Hi there,

We have a Redhat 6.2 machine running Mysql version 3.22.32 (RPM package)

With this the machine also has Apache installed toghether with a static module of PHP 
(version 4.0.4 if I remember correctly...)

This machine now alsmost has an uptime of  1 year (currently it is 344 days..)

We use MySQL very intensively on the machine to store all content from the website and 
user submissions + also use it as a Web-page logger...
MySQL is just great, the last time I looked the 'query' counter stood on a staggering 
20.000.000 !!! We only use PHP with Mysql, no PERL etc etc..

Till now I never had any troubles.. (Neither with Apache/Qmail and Mysql..)

Today I was busy and wanted to look something up on the machine through the 
web-server..
This didn't seem to work and when I telnetted to the machine and took a look at the 
prcocesses I noticed that one of the 
three mysql processes where defunct: eg, something like the following:
14848 ?        00:00:00 safe_mysqld
 9807 ?        00:00:00 mysqld
 9809 ?        00:00:00 mysqld -defunct-
 9810 ?        00:00:00 mysqld

I noticed that the system load was higher than normal.. But I think that this was the 
result of about 10 PHP scripts which are started through Cron 
and tried to connect to MySQL....

Because I didn't want to take a lot of time to see what was wrong I first tried to 
stop the daemon doing a /etc/rc/init.f/mysql stop..

This didn't quite work and then I did a: killall -9 mysql .. Yes, I know that when I 
have bad luck I can damage the Mysql tables.. 
Then I started mysql again and everything worked fine again..

Now the question: Did I did the right thing by doing a killall -9 mysql.. We (luckily) 
always make a backup every night at 1:00..
Was there a better way to get rid of the defunct process and get mysql running 
again... ??  (Eg: I want to get it running again FAST..)

Is this something to worry about.. ?? Did we do something wrong to mysql, or is it 
just the typically statistic chance that every 1000.000.000 hours 
there is a small error/bug in the hardware/software/kernel/processor etc etc and when 
this happens a programs dies and this wat the result that I saw...
(BTW: A year now this machine hasn't given me one sleepless night which I find 
unbelievable thinking that I have to restart windows at least 3 times a day..)

Should I forget this incidence, or are there chances that I need to make some 
changes.. I know that there are newer versions of MySQL.. But I was always told
to not fix things which aren't broken.. The current MySQL just works fine the way it 
is ... I only upgrade PHP a lot to get my hands on the new functions... 

Only awkward thing I can think I could think of was that we added some colums to a 
very large table.. 

Thanks for any suggestions..

Bye Bye
David Bouw

Reply via email to