Help! Windows Log Template/trigger have gone beserk.

[mechanum]

Hi!
Yesterday i upgraded to 1.8.8 (from 1.8.4) and compiled from source on Debian 5.04.
Everything seems to work ok except for one thing...
I added the Windows Logging template to my hosts (Windows Servers). And it began to constantly sending alerts.
I have removed the template from the hosts, deactived the triggers and finally deleted the template itself, restarted the agents, zabbixserver, and deactivated the alert but nothing helps.
Im getting desperate, where do i look? I have gone through all triggers and host.
The alerts is alernating versions OK and PROBLEM from the logs:
But no hostname or IP. i have {HOSTNAME} {IPADDRESS}
Application Log ERROR: OK
Application Log ERROR: PROBLEM
System Log ERROR: PROBLEM


Best Regards
Anders

[Robert Wagnon]

Probably a performance problem

That looks like the classic problem with Windows Event Logs from prior to 1.8.8. Look at your mysql performance - it is probably using 99% CPU. When this happens, zabbix begins failing all triggers in the system.

Zabbix 1.8.8 is intended to improve upon this situation, but perhaps it isn't a complete resolution.

[mechanum]

Re: Probably a performance problem

Hi.
Nope the CPU is almost idle, the problem is that im still getting alerts even when no hosts is linked to the template and i even deleted it but to no avail.
( before i deleted it i tested just to disable the triggers).
I also restarted the server. Nothing helps...
And the strange thing is that i cant se this in the Dashboard / Monitor Triggers view.
Best Regards
Anders

[Robert Wagnon]

CPU and mysql

Although my suggestion may still be wrong, let me clarify our situation.

We have zabbix on one server and mysql on another server.

The zabbix server will be at 5% utilization. The mysql server will be at 100%.

The Windows Event Log triggers work, but if too many come in the mysql can't keep up and all Alarms start changing between PROBLEM/OK until the load clears. Very disruptive to monitoring.