Possible Bug: Zabbix Server loses Connectivity to Agent on SNMP Monitored Hosts

[rob9988]

Version:
Zabbix 1.1 (server and agent)

Server OS:
SuSE Enterprise Linux 9 SP3 (9.3)

Client OS:
SuSE Enterprise Linux 9 SP3 (9.3)

Problem:
Zabbix server loses all connectivity to Zabbix agent when SNMP monitored items do not respond (i.e. the application which reports status via SNMP is not running).

Details:
- Server and agent were working fine with a mix of SNMP and agent items
- Server was rebooted to implement new changes to monitored app; the app failed to restart, so SNMP data were no longer available
- Zabbix server could no longer connect to the agent. Message: "Timeout while connecting to [host.ourdomain.com]:nnn" (where nnn=the port number which we use for SNMP monitoring)
- Telnet to the host on port 10050 worked, and returned a ZBX error

Workaround:
Disabling all SNMP items for the monitored host resulted in the Zabbix server being able to immediately establish a connection to the agent and once again start collecting data.

Anyone else experiencing this bug?

-Rob


PS: Here's a link to my response to another post about this problem:

http://www.zabbix.com/forum/showthread.php?t=3727

[dantheman]

I had this with a handful of cisco switches, I had added an SNMP item to monitor the temperature and then 7 of my 9 switches become unreachable and none of the other items were collected, once I deleted the temperature item (none of my switches supported it ) then I was getting the rest of the snmp data again. It's not my only unsupported SNMP item either.. I have others that are not causing problems.