Workers should not fail with exit status !=0 when coordinator/master refuses to respond

Have to handle the case of coordinator successfully exiting such that workers also exits successfully instead of failing with rpc error.

Look at the worker and coordinator/master communication code, likely in files named worker.go and coordinator.go. Understand the RPC error handling when the coordinator exits. The goal is to modify the worker's exit logic to succeed when the coordinator cleanly shuts down, rather than treating it as an RPC failure. Check for existing tests related to worker shutdown.

由索引模型根據 Issue 內容生成。