After configuring nginx, check the disk space and use df -lh to check that the system space is full. . . . .
/dev/xvda1 20G 20G 0G 100% /
Then, query the corresponding large file through du --max-depth=1 -h / and find that only 2.3G of space is occupied
I couldn't find a large file in various ways. Later, I saw in an article that it might be caused by deleting the corresponding log file when apache or tomcat was running. Suddenly I remembered that there was a small partner a few days ago. Said the space was full, I asked him to delete the log under tomcat, and then he did not restart. . . . . .
The reason for the problem:
When the apache/tomcat service is running, the log of the running service is cleared, which leads to the problem that /dev/xvda1 is full. In general, most services (including scripts) cannot delete log files that are currently being written while they are running.
Principle analysis:
Restart the related service or program of the log file, for example, restart apache for the apache log file (as shown in the figure below, the hard disk space is occupied normally after the restart).
refer to:
http://www.android100.org/html/201407/27/46936.html