We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Wayfinder-Foundry/gdal-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
0009-concurrency-and-memory-model.md•580 B
# ADR-0009-concurrency-and-memory-model.md
## Title
Concurrency, memory, and dataset lifecycle
## Status
Proposed
## Context
GDAL isn’t fully thread-safe; rasters can be large.
## Decision
* Use **windowed IO** and VRT composition to avoid large in-memory arrays.
* Prefer **`anyio.to_thread` / process pools** for CPU-heavy work; avoid blocking the event loop.
* **Always** close datasets promptly; encapsulate IO in context managers.
## Consequences
* Pro: Stable under load; bounded memory; responsive server.
* Con: Slight complexity managing pools and lifecycle.