Error: ZBX_TCP_READ() on newly added agent

[Wayne]

Hi guys,

Wondering if you can help with an issue that we're having with a new host that we just put online.

Zabbix Server/frontend 1.4
OS - CentOS 4.5

We just installed the latest agent (1.4.1) on a new host, and are seeing the following error in the zabbix_server.log file:

9994:20070804:195256 Timeout while answering request
9994:20070804:195256 Get value from agent failed. Error: ZBX_TCP_READ() failed [Interrupted system call]
9994:20070804:195256 Host [server99.host-99.biz]: another network error, wait for 15 seconds

Some notes:

- I've double and triple checked the configuration, as well as disabled the firewall on the agent server altogether to be sure it was not a firewall issue.
- IP for zabbix server is set correct in zabbix_agentd.conf on the agent server
- Bandwidth, CPU load, etc is all reported back to zabbix just fine on this host. But we get the above error, and "Server hostname is unreachable" triggers being sent out.
- All other hosts are reporting just fine (1.4 agents)

Any thoughts?
Does our zabbix server need to be upgraded to the latest build? I was under the impression that it was backwards compatible, but perhaps not. I'd just rather not upgrade at this point if it is not the solution.

Any help is really appreciated.

Thank you!

[Wayne]

As an update, we just added another host to the monitoring and fully configured it with zabbix agent 1.4.0 (same build as our zabbix server), and we are receiving the same error in the logs.

Anyone have some thoughts on what might be causing this? I have 6 hosts being monitored without issue, and these last 2 hosts are giving errors. I can't find anything different configuration-wise between the working and non-working hosts.

[RNHurt]

I was having this same problem on a newly added agent and the problem was that the server information was incorrect in the /etc/zabbix/zabbix-agentd.conf file. Because this box is going to be outside our firewall I used the external DNS name but not the internal IP address. Once I added the 192.168.x.x address of the server it got past this error. I'm having other problems but I think they are unrelated to this issue.

My suggestion is to re-double check your agent settings and make sure you can communicate to the server on both ports 10050 and 10051.

Later...
Richard

BTW: The server setting on the agent config file is very sensitive. Make sure you remove any blank spaces or it wort work.