Just a quick way to look at Inode Counts

Running out of inodes is one of those sneaky Linux problems that catches you off guard. Your df output shows plenty of free space, yet writes fail with “No space left on device.” The culprit is inode exhaustion, and the fastest way to find the offender is a single bash pipeline.

The One-Liner

find -maxdepth 1 -type d | while read dir; do echo -en "$dir - "; find "$dir" | wc -l; done | sed 's/^\.\///' | column -t | sort -rnk3 | head -n10

What It Does

1. find -maxdepth 1 -type d lists every directory at the current level.
2. The while loop counts all inodes (files, subdirectories, symlinks) inside each directory with find "$dir" | wc -l.
3. sed 's/^\.\///' strips the leading ./ so the output reads cleanly.
4. column -t aligns the results into a tidy table.
5. sort -rnk3 sorts numerically in reverse order by the third column (the count).
6. head -n10 keeps only the top ten offenders.

When to Use It

• You see ENOSPC errors but df -h reports free blocks.
• A service refuses to create new files or logs.
• You want a quick audit before running find / -xdev -printf '%i\n' | sort -n | uniq -c | sort -rn | head for a full filesystem scan.

Tips

• Run it from the mount point you suspect (e.g., cd /var before executing) to keep the scan fast.
• If a single directory dominates, drill into it by cd-ing there and re-running the same command.
• Pair the results with df -i to confirm the inode usage percentage for the affected filesystem.

• SysAdmin

• Bash
• Troubleshooting
• SysAdmin