What happened to all my Windows agents? 1.4.4

[barronj]

I installed late last week, and as of Friday I had about 12 Windows agents (1.4.4) set up and sending data to my Fedora8 Zabbix (1.4.4) server. WIth this pool of "pilot" agents, I intended to set about really getting Zabbix set up for my needs before moving forward with more installs.

I arrived on Monday and none of them were sending data any more.

None of the agent logs show a problem.

The Zabbix server log shows a line like this for every Windows agent:

2460:20080305:163200 Get value from agent failed. Error: ZBX_TCP_READ() failed [Connection reset by peer]
2460:20080305:163200 Host [SOMEHOSTNAME] will be checked after 60 seconds

I have searched the forum for similar problems, but nothing seems to suggest a solution.

I did find a thread here where someone mentioned that none of their Windows agents would work after a few days, so I installed on another Windows box to see if it would work for awhile, but it hasn't worked at all. Again, the agent log seems to indicate that the agent thinks everything is fine.

I didn't think there were any firewalls between me and the hosts, and a quick call to our "Network Services" department confirmed that this is so, and that nothing has been changed on the network that would affect my access. I *own* the hosts being monitored (from a support and configuration standpoint), so I know that nothing has been changed on that end.

The only Linux agent I'm running is on the Zabbix server itself, and that agent seems to be working fine.

Any ideas? My current log level is "3".

Thanks!

[Alexei]

Try to connect to the agent (TCP port 10050) from ZABBIX server box and from a local Windows server when the agent is running.

[barronj]

I was already certain that port was open, but on the way to double checking it I discovered that the FQDN of my Zabbix server (which I used in the .conf file for the agents) will no longer resolve on our WAN. I don't know why this is the case. However, changing the .conf file to use the IP address fixed the problem. I can't fathom why it was fine for a few days then stopped, but there you go...

I still have an issue to resolve, since the server uses DHCP, and I currently don't have the option of using a static address, but for now I can continue to get things set up.

Thanks!