Bug: DockerDeployment.start() times out when run in parallel with asyncio.gather

[vgakhreja-google]

When starting two DockerDeployment containers concurrently using asyncio.gather, the deployment.start() method hangs and eventually fails with a timeout.

This is a simple script to reproduce the issue:

import asyncio
import logging
from swerex.deployment.docker import DockerDeployment

logger = logging.getLogger(__name__)

async def start_container(name: str):
    logger.info(f"[{name}] Initializing deployment...")
    
    deployment = DockerDeployment(
        image="android-bench-env",
        docker_args=["--privileged"]
    )
    
    try:
        logger.info(f"[{name}] CALLING 'await deployment.start()'...")
        await deployment.start()
        logger.info(f"--- [{name}] SUCCEEDED ---")

    except Exception as e:
        logger.exception(f"--- [{name}] FAILED TO START ---")
    
    finally:
        logger.info(f"[{name}] Stopping deployment...")
        try:
            await deployment.stop()
        except Exception as e:
            logger.error(f"[{name}] Failed to stop deployment", exc_info=True)


async def main():
    logger.info("--- Starting two container deployments in parallel ---")
    
    await asyncio.gather(
        start_container("Container-1"),
        start_container("Container-2"),
        return_exceptions=True
    )
    
    logger.info("--- Both tasks finished. ---")

if __name__ == "__main__":
    asyncio.run(main())