False "unreachable" Alerts

[m0773w]

My Zabbix server monitors about 150 individual servers. Among the alerts that may be sent out is when the server becomes unreachable - usually if the server has crashed or has been rebooted.

Recently there have been "unreachable" alerts sent out for a handful of servers which are actually up and running. I tried stopping and restarting the Zabbix service on those boxes, but to no avail. I restarted one of them and this seemed to help, but afterwards a new bunch started sending out these alerts. And restarting every box that has this issue is, at this time, impractical.'

Does anyone know what may be causing this and what the remedy is?

Any help would be appreciated.

[smitty]

I have a similarly sized environment, running on zabbix 1.8.4. What I sometime see is that a machine will get "stuck" in host unreachable as well (typically the same one or two hosts). This may originally be due to an actual reboot, however the stays in "host unreachable" state until I manually disable the trigger, wait a minute, then re-enable the trigger. Seems similar to your issue..? I suspect that if I removed the host completely from zabbix and reinstalled that the problem would be fixed, but I haven't had a chance to do this. I'm not sure of the cause, but I suspect that it's a bug...