We have the 4.0.21 RPM's installed direct from MySQL.com on a Fedora Core 2 server. The previous incarnation of this server was a Fedora Core 1 server which ran flawlessly.
Since the new server has been put in to production, on a daily basis (random times), we are forced to restart mysql because all processes which attempt to get data out of the databases fail. We access the database through Apache+PHP+MySQL as well as via the Perl DBI interface..
Everything comes back after a restart...
In the logs, all we see is:
041018 20:32:50 mysqld started
041018 20:32:51 Warning: Asked for 196608 thread stack, but got 126976
041018 20:32:51 InnoDB: Started
/usr/sbin/mysqld: ready for connections.
Version: '4.0.21-standard' socket: '/var/lib/mysql/mysql.sock' port: 3306 Official MySQL RPM
041020 16:20:35 Error in accept: Too many open files
041020 16:33:23 /usr/sbin/mysqld: Normal shutdown
041020 16:33:24 InnoDB: Starting shutdown... 041020 16:33:27 InnoDB: Shutdown completed 041020 16:33:27 /usr/sbin/mysqld: Shutdown Complete
041020 16:33:27 mysqld ended
I did find that the Warning can be safely ignored. However, the problem definitely is at the "Error in accept: Too many open files" line... the rest of the system is behaving happily, so I am not sure where to go from here..
/proc/sys/fs/file-max comes back with an astronomical 76949, and sysctl shows nominal activity:
fs.file-max = 76949
fs.file-nr = 3760 0 76949
fs.inode-state = 17709 2629 0 0 0 0 0
fs.inode-nr = 17709 2629
A couple of related postings seemed to indicat that it had to do with the mysql process and/or the limits set upon the mysql user account, but that was all speculation.
Ideas, Comments, Suggestions are all more than welcome!
-Rich
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
