docker restart: Stop and Start a Container

Restart a container: stop it, then start it again.

docker restart stops and then starts a container in one command, reusing its config. This page covers the timeout and using it to recover a wedged service.

What it does

docker restart issues a stop (SIGTERM then SIGKILL after the grace period) followed by a start. The container keeps its configuration and writable layer, so it is the same container, just cycled.

Common usage

Terminal

docker restart web
docker restart -t 30 web                 # 30s grace period
docker restart $(docker ps -q)

Common errors in CI

If the container crashed on start it will exit again right after restart - fix the root cause rather than looping restarts. A container run with --rm is removed on stop, so docker restart errors with "No such container"; recreate it with docker run. In CI prefer fixing config over restart-as-a-retry.

Related guides

docker stop: Gracefully Stop ContainersHow docker stop works: sending SIGTERM then SIGKILL, the grace period, stopping multiple containers, and clea… docker start: Start Stopped ContainersHow docker start works: restarting an existing stopped container, attaching, and why start differs from run i… docker kill: Force-Stop Containers with a SignalHow docker kill works: sending SIGKILL or a custom signal to a container immediately, and how it differs from… docker ps: List Containers, Filters & FormattingHow docker ps works: listing running and stopped containers, filters, format strings, and reading exit status…