
- Next steps are to verify MCP interface accessibility and file indexing functionality