agent unreachable, zabbix_get works,

[kljl43]

I'm having a strange ubuntu (server) -> windows (agent) problem. The status "Zabbix agent on (windows desktop) is unreachable for 5 minutes" flaps on and off. The logs from the agent machine (2.1.1) show:

"active check configuration update from [192.168.1.151:10051] started to fail (cannot connect to [[192.168.1.151]:10051]: [0x0000274D] No connection could be made because the target machine actively refused it.)"

However, this error doesn't necessarily coincide with the error from the web frontend. In fact, this error hasn't occured in about an hour (after agent restart) but the error from the web frontend has flapped on and off 4 times already.

The server does show errors of the form:

 11963:20140413:005614.720 Zabbix agent item "weather.outdoordewpointf" on host "server" failed: first network error, wait for 15 seconds
 11966:20140413:005629.649 Zabbix agent item "weather.outdoordewpointf" on host "server" failed: another network error, wait for 15 seconds

However I can successfully run

$ zabbix_get -s 192.168.1.150 -k "agent.ping"
1
$ zabbix_get -s 192.168.1.150 -k "vfs.fs.size[D:,used]"
301073010688

from the server pointing to the agent and this is successful with or without the firewall on the agent machine on. But I have disabled the firewall during this debugging process.

I did install a switch that the two machines connect to (was a second router). I also reinstalled 2.2.2 on ubuntu (server), but did not recreate the database (didn't want to redo settings).

Any suggestions?

[kloczek]

I'm having a strange ubuntu (server) -> windows (agent) problem.
[..]
Any suggestions?

Check why you host refuses active agent connection from monitored host to your zabbix server.

PS. Checking item from zabbix server using zabbix_get simulates passive agent monitoring.

[kljl43]

Check why you host refuses active agent connection from monitored host to your zabbix server.

I'm not quite sure how to do this? Windows Eventlogs? Firewall logs? Sorry I'm not super experienced at this. But, the question has turned moot (except for others who come across this thread) because the problem seems to have resolved itself. But it is annoying that I don't know how.

[aib]

"active check configuration update from [192.168.1.151:10051] started to fail (cannot connect to [[192.168.1.151]:10051]: [0x0000274D] No connection could be made because the target machine actively refused it.)"

This message means that:
1) you configured ServerActive=192.168.1.151 in zabbix_agent.conf on windows PC.
2) You don't configured the firewall on server to accept any inbound connection to port 10051

And when the agent try to get the list of active checks, the Server's Firewall (selinux/whatever) doesn't allow him to connect.

Check the network settings about security and firewalling on server.

The server does show errors of the form:

 11963:20140413:005614.720 Zabbix agent item "weather.outdoordewpointf" on host "server" failed: first network error, wait for 15 seconds
 11966:20140413:005629.649 Zabbix agent item "weather.outdoordewpointf" on host "server" failed: another network error, wait for 15 seconds

However I can successfully run

$ zabbix_get -s 192.168.1.150 -k "agent.ping"
1
$ zabbix_get -s 192.168.1.150 -k "vfs.fs.size[D:,used]"
301073010688

from the server pointing to the agent and this is successful with or without the firewall on the agent machine on. But I have disabled the firewall during this debugging process.

There is the strange explanation of problem.
From the LOG-file I can see that the server cannot get the key "weather.outdoordewpointf".
But from the command line you try to check the absolutely different keys - "agent.ping" OR "vfs.fs.size[D:,used]"

Just imagine :
- the typical key "agent.ping" is implemented in the code of zabbix_agent and response in 1 second
- the User key "weather.outdoordewpointf" was created from scripts which include some external programs calls and will take up to 1+ minutes to respond.

When you will check "agent.ping" you will see NO any problem.
When you will check "weather.outdoordewpointf" you will see A lot of problems (because of Timeout for every call of this key).

My advice is:
- try to increase Timeout=30 in zabbix_server.conf (on server) and in zabbix_agent.conf(on client) and restart both program (server and client)
- when you debug something, try to be consistent and check the same key which create a problem.

OK?