Introduction

A full disk on a web server can trigger far more than one visible error. Sites may stop writing logs, sessions, uploads, caches, or database temp files, and once the host runs out of space, unrelated services can begin failing at the same time. The right fix is to identify what consumed the disk and free space without destroying the evidence you need to prevent the same outage from returning.

Symptoms

  • Deploys, uploads, or application writes fail unexpectedly
  • Services crash or restart repeatedly because temp or log writes fail
  • The site becomes slow, unstable, or partially unavailable
  • Monitoring shows storage usage pinned near or at 100 percent
  • The problem started after log growth, backup buildup, or a deployment change

Common Causes

  • Logs, backups, cache files, or temporary files grew without rotation or cleanup
  • Failed deployments or build artifacts accumulated on the server
  • Database, mail, or queue data grew beyond the allocated disk size
  • Container layers or snapshots consumed space outside the expected app path
  • A runaway process created many files or exhausted inodes instead of raw bytes alone

Step-by-Step Fix

  1. Confirm which filesystem is full and which services depend on it for runtime writes.
  2. Identify the largest growth source before deleting anything, especially if logs or crash data explain the outage.
  3. Check both byte usage and inode usage so you do not miss a small-file explosion.
  4. Remove or archive only non-critical data first, such as stale caches or obsolete build artifacts, while preserving recent evidence.
  5. Review log rotation, backup retention, and deployment cleanup behavior to understand why the disk filled.
  6. Verify applications, databases, and queues can write normally again after free space is restored.
  7. Retest key workflows such as uploads, admin actions, and deploy hooks that failed during the incident.
  8. Expand disk capacity only if the workload genuinely outgrew the current allocation rather than because of preventable sprawl.
  9. Add alerts and retention controls so future storage growth is caught before it reaches production-impacting levels.