problem monitoring/restarting zabbix_server and agents

[ShivaS]

Hi, we've got a very crytical system that i am monitoring with zabbix.
Lately i have a problem that zabbix_server fails.
However i did a simple crontab script to try start server every 10 minutes.
In case it runs it's ok.
In case not, then i have a problem (also with agents)
there is a pid file that is never deleted.and i need delete it manually.
However i agree that i may write a script to check server/agent and if not running then try delete pid file and start it again.
But may be there is another solution.
Also we agents i noticed that it usually takes time to start it.
i always delete pid, start and check process. and it's not running.
again deleting pid and again the same loop until i notice it running.
The thing makes me crazy.
Is there any solution for it?
does someone experienced similar problem and have custom script?
thanks

[elkor]

Your problem is most likely network sockets that haven't closed yet.

You can try my startup script from the end of this thread (http://www.zabbix.com/forum/showthread.php?t=1281) and see if that helps. I wrote it to parse startup parameters from the zabbix config file and to wait for open sockets to close before attempting agent startup in order to get around that issue.

it assumes your agent is installed in \zabbix\bin and that your config files are in \zabbix\etc. With minor editing it should work for the server as well.