Watchdog and MySQL

[MassimoP]

For a strange reason, after a reboot the Watchdog module does not recognize the mysql process (even if it is started and it is working) and tries to restart it.
Actually a new mysql process is started thus blocking mysql, the following lines appear in the mysql log file:

InnoDB: Unable to lock ./ibdata1, error: 11
InnoDB: Check that you do not already have another mysqld process
InnoDB: using the same InnoDB data or log files.

After this point, mysql stops to accept connections.

I don't understand why, but it is a weird behaviour of the mysql script which usually does not start another service instance.

The sequence of operations is:
1) start mysql at boot, it works perfectly
2) watchdog checks mysql, does not recognize its presence and tries to restart it
3) mysql start a new instance that block connections

any idea?

 

[IgorG]

Check in /usr/local/psa/etc/modules/watchdog/monitrc that path to MySQL pid file is correct.

 

[MassimoP]

Thanks, I checked the path to the pid file and it is correct...

The problem arises only after a reboot.

 

[MassimoP]

Just to add some information... The problem does not arise if the wathdog is disabled...

 

[MassimoP]

Main caracteristics of the server in which I have this problem are
CPU AuthenticAMD, Quad-Core AMD Opteron(tm) Processor 2372 HE
Version Parallels Plesk Panel v10.0.1_build109101029.19 os_SuSE 11.1
OS Linux 2.6.27.54-0.1-pae

 

[Justin Clarke]

For a strange reason, after a reboot the Watchdog module does not recognize the mysql process (even if it is started and it is working) and tries to restart it.
Actually a new mysql process is started thus blocking mysql, the following lines appear in the mysql log file:

InnoDB: Unable to lock ./ibdata1, error: 11
InnoDB: Check that you do not already have another mysqld process
InnoDB: using the same InnoDB data or log files.

After this point, mysql stops to accept connections.

I don't understand why, but it is a weird behaviour of the mysql script which usually does not start another service instance.

The sequence of operations is:
1) start mysql at boot, it works perfectly
2) watchdog checks mysql, does not recognize its presence and tries to restart it
3) mysql start a new instance that block connections

any idea?

I am using Monit (instead of the built-in Watchdog) and this does exactly the same thing! I have to setup Monit on an alert only basis for MySQL otherwise it breaks when it tries to restart that process.

Every time I start the Monit process (either manually or by rebooting my server) I get an alert from Monit saying that:

Connection failed Service mysql

Date: Thu, 18 Nov 2010 13:40:27 +0000
Action: alert
Host: myserver.info
Description: failed, cannot open a connection to INET[127.0.0.1:3306] via TCP

Your faithful employee,
Monit

 

[MassimoP]

I think that Whatdog and Monit are the same software...

As requested by the forum moderator, please post your server main information, in this way a resolution of this bug could be faster.

 

[Justin Clarke]

I think that Whatdog and Monit are the same software...

As requested by the forum moderator, please post your server main information, in this way a resolution of this bug could be faster.

Plesk Parallels Panel v10.0.1

CentOS - Linux 2.6.9-023stab052.4-smp

VPS running AuthenticAMD, Quad-Core AMD Opteron(tm) Processor 2352

1GB RAM (my share)

10GB HDD (my share)

 

[German Wider]

Hi! Have you had any luck solving the problem? Same issue here.

 

[BenjS]

Same here with Debian 6.0 and Plesk 10.2.
On every reboot, something wants to start a second mysql process. When I stop all mysql processes and start it again, it works.

Any solution?