Timeouts when adding multiple hosts

[Wayne]

I'm just starting to get Zabbix up and configured for our network, but having an issue anytime I add a second monitored host into the mix.

I added one server, and have been logging data for 24 hours successfully and as expected. Whenever I start to monitor a second server, data immediately stops being logged for the first server, and the second will log very spotty (if at all).

Now, at first I hadn't configured zabbix_agentd.conf properly on each monitored host, as I left 'localhost' in as the hostname. This was working ok with one host, but when I add the second I figured that was causing the issues. I have since corrected this, but unfortunately am having the same results.

Logging on the zabbix server shows timeouts. There are quite a few timeouts actually when only running the one server, but as soon as I add the second they both timeout.

zabbix_agent.conf is configured on each host with the Zabbix server IP, the local short hostname, and I have increased timeout to 25 on each of the hosts.

All of these servers are on the same network, and it is not a firewall issue as I have added ports 10050 and 10051 to allow connections, and even disabled the firewall on each machine at one point to rule that out.

Anyone have an idea of what I can check config wise?

[Wayne]

Ok, I've got the timeout situation handled, it seems there was something I missed on the firewall settings.

I've enabled debug on all machines, and it data is passing back and forth between the Zabbix server, and both hosts (both monitored at once).

Maybe it's a graphing issue only. Hmm..

[Wayne]

Ok, I've spent some time watching the logs on all servers, and now that I've got my firewall issue squared away, monitoring seems to be fine between the hosts and Zabbix server.

Data values are sent TO Zabbix server just fine, I can see them in the output from the hosts log via tail.

Graphing bandwidth is a clear difference in accuracy however once I switch from monitoring one host, to monitoring two hosts. Is this normal?

On both of these screenshots below, the portion of the graph on the left is when I was *only* monitoring this host, the portion on the right is when I started monitoring *two* hosts. The closeup shot is hourly detail (where you can barely see any data after I started monitoring 2 hosts), the other is a 'screen' with both incoming and outgoing bandwidth - same comparison of before and after... but you can clearly see more data in this view than looking at an hourly view.

Maybe it's just me, and this is normal?