Zabbix Server doesn't Start on startup

[alexandrevvo]

I'm having troubles in making the zabbix-server auto-start to work.

Looking at the log and researching, I could see that the problem is: the initialization fails because the Mysql didn't have started already;

783:20110324:091203.388 Starting Zabbix Server. Zabbix 1.8.2 (revision 11211).
   783:20110324:091203.392 **** Enabled features ****
   783:20110324:091203.392 SNMP monitoring:       YES
   783:20110324:091203.392 IPMI monitoring:       YES
   783:20110324:091203.392 WEB monitoring:        YES
   783:20110324:091203.392 Jabber notifications:  YES
   783:20110324:091203.392 ODBC:                   NO
   783:20110324:091203.392 SSH2 support:           NO
   783:20110324:091203.392 IPv6 support:          YES
   783:20110324:091203.392 **************************
   783:20110324:091203.392 [Z3001] Connection to database 'zabbix' failed: [2002] Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)

I saw that in some cases, configuring the script os zabbix-server on /etc/init.d solved the problem. And I did the indicated:

#! /bin/sh
### BEGIN INIT INFO
# Provides:          zabbix-server
# Required-Start:    $remote_fs $network
# Required-Stop:     $remote_fs
[U][B]# Should-Start: mysql
# Should-Stop:  mysql[/B][/U]
# Default-Start:     2 3 4 5
# Default-Stop:      0 1 6
# Short-Description: Start zabbix-server daemon
### END INIT INFO

But didn't work..
The problem happens after reboot or power off.

running the script manually works fine. (sudo /etc/init.d/zabbix-server);

Any ideas???

Thanks.

ps.: Zabbix 1.8.2 on Ubuntu server 10.10

[DSon]

Workaround..

I had this problem, so I:

- deleted the startup symbolic links (not the shutdown ones though) for zabbix server and agent
- added two startup commands for zabbix-server and agent in /etc/rc.local

Since the rc.local file is the last to be run in the boot up sequence, there should be time for the mysql service to get started.

Hope this helps,
Danny.

[Colttt]

can you please send the output from this command:

ls /etc/rc*.d/

and i would like to know which runlevel do you use..

[DSon]

Oops!

Hi there, sorry for the late reply.

I just realised something!

The method I suggested (using rc.local) is primarily for versions up to and including 8.04.

I noticed that you're using 10.10 - doesn't this make use of "upstart" scripts?

If so, you need to do what I did, which was to search this forum for:

ubuntu upstart script

You should find all you need to know in that thread (there is only one that I could find).

Hope this helps,
Danny.

[alexandrevvo]

Tks

Thanks DSon..

I think i solved using the rc.local, but I had to put a sleep 2 at the script before calling the service..
Its not a good solution but worked..

By the way I'll search for: ubuntu upstart script, and see how it will work..

Thanks a lot for the replies...

Anything new I'll post here..

[Tekkell]

startup

alexandrevvo,

Make sure your MySQL is running before your zabbix_server starts.

in my /etc/rc3.d MySQL starts at S64 but my zabbix_agent and zabbix_server start at S55

This is after a clean install and all default. Zabbix wants the database to be running when it starts up, but as 55 is lower then 64 it does not and complains.
My installation is on CentOS 5.6

changing the startup sequence:

# chkconfig mysqld off
# vi /etc/init.d/mysqld

find the line with chkconfig and change it to 50 50 (line:6)

# chkconfig mysqld on

at next reboot the database starts before the zabbix server.

maybe your distribution is different, but the concept is the same

rgrds,

Wim

[efrain02]

I have this problem too. The difference is that zabbix only is connected a few seconds and then automatically gets disconnected. Hope someone could help me with this problem.

[Tekkell]

efrain02,

Please be more specific. Logs?

rgrds,

[efrain02]

This is my zabbix-server.log file:

[Tekkell]

efrain02,

First of all I think you should remove your zabbix pid-files in /tmp while your zabbix is stopped.
The bottom things are (temporary?) connection problems to devices you want to monitor.
I cannot see zabbix failing.

rgrds,

[efrain02]

Ok thanks . Now i don't have any pid-files. And yes, the bottom things are devices i want to monitor. Now my log file is like this:

In the blue rectangle area, it tells me to increase the cache size and i have already increased to 38M just to see what happens. But the screen shows the same. And i can't see any other error.

[Tekkell]

Please try the TrendCacheSize, that one is 4M.

http://www.zabbix.com/documentation/.../zabbix_server

rgrds,

[g.vecchi]

Here my /etc/init/zabbix-server.conf:

# zabbix-server - Start zabbix server
description     "Zabbix Server"
author          "Vecchi Giovanni"

pre-start script

	if [ ! -d /var/run/zabbix-server ]; then
        	mkdir -p /var/run/zabbix-server
    		chown zabbix:zabbix /var/run/zabbix-server
    		chmod 755 /var/run/zabbix-server
	fi

	ZabbixConfig=/etc/zabbix/zabbix_server.conf
	DBUser=""
	DBPassword=""
	DBName=""


	if [ -f $ZabbixConfig ]
	then
		DBUser=`cat $ZabbixConfig | grep '^DBUser='`
		DBUser="${DBUser##*=}"
		DBPassword=`cat $ZabbixConfig | grep '^DBPassword='`
		DBPassword="${DBPassword##*=}"
		DBName=`cat $ZabbixConfig | grep '^DBName='`
		DBName="${DBName##*=}"

		if [ -n "$DBUser" ] && [ -n "$DBPassword" ]
		then

			while [ ! -S /var/run/mysqld/mysqld.sock ]
			do
				sleep 1
			done

			i=0

			while [ $i -le 5 ]
			do
			mysql -u$DBUser -p$DBPassword $DBName -e "SELECT \"1\"" &> /dev/null
			MYSQL_EXIT_CODE=$?

				if [ $MYSQL_EXIT_CODE=0 ]
				then
			        	i=6
		        	else
			        	sleep 5
			        	i=`expr $i + 1`
		        	fi
			done
    		fi
  	fi

end script

start on (runlevel [2345] 
	  and started mysql)
stop on runlevel [016]
respawn
expect daemon
exec /usr/sbin/zabbix_server

it works on Ubuntu 10.04.4