Why Inodes Fill Up Before Disk Space? Diagnose and Fix Linux Filesystem Limits

1. What Is an Inode?

In Linux, file data is stored in blocks, while metadata such as owner, timestamps, and size is kept in a structure called an inode (index node).

During formatting, the disk is divided into a data area and an inode table area that stores these metadata structures.

Typical inode sizes are 128 bytes or 256 bytes, and the total number of inodes is fixed at format time (e.g., one inode per 1 KB or 2 KB). On a 1 GB disk with 128‑byte inodes and one inode per 1 KB, the inode table would occupy about 128 MB (12.8% of the disk).

2. Inode Exhaustion

Running out of inodes is similar to running out of disk space: you cannot create new files or execute certain commands.

Even when free space remains, a large number of tiny files can fill the inode table.

Because disk space usually runs out long after inodes, inode usage is often overlooked.

Example commands to view usage:

Check inode usage with:

Using

df -h

shows 71% disk usage, while

df -i

reveals 100% inode usage.

3. Solving Inode Exhaustion

The number of inodes is determined by partition size; larger partitions allocate more inodes.

Root partitions are often small, and periodic creation of tiny files can quickly consume inodes.

Steps to resolve:

(1) Identify directories with the most files Run: for i in /*; do echo $i; find $i | wc -l; done In the example, the /var/spool/postfix/maildrop directory contained many small files generated by cron email notifications that failed to send.

(2) Delete excess files Use: ls | xargs -n 1000 rm -rf The xargs utility ensures batch deletion without errors.

4. Preventive Measures

Disable unwanted mail output in cron by adding

MAILTO=""

to the first line of

crontab -e

.

Redirect output of non‑essential cron jobs to

/dev/null

(e.g.,

*/10 * * * * /tmp/test.sh >/dev/null 2>&1

).

Schedule periodic cleanup of old files, e.g.,

find /path -type f -mtime +30 | xargs -n 1000 rm -f

.

Monitor inode usage regularly with

df -i

or monitoring tools.

By understanding inode allocation and regularly cleaning up unnecessary small files, you can avoid unexpected service disruptions caused by inode exhaustion.