Find inode usage

Disk space issues can be caused by large files or too many inodes. Inodes store information about files and directories (folders), such as file ownership, access mode (read, write, execute permissions), and file type. The maximum number of inodes, like disk space, is set when the server is created.

Symptoms

Some symptoms of too many inodes will be the same as disk space issues.

• You may receive emails from Plesk or WHM that inode limits are being approached or have been exceeded.
• You may receive error messages like Out of disk space. Unable to write content to file or failed: No space left on device.
• You may receive these errors when, in fact, there is plenty of physical disk space available.
• The server may hang or take a long time to respond when attempting to view the contents of a directory.
  
  

1. Enable administrator access on your Gen 3 or Gen 4 server if you haven't already done so.
2. Connect to my server with SSH (Secure Shell).
3. Switch to the root user.
4. For an overview of disk usage on the server, use the df command.
   
   

   [root@server[~]: df
   Filesystem       1K-blocks    Used   Available  Use%  Mounted on
   /dev/ploop29904p1 125684164 11693260 107699656  10%   /
   devtmpfs           524288     60  524228    1% /dev
   tmpfs              524288      1  524287    1% /dev/shm
   tmpfs              524288    311  523977    1% /run
   tmpfs              524288     10  524278    1% /sys/fs/cgroup
   tmpfs              524288      1  524287    1% /run/user/1000
   

5. If you have exhausted the inodes within your main file system, it may look similar to the example below:

   [root@server[~]: df -i
   Filesystem         Inodes   IUsed   IFree IUse% Mounted on
   /dev/ploop29904p1 7864320 7864222      82  100% /
   devtmpfs           524288      60  524228    1% /dev
   tmpfs              524288       1  524287    1% /dev/shm
   tmpfs              524288     311  523977    1% /run
   tmpfs              524288      10  524278    1% /sys/fs/cgroup
   tmpfs              524288       1  524287    1% /run/user/1000
   

6. To view the inode distribution within the current working directory:

   find * -maxdepth 0 -type d -exec sh -c "echo -n {} ' ' ; ls -lR {} | wc -l" \;
             

7. So, if you switched to the root directory of the server and ran the command, it would produce output that looks like this:

   [root@server[~]: cd /
   root@server[/]: find * -maxdepth 0 -type d -exec sh -c "echo -n {} ' ' ; ls -lR {} | wc -l" \;
   backup  2
   boot  7
   dev  78
   etc  7769
   home  1448
   lost+found  2
   media  2
   mnt  2
   opt  11749
   proc  21481
   root  56
   run  393
   srv  2
   sys  1643
   tmp  11
   usr  231243
   var  7468179
   

8. In the example, you can see that the var directory contains the bulk of the inodes. For more specifics, move into /var and run the command again.

   Note: It is important to drill down enough in the filesystem to separate what can be removed (like cache files) from what can't be removed (like your email content).

   [root@server[/]: cd /var
   root@server[/var]: find * -maxdepth 0 -type d -exec sh -c "echo -n {} ' ' ; ls -lR {} | wc -l" \;
   adm  2
   cache  779
   cpanel  13698
   db  20
   empty  6
   games  2
   gopher  2
   installatron  48856
   kerberos  10
   lib  13004
   local  2
   log  419
   named  49
   nis  2
   opt  2
   preserve  2
   spool  7391257
   tmp  29
   www  18
   yp  2
             

In the example above, you can see that the /var/spool directory is where the bulk of the inodes are being used. We know this is where the email queue is located on a WHM/cPanel server, so we can come to the conclusion that we have an issue with excessive email messages that needs to be addressed.

Not out of inodes?

Too many inodes in a single folder can cause performance issues, even if inodes aren't exhausted on the server. This may be more noticeable if you have several directories with more than 1,024 inodes.

To scan all of the directories on the server, and list the 20 directories with the largest amount of inodes, you can use this command:

find / -xdev -printf '%h\n' | sort | uniq -c | sort -k 1 -n | tail -20
          

The output of the command will look like this:

[root@server ~]# find / -xdev -printf '%h\n' | sort | uniq -c | sort -k 1 -n | tail -20
   2180 /home/onecool/public_html/wp-content/uploads/2018/05
   2180 /home/onecool/public_html/wp-content/uploads/2019/07
   2271 /home/onecool/public_html/wp-content/uploads/2019/05
   2683 /home/onecool/public_html/wp-content/uploads/2020/11
   2768 /home/onecool/public_html/wp-content/uploads/2020/09
   2822 /home/onecool/public_html/wp-content/uploads/2019/12
   2929 /home/onecool/public_html/wp-content/uploads/2020/06
   3064 /home/onecool/public_html/wp-content/uploads/2020/04
   3100 /home/onecool/public_html/wp-content/uploads/2020/07
   3186 /home/onecool/public_html/wp-content/uploads/2020/08
   3332 /home/onecool/public_html/wp-content/uploads/2020/01
   3354 /home/onecool/public_html/wp-content/uploads/2019/11
   3445 /home/onecool/public_html/wp-content/uploads/2020/02
   3706 /home/onecool/public_html/wp-content/uploads/2019/09
   3743 /home/onecool/public_html/wp-content/uploads/2020/03
   3846 /opt/cpanel/ea-openssl11/share/doc/openssl/html/man3
   3846 /opt/cpanel/ea-openssl11/share/man/man3
   3964 /home/onecool/public_html/wp-content/uploads/2019/08
   4845 /home/onecool/public_html/wp-content/uploads/2020/05
   5078 /home/onecool/public_html/wp-content/uploads/2020/10
          

9. Once you have identified where the use is, you can remove content and work to prevent the issue from happening again.

Related steps

• Move on to remove excessive inodes.
• Return to Disk Space Audit.
• It's important to identify and correct the root issue that caused the disk space to become exhausted. Move on to preventing disk usage issues.

More info

• Our server experts can perform these steps for a fee. For more information about our Expert Services, please visit our Expert Service menu.