Bug #9086 crond mysqld_safe <defunct> process

• Submitted: 10 Mar 2005 3:42
• Modified: 31 May 2005 15:30
• Reporter: Faris Xiao
• Status: Can't repeat
• Category: MySQL Server
• Severity: S3 (Non-critical)
• Version: 4.0.23a
• OS: Linux (RedHat AS 3)
• CPU Architecture: Any

[10 Mar 2005 3:42] Faris Xiao

Description:
I use crond to start mysql.

35 11 * * * /app/mysql/bin/mysqld_safe  --user=mysql &

but after the crond execute this,my Operating system appear several 
<defunct> processes.

ps -efw

root      2330  1416  0 11:35 ?        00:00:00 CROND
root      2331  2330  0 11:35 ?        00:00:00 [sh <defunct>]
root      2332     1  0 11:35 ?        00:00:00 /bin/sh /app/mysql/bin/mysqld_safe --user=mysql
mysql     2358  2332  0 11:35 ?        00:00:00 /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.

why there is a [sh <defunct>]?

Is this a bug of mysqld_safe ?

How to repeat:
always repeat

[13 Mar 2005 20:27] Aleksey Kishkin

Hi! could you please run

ps axf

and provide here  output of this command with defunct processes?

[14 Mar 2005 4:25] Faris Xiao

3612 ?        S      0:00 crond
28753 ?        S      0:00  \_ CROND
28754 ?        Z      0:00      \_ [sh <defunct>]
28776 ?        Z      0:00      \_ [crond <defunct>]
28755 ?        S      0:00 /bin/sh /app/mysql/bin/mysqld_safe --user=mysql
28779 ?        S      0:00  \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mysql --pid
28780 ?        S      0:00      \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mysql -
28781 ?        S      0:00          \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mys

[14 Mar 2005 4:25] Faris Xiao

3612 ?        S      0:00 crond
28753 ?        S      0:00  \_ CROND
28754 ?        Z      0:00      \_ [sh <defunct>]
28776 ?        Z      0:00      \_ [crond <defunct>]
28755 ?        S      0:00 /bin/sh /app/mysql/bin/mysqld_safe --user=mysql
28779 ?        S      0:00  \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mysql --pid
28780 ?        S      0:00      \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mysql -
28781 ?        S      0:00          \_ /app/mysql-4.0.23/libexec/mysqld --basedir=/app/mysql-4.0.23 --datadir=/app/mysql-4.0.23/var --user=mys

[14 Apr 2005 23:00] Bugs System

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

[15 Apr 2005 14:42] Sergei Golubchik

reopened

[31 May 2005 15:30] Jorge del Conde

Thanks for your bug report.

I was able to reproduce this with Fedora Core2 and sysV startup script:

There where the lines that I used for my crontab:

52 * * * * root /etc/rc.d/init.d/mysql.server start

root-/etc# ps fax | grep mysql
11721 pts/10   S      0:00                      \_ grep mysql
11685 ?        Z      0:00      \_ [mysql.server] <defunct>

But I was unable to reproduce it using mysqld_safe directly (which is used by my 'mysql.server' script).