Without warning, MySQL went on strike. After starting with the Mysql.server startup script, there is no hint that the message ends, and the mysqld process is naturally not up.
Put the mysql.server inside $bindir/mysqld_safe--datadir= $datadir--pid-file= $server _pid_file >/dev/null This sentence output redirect after the operation, the results found error: File size limit exceedednohup ...
So it seems natural that there are too big a file. But look at the MySQL data directory related to the database files and log files, the largest is hundreds of trillion. Ulimit-a View also does not have any restrictions on file size.
After N minutes of depression, decided to use Find/-size +2048000k search the entire hard disk suspicious files. The result:/var/log/pacct this large 2G file. Obviously, this is the log file that was generated after the SA activated the Accton and did not roll back. Apart, empty/var/log/pacct and start again MySQL, everything back to normal!
What I don't understand is why MySQL's startup scripts are affected by such a trivial process statistic log that the mysqld process does not start completely. Estimating is still a problem with Mysql.server/mysqld_safe scripting logic.