Strange Mysql Problem

A very stange thing has been happening with my mysql service. About three weeks ago, with no apparent changes in mysql, my server locks as a result of mysql dying. The reason for the lock is that once mysql dies, mysql starts loading multiple copies until all of my memory is consumed. I can duplicate the problem by killing mysqld with the 'kill -9 pid' command.

DA sends a message that the mysqld service is down. Several minutes later, so is my server. This is now happening about once every 24-48 hours. I am not so much concerned about mysql dying more the server lockup when it trys to reload, and reload, and reload. I can start and stop the mysql service from the DA control panel without a problem and it comes up just fine at boot time.

I run a simple and standard mysql server. There is no 'my.cnf' file. mysqld starts at bootup from the /etc/rc.d/init.d/mysqld script.
mysqld_safe is also a running process and the script is located in /usr/bin. I updated to 4.0.18 to see if that would fix the problem but no luck. I'm running Rehat 9 with the latest version of DA.

I couldn't find anything searching and am wondering if this is some issue with DA trying to restart mysql.

Any help would be appreciated.

I assume you had a fresh OS install?

Check the error logs /var/log/

If you don't see something abnormal. I would assume that something is very wrong somehwere and remove ALL traces of MySQL from the system. Then go:

/usr/local/directadmin/scripts/ and run the ./mysql.sh to reinstall.

Hello,

I'm not too sure about the mysql problem, but if you want to turn off DA's check for mysql running, just set it to OFF in the /usr/local/directadmin/data/admin/services.status file.

John

I have almost the same problem

I have a my.cnf file that keeps the mysql running longer with it it dies within 14-24hrs and run like you said restart it self

after 2 restarts I need to kill all mysqld procs and start it manualy

I have two machines one with DA, and one without

on the DA machine SQL dies

and the other I got more then 300+ Users no my.cnf file works great only 5-6 mysqld procs running

on the DA machine 25+ mysqld procs ( and also apache )

why?

Hello,

The mysql rpms are stock from mysql.com .. DA only adds/removes users.. I couldn't say why it's running with more processes. Generally that just means a busier site.

John

I don't know if this cured the problem or not but I have been up for three days now without a problem.

DA checks for mysql running had already been set to OFF.

What I noticed what that both init and mysqld_safe were trying to reload mysqld when it died. Because of my lack of Linux experience, I can't expain why either mysql_safe or init tries then to reload multiple mysql servers on the system thereby consuming all memory very quickly.

Regardless, I disabled the stock init mysqld startup script and call


./usr/bin/mysqld_safe &

as a command in /etc/rc.d/rc.local

So far so good but it's only been running for a few days.

The site is not busy enough to cause these types of problems but I will keep my eye on that condition.

Thanks to everyone for their input. I will keep you posted.

I think my problem is the apache 1.3.29 because I got mysqld running as much as the apache mybe even more more apache pids running more mysqld pids

in rc.local I dont have mysqld there so
I dont know how to fix it