Storage Capacity, Health Thresholds & Filesystem Triage

Disk usage warning (>80%) and critical (>95%) thresholds, SSD endurance, TRIM, Linux inode exhaustion (df -i), and NTFS dirty bit recovery.

Disk Capacity Thresholds

Netscrew evaluates filesystem health using established storage engineering thresholds:

• < 75% Used [OK]: Optimal performance. File system has plenty of contiguous free blocks to prevent fragmentation.
• 75% - 85% Used [OK / Caution]: Normal operating condition. Consider scheduling maintenance or cleaning old log archives.
• > 85% Used [WARNING]: Free space is constrained. On SSDs, wear-leveling algorithms lose efficiency and write amplification increases.
• > 95% Used [CRITICAL]: Emergency threshold. Background updates, database checkpoints, and system restore snapshots will begin failing.
• > 99% Used [EMERGENCY]: Kernel crashes, database corruption, or system lockup can occur as swap files and temporary files fail to write.

Special Trap: Linux Inode Exhaustion (df -i)

The Symptom:

A Linux server returns:

touch: cannot touch 'test.txt': No space left on device

Yet running df -h reports 50 GB of free space!

The Cause:

Every file, directory, and symlink on a Linux filesystem requires an inode. If millions of small zero-byte files (such as PHP session files, mail queue spools, or cache files) are generated, the filesystem can run out of inodes long before running out of disk bytes.

The Fix:

Check inode usage with df -i. Locate and delete directory trees with excessive file counts:

find /var/spool -type f | wc -l