Skip to content

docker-cannot-connect-to-docker-daemon.md

Latest commit

 

History

History
144 lines (106 loc) · 4.54 KB

docker-cannot-connect-to-docker-daemon.md

File metadata and controls

144 lines (106 loc) · 4.54 KB
Raw
title Docker Cannot Connect to the Docker Daemon
slug docker-cannot-connect-to-docker-daemon
technologies
docker
severity high
tags
docker
daemon
socket
connectivity
production
related
permission-denied-docker-socket
oci-runtime-create-failed
last_reviewed 2026-06-27

Docker Cannot Connect to the Docker Daemon

Error Message

Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?

error during connect: Get "http://%2F%2F.%2Fpipe%2Fdocker_engine/v1.45/version": open //./pipe/docker_engine: The system cannot find the file specified.

Description

The Docker CLI (docker) is a thin client that talks to a separate long-running server, dockerd. This error is emitted by the client when it cannot reach the daemon's API endpoint — by default the Unix socket /var/run/docker.sock on Linux or the named pipe //./pipe/docker_engine on Windows. It does not mean Docker is broken; it means the client could not open a connection to wherever it expects the daemon to be listening.

Technologies

  • docker (CLI client, dockerd daemon, systemd)

Severity

high — no Docker operation works while the client cannot reach the daemon. On a CI runner or a host whose workloads are managed by Docker, this halts builds and can take running services with it if the daemon itself has crashed.

Common Causes

  1. The docker daemon is not running (stopped, crashed, or never started).
  2. DOCKER_HOST is set to a host/socket that is wrong or unreachable.
  3. The daemon listens on a non-default socket and the client wasn't told about it.
  4. The user lacks permission to the socket (this usually surfaces as permission denied — see related error — but a missing socket reads as this).
  5. Inside a container, /var/run/docker.sock was not mounted (Docker-in-Docker).

Root Cause Analysis

The client resolves its endpoint in this order: -H/--host flag, the DOCKER_HOST environment variable, the active docker context, then the built-in default socket. If the resolved endpoint has no listener — because dockerd is down or the path is wrong — the connect syscall fails and the CLI prints this message. Distinguishing "daemon down" from "wrong endpoint" is the whole job: check the service state, then check what endpoint the client is actually using.

Diagnostic Commands

# Is the daemon service up?
systemctl status docker

# What endpoint is the client trying to use?
docker context inspect
echo "$DOCKER_HOST"

# Does the socket exist and who owns it?
ls -l /var/run/docker.sock

# Is anything listening on the socket?
ss -lx | grep docker.sock

# Recent daemon logs (why it failed to start, if it did)
journalctl -u docker --no-pager -n 50

Expected Results

$ systemctl status docker
   Active: inactive (dead)        # daemon is down -> start it

$ ls -l /var/run/docker.sock
ls: cannot access '/var/run/docker.sock': No such file or directory

# Healthy host:
$ systemctl status docker
   Active: active (running)
$ ss -lx | grep docker.sock
u_str LISTEN 0 4096 /var/run/docker.sock 12345 * 0

Resolution

  1. If the service is down, start and enable it:

    sudo systemctl start docker
sudo systemctl enable docker

  2. If dockerd fails to start, read journalctl -u docker for the real reason (corrupt daemon.json, a port/socket conflict, storage-driver error) and fix that before retrying.

  3. If DOCKER_HOST or the active context points at the wrong place, unset/correct it:

    unset DOCKER_HOST
docker context use default

  4. Inside a container that needs the host daemon, mount the socket: -v /var/run/docker.sock:/var/run/docker.sock.

Validation

docker info
# Expect server fields (Server Version, Storage Driver) printed without error.

Prevention

  • Enable the service so it survives reboots: systemctl enable docker.
  • Validate daemon.json with dockerd --validate before restarting in CI.
  • Standardize the endpoint via docker context rather than ad-hoc DOCKER_HOST.
  • Add a docker info readiness check at the top of CI jobs that use Docker.

Related Errors

References

Tags

docker · daemon · socket · connectivity · production