One client out of 221 refuses to connect

[slinx]

Hello,

We have a ZABBIX Server (daemon) v1.6.5 (revision 7442) (3 April 2009) server monitoring 221 clients.

All of the clients have the same configuration, running ZABBIX Agent (daemon) v1.4.5 (25 March 2008).

All of them report into the server EXCEPT ONE.

I can telnet to port 10050 on the client from the server, send a zabbix key name, and get a response.

I can telnet to port 10051 on the server from the client.

Here's the client's log:

  9686:20100216:142349 In refresh_metrics('10.98.2.35',10051)
  9686:20100216:142349 get_active_checks('10.98.2.35',10051)
  9686:20100216:142349 Sending [ZBX_GET_ACTIVE_CHECKS ta365]
  9686:20100216:142349 Before read
  9686:20100216:142349 In parse_list_of_checks() [ZBX_EOF]
  9686:20100216:142349 In disable_all_metrics()
  9686:20100216:142349 Parsed [ZBX_EOF]
  9686:20100216:142349 In process_active_checks('10.98.2.35',10051)
  9686:20100216:142349 In get_min_nextcheck()
  9686:20100216:142349 Sleeping for 60 seconds
  9686:20100216:142449 In process_active_checks('10.98.2.35',10051)
  9686:20100216:142449 In get_min_nextcheck()
  9686:20100216:142449 Sleeping for 60 seconds
  9686:20100216:142549 In refresh_metrics('10.98.2.35',10051)
  9686:20100216:142549 get_active_checks('10.98.2.35',10051)
  9686:20100216:142549 Sending [ZBX_GET_ACTIVE_CHECKS ta365]
  9686:20100216:142549 Before read
  9686:20100216:142549 In parse_list_of_checks() [ZBX_EOF]
  9686:20100216:142549 In disable_all_metrics()
  9686:20100216:142549 Parsed [ZBX_EOF]
  9686:20100216:142549 In process_active_checks('10.98.2.35',10051)
  9686:20100216:142549 In get_min_nextcheck()
  9686:20100216:142549 Sleeping for 60 seconds

Here's the agentd.conf file (blank lines and comments stripped):

Server=10.98.2.35
Hostname=ta365
StartAgents=5
DebugLevel=4
PidFile=/var/tmp/zabbix_agentd.pid
LogFile=/tmp/zabbix_agentd.log
Timeout=3

Yet this is what I see in the server log for this client:

 30013:20100216:143956 Item [ta365:controllerRollupStatus.1] error: Timeout while connecting to [ta365.ta.com:161]

Processes look normal:

[root@ta365 ~]# ps -ef | grep zabbix
zabbix    9680     1  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
zabbix    9681  9680  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
zabbix    9683  9680  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
zabbix    9684  9680  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
zabbix    9685  9680  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
zabbix    9686  9680  0 Feb15 ?        00:00:00 /usr/local/sbin/zabbix_agentd
root     23099 14409  0 14:39 pts/3    00:00:00 grep zabbix

All the other clients are working fine. What could I be missing?

Thanks!

[Alexei]

Zabbix server tries to connect to port 161, which is supposed to be used by SNMP agents! You may want to check what port number is configured for this host in Zabbix GUI.

[slinx]

In the GUI, it is configured to use port 10050.

[Alexei]

What about item with a key 'controllerRollupStatus.1'. It seems to be SNMP based, right?

In version 1.6.x Zabbix considers host as unavailable if at least one check constantly timeouts. So, make this item (and any other broken items) work or disable it.

[slinx]

What about item with a key 'controllerRollupStatus.1'. It seems to be SNMP based, right?

In version 1.6.x Zabbix considers host as unavailable if at least one check constantly timeouts. So, make this item (and any other broken items) work or disable it.

Thanks Alexei. I had to install the net-snmp packages and run /etc/init.d/dataeng enablesnmp , and restart the dataeng script. It is working now.