GitLab CI "no space left on device" (Runner Disk Full)

By Kaveh Alemi·Latchkey

A step failed because the runner ran out of disk. Accumulated Docker images, caches, build outputs, or artifacts filled the volume mid-job.

What this error means

A build, checkout, or artifact step dies with "no space left on device" / "write error: No space left on device". It often starts intermittently on a busy runner and worsens over time.

tar: write error: No space left on device
ERROR: Job failed: exit code 2
# or
failed to register layer: write /usr/lib/...: no space left on device

Diagnose it: which rule matched, and on which runner?

GitLab evaluates rules: top to bottom and the first match wins, including one that sets when: never. A job that does not run, or runs when you did not expect it to, is nearly always matching an earlier rule than the one you are reading.

# validate the definition against the project
curl -s --header "PRIVATE-TOKEN: $TOKEN" \
  "https://gitlab.com/api/v4/projects/$CI_PROJECT_ID/ci/lint" \
  --data-urlencode "content=$(cat .gitlab-ci.yml)"

# what the job actually sees
script:
  - env | grep -E "^CI_(PIPELINE_SOURCE|COMMIT_REF_NAME|RUNNER)" | sort

Common causes

Runner host disk filled

Long-lived runners accumulate dangling Docker images, build caches, and old artifacts until the volume is full. The next job fails on the first large write.

A single job writes too much

A job that unpacks huge images, generates large build outputs, or downloads big datasets can exhaust even a freshly cleaned disk.

How to fix it

Reclaim disk on the runner host

Prune unused Docker data and stale build directories.

docker system prune -af --volumes
df -h
# remove old build dirs under the runner's builds_dir if safe

Reduce per-job footprint

1. Use smaller base images and multi-stage builds to shrink layers.
2. Limit cache and artifact sizes; expire old artifacts sooner.
3. Schedule periodic cleanup, or use ephemeral runners that start with a clean disk.

How to prevent it

• Run docker system prune on a schedule for long-lived runners.
• Prefer ephemeral runners that reset disk between jobs.
• Keep cache and artifact sizes bounded with short expiry.

Frequently asked questions

Related guides

References

• GitHub Actions documentation