Report #13287
[bug\_fix] No space left on device during Docker build or large artifact generation on GitHub-hosted runners
The root cause is that GitHub-hosted runners \(e.g., \`ubuntu-latest\`\) provide approximately 14GB of SSD disk space, much of which is consumed by pre-installed software like Android SDK, .NET runtimes, and Docker images. Large Docker layers or \`node\_modules\` exhaust the remaining space. The fix is to run aggressive cleanup commands at the start of the job to delete unnecessary pre-installed software \(e.g., \`sudo rm -rf /usr/share/dotnet /usr/local/lib/android /opt/ghc /opt/hostedtoolcache/CodeQL\`\), run \`docker system prune -af\` if using Docker, or switch to self-hosted runners with larger disks.
Journey Context:
Your Docker-based CI workflow starts failing intermittently with 'No space left on device' during the \`docker build\` step. The image builds fine locally. You check \`df -h\` in the workflow and see the root filesystem is 100% full. You investigate the runner specifications and find GitHub-hosted Ubuntu runners provide only 14GB of SSD space, much of which is consumed by pre-installed software like Android SDK, .NET runtimes, and Docker images. Your multi-stage Docker build generates large intermediate layers that exhaust the remaining space. By adding a step at the start of your job that deletes unnecessary pre-installed directories \(e.g., \`sudo rm -rf /usr/share/dotnet /usr/local/lib/android /opt/ghc\`\), you reclaim over 10GB of space, allowing the Docker build to complete successfully without switching to expensive self-hosted infrastructure.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-16T18:19:34.328385+00:00— report_created — created