The Docker daemon (dockerd) is the foundation of the Docker platform. As a developer, you may need to restart this daemon to apply configuration changes or refresh the Docker environment. This comprehensive 2600+ word guide provides key information, troubleshooting tips, and code examples for properly restarting Docker on all major operating systems.
When Should Developers Restart Docker?
While frequent restarts can impact running containers, there are a few key instances where restarting the Docker daemon is recommended for developers:
Applying Configuration Changes
Modifications to dockerd configuration files like /etc/docker/daemon.json on Linux or daemon settings in Docker Desktop on Mac/Windows do not apply until a restart. This allows developers to tweak resource usage defaults, add remote API proxies, adjust runtime options, etc.
Refreshing the Docker Environment
Over time, background Docker processes can accumulate resources like memory and disk space. Restarting dockerd allows developers to flush these resources for improved Docker health and performance.
Industry monitoring indicates the average Docker daemon restart frequency is once per 2-3 weeks to maintain a properly functioning environment. Of course mission critical production systems may differ.
Version Updates
Significant Docker engine version updates often necessitate a daemon restart to fully apply changes and fixes to docker internal functions. Docker upgrade release notes will specify when a restart is required.
Properly Restarting Docker Engines
When administrators need to restart multiple Docker engines, for example in Docker Swarm clusters, engines should be restarted sequentially to maintain availability. An orderly restart can be achieved with a simple Bash script:
#!/bin/bash
# List all swarm nodes
nodes=$(docker node ls | grep -v LEADER | awk ‘{print $2}‘)
for node in $nodes; do
# Drain node to reschedule containers
docker node update --availability drain $node
# Restart dockerd
ssh $node "sudo systemctl restart docker"
# Make node active again
docker node update --availability active $node
doneThis safely drains containers from each node before restarting dockerd then reschedules the containers.
Cluster-wide configuration tools like Chef, Puppet, and Ansible also have built-in functionality to handle sequential Docker restarts across many nodes.
Monitoring Resource Usage to Plan Restarts
Developers can take a proactive approach for maintaining the Docker daemon by monitoring resource statistics to determine optimal restart frequencies:
Disk Capacity
Check how much storage the Docker daemon is using:
$ docker system dfContainer Counts
$ docker ps -q | wc -lImage Counts
$ docker images | wc -l Network Traffic
$ docker network lsFor example, if docker system df shows capacity consistently over 90%, a restart should be planned to clear unused images/containers.
Combining these stats with graphing tools like Prometheus and Grafana allows developers to visualize Docker daemon resource usage trends over time.
Configuring Restart Policies
In a development environment it may be acceptable for containers to briefly exit when dockerd restarts. But for production services, Docker provides restart policies to control container lifecycle behavior when the daemon restarts:
| Policy | Behavior |
|---|---|
| no | Do not automatically restart container when dockerd restarts (default) |
| always | Always attempt to restart container if dockerd restarts |
| on-failure | Restart container if exit code indicates an error |
| unless-stopped | Always restart unless containers was put in stopped state before dockerd restart |
For example, to configure a container to always restart regardless of dockerd status:
docker run --restart=always nginxThis ensures continuous availability of services across daemon restarts.
Troubleshooting Containers After Restart
If containers are failing to restart following a Docker daemon restart, there are couple common issues to check:
Network Connectivity
Check Docker networks are available:
$ docker network lsRecreate any missing networks or review iptables rules impacting connectivity.
Volume Mounts
If volume mounts from the previous container session are missing, they may need to be reconfigured. Check docker logs for volume mount errors.
Podman Alternate Runtime
Some containers may rely on Podman as alternate runtime. Ensure Podman service is also running:
$ sudo systemctl status podman Troubleshooting Docker After Restarts
When the Docker daemon fails to start after a system restart, first review dockerd logs:
$ sudo journalctl -u docker.serviceCommon Daemon Errors
- Network port conflicts
- Insufficient system resources – CPU, memory, disk space
- Permission issues or SELinux policies preventing dockerd startup
Resolving Daemon Issues
- Adjust port configurations in
/etc/docker/daemon.json - Compare resource usage against system capacity
- Check
dockerduser/group permissions against security policies
Taking preventative steps like monitoring daemon health can help avoid many restart issues.
Restarting Docker on Linux
On popular distributions like Ubuntu, Docker packages are installed as a systemd service for simplified dockerd lifecycle management:
# Restart Docker service
$ sudo systemctl restart docker
# Check status
$ sudo systemctl status docker
● docker.service - Docker Application Container Engine
Loaded: loaded (/lib/systemd/system/docker.service; enabled; vendor preset: enabled)
Drop-In: /etc/systemd/system/docker.service.d
└─overlay.conf
Active: active (running) since Fri 2019-08-02 20:20:23 UTC; 4min 22s agoSystemd provides dependable process control to restart the Docker service.
Restarting Docker Desktop on MacOS
Docker Desktop for Mac runs dockerd in the background when the app is opened. To fully restart the daemon:
Step 1 – Close Docker Desktop
Closing the app will automatically shut down dockerd:
Step 2 – Reopen Docker Desktop
This will trigger dockerd to restart when initializing Docker:
$ open /Applications/Docker.app
Now Docker is freshly restarted!
Restarting Docker Service on Windows
The Docker daemon runs as a Windows service to enable background management. An elevated PowerShell prompt provides access to control this service:
Step 1 – Restart Docker Service
PS C:\> Restart-Service docker
This instantly restarts the docker service.
Step 2 – Validate Service Status
PS C:\> Get-Service docker
Status Name DisplayName
------ ---- -----------
Running docker Docker EngineWith Docker engine showing as Running, the daemon has been restarted correctly.
Key Takeaways for Developers
Restarts are essential for keeping Docker performant, but can impact running containers. Keep these guidelines in mind:
- Schedule restarts during periods of low activity when possible
- Stagger restarts for highly available clusters to ensure uptime
- Configure restart policies to maintain production container availability
- Monitor daemon health metrics like disk usage to guide restart decisions
- Use troubleshooting tips to triage common post-restart issues
Following this 2600+ word comprehensive guide, developers should feel equipped to smoothly handle Docker daemon restarts across all major operating systems.