CI runner disk full

What this failure means

The CI runner has run out of available disk space. The build, test, or artifact step cannot write files and fails with ENOSPC.

Symptoms

Faultline looks for one or more of these log fragments:

No space left on device
disk quota exceeded
Write failed: No space left
not enough disk space
insufficient disk space
out of disk space
Disk is full
Error processing tar file

Diagnosis

The CI runner has run out of available disk space. The build, test, or artifact step cannot write files and fails with ENOSPC.

Fix steps

1. Check remaining space before diagnosing: df -h and du -sh /* | sort -rh | head -20.
2. Remove Docker build cache and dangling images: docker system prune -af --volumes.
3. Clear dependency caches: rm -rf ~/.npm ~/.gradle/caches ~/.m2/repository ~/.cache/pip.
4. In GitHub Actions, add the jlumbroso/free-disk-space action before large build steps.
5. In GitLab CI, prune images in before_script and reduce artifact retention.
6. In CircleCI, choose a larger resource_class or clean up large workspace directories.
7. Upgrade to a runner or agent tier with more disk, or split the job across multiple stages.
8. Reduce .dockerignore exclusions to keep the build context small.
9. Add a cheap guard near the top of the job so storage problems fail fast with a clear message instead of surfacing mid-build.

Validation

• Re-run the failing workflow step after confirming free space with df -h.
• Confirm the original failure signature for CI runner disk full is gone.

Why it matters

Disk space is consumed by: Docker image layers pulled during the build, large build artifacts not cleaned up between jobs, excessive logging, test fixtures, dependency caches, or container image builds that layer many files.

Prevention

• Add a disk space check at the start of long jobs: df -h.
• Prune Docker build cache between jobs in your CI runner configuration.
• Use remote layer caching (BuildKit registry cache or kaniko) to avoid re-downloading large layers.
• Set explicit artifact/cache expiry policies on your CI platform to avoid accumulation across runs.

Try it locally

df -h
docker system df
df -h

How Faultline detects it

Use faultline explain runner-disk-full to see the full playbook.

faultline analyze build.log
faultline explain runner-disk-full

Generated from playbooks/bundled/log/ci/runner-disk-full.yaml. Do not edit directly.