Troubleshooting
Server disk full: how to find what used the space
A server can stop behaving normally when a filesystem reaches capacity. Before deleting the largest thing you can find, identify which filesystem is full, whether the problem is bytes or inodes, and what application is creating the growth.
Check the actual filesystem and inode pressure
A mount can be full while another disk still has space. Also check inode/file-count exhaustion; millions of tiny files can prevent new files from being created even when the byte total does not look extreme.
Find large directories and recent growth
Logs, databases, uploads, caches, package files, container layers, backups, temporary files, and application queues are common consumers. Compare current usage with monitoring history when available to find what changed.
Watch for deleted files still held open
A process can keep disk blocks allocated after a file is deleted until that process closes the file. This can make directory scans disagree with filesystem usage. Restarting the correct process after preserving logs may release the space.
Fix retention and alerts after recovery
Once service is stable, set log rotation, backup retention, cache cleanup, application limits, or capacity changes appropriate to the cause. Add warning thresholds so the next incident is detected before the filesystem reaches zero free space.