False Alerts

[warren markley]

Hi,

We are receiving false alerts that a service is down when it is in fact running. This is only cleared when the zabbix service is restarted.

The logs contain the following message:
32458:20141106:011936.699 [24] Too many open files
32458:20141106:012003.203 [24] Too many open files
32458:20141106:012003.747 [24] Too many open files
32457:20141106:012338.564 [24] Too many open files

I have ran a "lsof" against the zabbix pid's where the following appears:
zabbix_ag 25820 zabbix 322u sock 0,6 0t0 50825793 can't identify protocol
zabbix_ag 25820 zabbix 323u sock 0,6 0t0 50826900 can't identify protocol
zabbix_ag 25820 zabbix 324u sock 0,6 0t0 50828415 can't identify protocol
zabbix_ag 25820 zabbix 325u sock 0,6 0t0 50830207 can't identify protocol
zabbix_ag 25820 zabbix 326u sock 0,6 0t0 50831969 can't identify protocol
zabbix_ag 25820 zabbix 327u sock 0,6 0t0 50833873 can't identify protocol
zabbix_ag 25820 zabbix 328u sock 0,6 0t0 50835141 can't identify protocol
zabbix_ag 25820 zabbix 329u sock 0,6 0t0 50836270 can't identify protocol
zabbix_ag 25820 zabbix 330u sock 0,6 0t0 50837761 can't identify protocol

Does anyone have any insight into this?

Cheers.

[alxumuk]

same issue(