When your Linux system reports 100% inodes usage in the root directory (/), it's often more critical than disk space exhaustion. Inodes store metadata about files, and when they're depleted, you can't create new files - even if there's available disk space.

First, verify the inodes situation:

df -i
Filesystem     Inodes  IUsed  IFree IUse% Mounted on
/dev/sda2     732960 727804   5156  100% /

Then identify which directories consume the most inodes:

for i in /*; do echo $i; find $i | wc -l; done

1. Small Files Accumulation

Millions of tiny files (like session files, cache, or logs) can exhaust inodes without consuming much space.

To find directories with most files:

find / -xdev -type f | awk '{print $NF}' | \
awk -F/ 'BEGIN {depth=3} {for(i=1;i<=depth;i++) printf "/%s",$i; print ""}' | \
sort | uniq -c | sort -n

2. Orphaned Docker Containers

Docker can leave behind layers and volumes:

docker system prune -a --volumes

3. Email Queues

Mail servers often create thousands of small files:

# For Postfix
postsuper -d ALL

Bulk Deletion of Cache Files

# For package manager cache
apt-get clean
# Or for yum
yum clean all

# For general cache
find /var/cache -type f -delete

Log File Rotation

Configure logrotate more aggressively:

vim /etc/logrotate.conf
# Set:
rotate 3
compress
delaycompress

Implement monitoring with this bash script:

#!/bin/bash
THRESHOLD=90
CURRENT=$(df -i / | awk 'NR==2 {print $5}' | sed 's/%//')

if [ "$CURRENT" -gt "$THRESHOLD" ]; then
    echo "Warning: Inodes usage exceeded threshold ($THRESHOLD%) - Current: $CURRENT%"
    # Add cleanup commands here
fi

If you consistently hit inodes limits, consider:

# For ext4 filesystems (check first)
tune2fs -l /dev/sda2 | grep Inode
# To resize (unmount first)
resize2fs /dev/sda2

Remember that some operations require booting from a live CD/USB for root partition operations.

---

When your Linux system reports 100% inode usage in the root directory while disk space appears available, you're facing a filesystem metadata exhaustion. The df -i output reveals the critical situation:

Filesystem  Inodes  IUsed  IFree  IUse%  Mounted on
/dev/sda2  732960  727804   5156   100%  /

From your logs, two directories dominate inode consumption:

• /proc: 10,937 inodes
• /sys: 22,504 inodes

However, these are virtual filesystems and shouldn't be your primary targets. The real issues likely hide elsewhere.

Use these commands to locate inode-heavy directories:

# Top-level directory scan
sudo find / -xdev -type f | cut -d "/" -f 2 | sort | uniq -c | sort -n

# Detailed directory analysis
sudo find /path/to/search -xdev -type f | awk -F/ '{print $2}' | sort | uniq -c | sort -nr | head -20

Common inode hogs and their solutions:

# Clean package cache (especially effective after failed apt operations)
sudo apt-get clean
sudo apt-get autoclean

# Remove old kernel versions
sudo apt-get purge $(dpkg -l | grep linux-image | awk '{print $2}' | grep -v $(uname -r))

# Clear systemd journal logs
sudo journalctl --vacuum-size=50M

When standard cleanup isn't enough:

# Force remove stale files (careful with this!)
sudo find /tmp -type f -atime +10 -delete

# Identify and remove small files (common inode wasters)
sudo find / -xdev -type f -size +0c -size -10k -exec ls -la {} \; | sort -n -k 5

• Implement log rotation for all services
• Monitor inode usage with cron jobs
• Consider separate partitions for /var, /tmp

# Sample monitoring script
#!/bin/bash
INODE_USAGE=$(df -i / | awk 'NR==2 {print $5}' | tr -d '%')
if [ $INODE_USAGE -gt 90 ]; then
    echo "Warning: Inode usage at $INODE_USAGE% on $(hostname)" | mail -s "Inode Alert" admin@example.com
fi