-
Type:
Incident report
-
Resolution: Fixed
-
Priority:
Major
-
None
-
Affects Version/s: 1.8.3
-
Component/s: Agent (G), Server (S)
-
None
-
Environment:Server: 1.8.3
Agent(active): 1.8.3 on Windows XP, multihomed(eth0 -LAN, eth1 - VPN connection)
Network devices in between: router-modem Anydata AWR500a
We have a bunch of remote hosts with the following:
1) PC is multihomed, two interfaces, one is real (LAN) , other one is virtual - VPN. Zabbix agent can reach zabbix server both ways: over LAN or over VPN. When the zabbix server was 1.8.2 , zabbix agent was configured to use LAN address for server: zabbix.company.com
2) After server was upgraded to 1.8.3 - all those agents went offline: restarting services or network devices did nothing. The solution was to change server name in zabbix agent.conf to VPN address of the server: 10.1.0.1. Hovewer:
3) Now, once in a while, all those agents go offline: the solution is to connect to host via telnet and issue the command:
net stop "zabbix agent" && net start "zabbix server"
4) In logs with debugging on I see countless:
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
In collect_cpustat()
5) Also, all those PCs are connected to Internet (and zabbix server) via 3G-like modem-router Anydata AWR500a. That's to note since there are other similar multihomed PC with the only difference that they are connected via Wimax routers and they are fine.
