Host uptime changes by +/- 3 days everytime it is checked

[lodperera]

Hi Boys,

I've got a weird one.

I have several linux servers monitored using Template Active OS Linux.
But only 1 server(megachile) reports incorrect uptime everytime zabbix runs the check.

There is a pattern where the actual uptime is 3 days, 23:21.
Zabbix says it is 3 days, 23:21 in latest data and changes it on the next check to 6 days, 23:28:49

So the megachile uptime graph looks like a zig zag(attached).
All the other servers do not have this issue.

I ran the following commands on the monitored server(megachile) to figure out whether the agent is playing up.

root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344192]
root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344201]
root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344212]
root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344222]
root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344232]
root@Megachile:~# zabbix_agentd -t system.uptime
system.uptime                                 [u|344245]

root@Megachile:~# zabbix_agentd -V
zabbix_agentd (daemon) (Zabbix) 3.0.3
Revision 60173 18 May 2016, compilation time: May 23 2016 09:44:47

[root@zabbix ~]# zabbix_server -V
zabbix_server (Zabbix) 3.0.3
Revision 60173 18 May 2016, compilation time: May 23 2016 08:57:57

Tried to figure out for almost a week, any help to figure out the issue is most appreciated.

[zabx_ks]

Hey lodperera,

Have you checked time on your host? Maybe there is not correctly set up date.

Which Linux distribution is using that host?

And is there any other information in host agent log?

zabx_ks

[LenR]

I've seen this where the items were agent active and 2 hosts were configured with the same host name.

Both poll Zabbix for the list of items to check, both update, zig/zag because they have different values.

Easy to find, graph the uptime, then change to last 500 values, if you have item update more frequently than your update interval, you have twins.

[lodperera]

two servers with the same hostname

Hi LenR,

That was exactly the issue.
There were two servers with same hostname. Actually I cloned one for testing and forgot to disable zabbix-agent on the cloned server.
Thanks mate,

PS: zabx_ks thank you for replying.