M1 Mac执行docker-compose run报错：无法创建新OS线程

I’ve run into this exact thread-limiting issue on my M1 Mac when running AMD64 containers, so I know how frustrating it can be. The error happens because the QEMU emulation layer (used to run x86/amd64 images on ARM) has strict thread limits by default, and your existing fixes only covered part of the problem. Let’s walk through the targeted fixes that should resolve this:

• Bump Docker Desktop Resource Allocation
  By default, Docker Desktop on M1 allocates minimal CPU and memory to emulated containers, which can trigger thread creation failures. Go to Docker Desktop > Settings > Resources > Advanced, set CPU to at least 4 cores and Memory to 8GB (or higher if you can), then restart Docker. This gives the emulation layer enough headroom to spawn threads.

• Add Platform Configuration to docker-compose.yml
  You added --platform=linux/amd64 to your Dockerfile, but you also need to enforce this at the compose level to ensure every service uses the correct emulation context. Update your docker-compose.yml:

  services:
    your-service-name:
      build: .
      platform: linux/amd64
      # Keep your other service configs here
  

• Reset QEMU Binfmt Configuration
  Sometimes the QEMU emulation state gets corrupted. Run these commands to reset it, then restart Docker Desktop:

  docker run --privileged --rm tonistiigi/binfmt --uninstall all
  docker run --privileged --rm tonistiigi/binfmt --install amd64
  

• Increase Container Thread Limits
  The error explicitly mentions hitting a thread count cap. Add ulimit settings to your service in docker-compose.yml to raise the limits:

  services:
    your-service-name:
      # Other configs
      ulimits:
        nproc: 65535
        nofile:
          soft: 65535
          hard: 65535
  

• Use Rosetta for Emulation (Recommended)
  Docker Desktop has a Rosetta integration that’s far more efficient than QEMU for x86/amd64 containers on M1. Enable it by going to Docker Desktop > Settings > Features in Development, check "Use Rosetta for x86/amd64 emulation on Apple Silicon", then restart Docker. This drastically reduces thread-related bottlenecks.

Try these steps in order—most likely the combination of resource allocation and Rosetta emulation will fix the issue right away.

内容的提问来源于stack exchange，提问作者VasiliiSikora