New 1.4 install with lots of strange problems

[nelsonab]

I recently installed 1.4 on a CentOS 4.4 box.

I'm getting a large host of errors the first one is:

10095:20070620:123910 Got empty string from [netmon] IP [0.0.0.0] Parameter [agent.ping]
10095:20070620:123910 Assuming that agent dropped connection because of access permissions
10095:20070620:123910 Host [netmon] will be checked after 60 seconds

netmon is the machine which zabbix is running. The netmon host is linked to the default linux template.

Also I'm getting A LOT of the following type errors

10054:20070620:124030 Expression [{11991}>150000] cannot be evaluated [Unable to get value for functionid [11991]]
10054:20070620:124030 Expression [{12051}>150000] cannot be evaluated [Unable to get value for functionid [12051]]
10056:20070620:124031 Expression [{11992}>150000] cannot be evaluated [Unable to get value for functionid [11992]]
10057:20070620:124032 Expression [{11993}>150000] cannot be evaluated [Unable to get value for functionid [11993]]

How do I track down what this is related to? It appears that the number in the {} is incrementing.

This is a fresh install. I only have two host entries, one SNMP for a network switch and the other for the Zabbix host itself.

What am I missing?

[Alexei]

The netmon has an interesting IP address, 0.0.0.0 ....

[nelsonab]

The netmon has an interesting IP address, 0.0.0.0 ....

Yes that is true. It is set to use the DNS name not the IP.

When I set the ip address to the proper host I get the following error from within the Zabbix UI

Got empty string from [Netmon] IP [192.168.100.65] Parameter [agent.ping]

[grabber]

Hi Nelson,

did you manage to get to the bottom of this. I have the same issue


cheers

Ian

[nelsonab]

Nope still having the same problem. I can monitor SNMP and windows hosts just fine, but no matter what I do the Linux boxen don't come up.

The two linux boxes are Ticket and Netmon, here's the error.

18025:20070727:214203 Got empty string from [Netmon] IP [192.168.100.65] Parameter [agent.version]
18025:20070727:214203 Assuming that agent dropped connection because of access permissions
18025:20070727:214203 Host [Netmon] will be checked after 60 seconds
18025:20070727:214203 Got empty string from [ticket] IP [0.0.0.0] Parameter [agent.ping]
18025:20070727:214203 Assuming that agent dropped connection because of access permissions
18025:20070727:214203 Host [ticket] will be checked after 60 seconds

The windows agents work fine without the ip address, and I've never had to add them using 1.1.7.

I did an nmap on both hosts and show the zabbix agent running on the proper port. I've tried to telnet to the agent but the telnet connection closes right away, this is not the case for the windows agents. The windows agents return "ZBXDZBX_NOTSUPPORTED" when I press enter which is expected. I also show the zabbix agent as running 6 processes which is expected and the agent log file shows nothing out of the ordinary.

I'm out of ideas and searching the forum turns up nothing. What am I missing?

[nelsonab]

bump...

Any ideas for troubleshooting?

[Alexei]

Perhaps Server in zabbix_agentd.conf does not match ZABBIX server IP address.

[nelsonab]

I will double check this just in case. I know the agent port numbers are commented out so the default 10051 is used (if my memory is working correctly). One would expect that you could telnet to the agent right? I know I've been able to do this before to ensure the agent is working.

I've also checked the DNS from the command line on the server and the addresses returned are correct for the systems in question.

What else could I be missing?

[nelsonab]

Ok turns out it was a multiple of things that was happening.

I reset the server string to be the name of the system, not the IP address and I made sure to set the ListenIP string to be the main IP of the host. Before the server string was the IP of the server and the ListenIP was commented out. Why this would cause a change does not make sense to me as it would seem both configurations would cause the same results, unless our DNS was not returning the hostname from the IP name which is quite possible.

On our "ticket" system I did much of the same thing and it turns out I gorked the IP address for the server anywho. What I wasn't seeing in the logs until I bumped up the debug level was a permissions error. Since I wasn't telneting in from netmon (Zabbix server) the agent would refuse the connection. I like that, but perhaps such messages might be useful at debut level 3?

In the end I take much of the blame (as expected on my part). I was getting confused by a lack of coherent debug information and understanding of the need for a correct server string for security reasons. Lessons learned. :-)