ZABBIX Agent seems broken

[froztbyte]

Hi

Perhaps someone else can help with this. I've got an installation based on 1.6.4 with agents running on the servers I'm monitoring. One server I have continues to go up and down in ZABBIX with the alert. The cause of this is that ZABBIX checking the agent fails to see that it is alive. The server is up, and has not been down in any of the times ZABBIX indicates it is. The agent *is* responding and an open Telnet session to it stays connected for a full hour without issues. The ZABBIX server runs a the host.status check, and then cycles between Unreachable and Up as follows (this paste is from the "Plain Text" display under graphs):

2009-06-18 10:13:12 1245312792 0
2009-06-18 10:10:11 1245312611 2
2009-06-18 10:06:35 1245312395 0
2009-06-18 10:05:34 1245312334 2
2009-06-18 10:04:43 1245312283 0
2009-06-18 10:02:43 1245312163 2
2009-06-18 10:01:43 1245312103 0
2009-06-18 09:59:42 1245311982 2
2009-06-18 09:58:42 1245311922 0
2009-06-18 09:55:41 1245311741 2
2009-06-18 09:54:47 1245311687 0
2009-06-18 09:52:47 1245311567 2
2009-06-18 09:51:47 1245311507 0
2009-06-18 09:48:46 1245311326 2
2009-06-18 09:47:56 1245311276 0
2009-06-18 09:45:54 1245311154 2

Has anyone experienced similar behaviour? This started about two days ago and only happens with this host.

[MrKen]

For an explanation of host.status see this post by Alexei: http://www.zabbix.com/forum/showpost...2&postcount=10

Probably a better way to check whether the host is up is just to use agent.ping, with a trigger like {host:agent.ping.last(0)}#1

MrKen