How to troubleshoot the problem of too many inodes

When it comes to The server Inodes Call the police , The troubleshooting process is as follows

Operating system environment ：CentOS6.8

One 、 Use df -i Check which partition uses the most • In this way, the problem partition is located first

Two 、 Enter this partition , Use the file count script

# File count script

#!/bin/bash

for i in `ls`; do   file_num=`find ./$i/ -type f -print | wc -l`   echo "$i,$file_num" done

• This script will count the number of files in each directory under the current directory , In this way, you can find the directory with the most files , Then check layer by layer , You can locate the problem directory .

3、 ... and 、 Delete in small batches

ls | grep '^' | xargs -n 100 rm -f  # Delete the listed files , With 100 One for a group

# This instruction is used to avoid "argument list too long " • Finally, don't forget to check the cause of the problem

Finally, share the problems encountered by the writer • I used the method of appeal to find out /var/spool/postfix/maildrop/ There are a lot of documents

The reason is that the system is executing cron When the task , Will cron Execute the output information in the script (output) And alarm information (warning) Send by email to cron owner , And my The server It's closed postfix , Cause mail sending failure , Small files are piled up maildrop Under the table of contents . If sendmail or postfix The normal operation , Will be in /var/mail A large number of mail files are piled under the directory . • resolvent ： modify crond The configuration file

vim /etc/crontab take MAILTO=root Replace with MAILTO="", Need to restart crond Service is OK