Check Inode Counts with Formatting

Inode exhaustion is one of those silent failures that can take down a service long before disk space runs out. A quick per-directory inode count can point you straight at the culprit. The snippets below give you a fast, formatted view of where your inodes are going, plus a disk-usage summary for the top directories.

Count Inodes per Directory (Simple)

This loop walks each top-level directory and prints the file count:

for d in *; do echo -n "Directory: "$d": "Inode Count: "; find $d -type f | wc -l; done

Count Inodes with Formatted Output

A slightly more structured version that uses printf for tab-aligned columns and prints a running total:

for d in `find -maxdepth 1 -type d | cut -d\/ -f2 | grep -xv . | sort`; do c=$(find $d | wc -l); printf "$c\t\t- $d\n"; printf "Total: \t\t$(find $(pwd) | wc -l)\n"; done

Disk Usage for Top Directories (with Progress)

This one-liner uses du to report the largest directories, applies a nice/renice adjustment so it does not starve other processes, and shows a small spinner while it works:

many=25; niceness=-13; echo; on=$(nice); renice $niceness -p $$ 2>&1>/dev/null; echo -e "## [\e[1;36mspace/deep\e[0m] Top $many from \e[1;37m$PWD\e[0m"; c=0; echo -n "Please wait... " 1>&2; find -type d 2>/dev/null | while read stuff; do du -Ss ${stuff#\.\/} 2>/dev/null; let "c+=1"; case $c in 1) echo -ne "\b\\" 1>&2;; 25) echo -ne "\b|" 1>&2;; 50) echo -ne "\b/" 1>&2;; 75) echo -ne "\b-" 1>&2;; 100) c=0;; esac; done | sort -rn | head -n$many | awk '{total+=$1; i++; if(i==5){top5=total}; if($1/1024>=1) printf("\r\033[1;33m%\0478i\033[0m MB - %s\n", $1/1024, $2);} END{printf("\n\033[1;33m%\0478i\033[0m MB in top %i listed directories\n", total/1024, i); printf("\033[1;33m%\0478i\033[0m MB in top 5 listed directories\n", top5/1024);}' ; renice $on -p $$ 2>&1>/dev/null; echo;

What it does, step by step

1. Sets many=25 (number of top directories to report) and niceness=-13 (a higher priority for the scan).
2. Saves the current nice value, applies the new one, and restores it when done.
3. Walks every directory under the current path with find -type d, runs du -Ss on each, and cycles a spinner character every 25 iterations.
4. Pipes the results through sort -rn | head to keep only the largest entries.
5. Uses awk to convert kilobytes to megabytes, highlight values in yellow, and print a summary of the top 5 and overall total.

When to Use These

• Inode counts: Run the first two snippets when df -i shows high inode usage and you need to identify which directory is consuming the most inodes.
• Disk usage: Use the du one-liner when you want a quick, human-readable breakdown of the largest directories without installing additional tools.

Both approaches are read-only and safe to run on live systems.

• SysAdmin

• Bash
• SysAdmin