strange issue with host discovery

[stuartjcarter]

sadly, no more machines are appearing as manageable in the monitoring/discovery console

[untergeek]

Not green? How does the manual discovery test work, with the zabbix_get?

[stuartjcarter]

trying to zabbix_get my desktop, it times out.

[untergeek]

Do you have the Zabbix agent running on your desktop?

[stuartjcarter]

I do indeed. Running as a service.

[untergeek]

What is the response if you try to telnet to 10050 from your zabbix server to your desktop?

[untergeek]

Also, do you have logging enabled? What is in the log file on your desktop?

The bottom line is that if no more hosts have green in the discovery monitoring pages then your hosts or network are to blame still, and/or the configuration of your Zabbix agents.

[stuartjcarter]

telnet mycomputer.domain. 10050
Trying correct_IP_address...
telnet: connect to address correct_ip_address: Connection timed out
telnet: Unable to connect to remote host: Connection timed out

[untergeek]

What that tells me is that Zabbix agent isn't running on 10050, or isn't running at all. You ought to check thoroughly (logs, etc.).

[stuartjcarter]

the log file on my desktop has a LOT of entries like

2008:20110518:134146.652 Can't parse list of active checks. Can't open JSON object ""

[untergeek]

are there any other messages besides those?

[stuartjcarter]

right at the beginning of the log file is

2812:20110513:151904.769 Starting Zabbix Agent. Zabbix 1.8.5 (revision 19047).
348:20110513:151904.789 agent #0 started [collector]
2652:20110513:151904.789 agent #1 started[listener]
2580:20110513:151904.789 agent #2 started[listener]
2384:20110513:151904.789 agent #3 started[listener]
2408:20110513:151904.789 agent #4 started[listener]
776:20110513:151904.799 agent #5 started[listener]
2780:20110513:151904.809 agent #6 started [active checks]

this sequence is repeated once, but the rest of the 3571 entries in the log file are the JSON issue

[untergeek]

Is your zabbix server actually running? The frontend will work independently of the actual Zabbix programs.

From your desktop:

telnet zabbixserver.domain. 10051

If you can't connect, that's the first problem.

[stuartjcarter]

according to the services manager in Centos, the zabbix server and client are both working.

Telnet doesn't work, as we use ssh instead.

[untergeek]

Telnet is only for doing a port test, not to log in to the server. Please execute the telnet command from your desktop to the zabbix server as previously outlined.

You can't rely on the services manager. Look in the log files, ps -ef | grep zabbix_server, etc.

--Aaron M