Zabbix 1.5.5 build 5973 - All Zabbix active checks failing

[disgruntleddutch]

Currently using 1.5.5 build 5973 on the server and Zabbix 1.5.3 on the host (1.5.4 crashes....).

I changed all system.cpu.util to Zabbix active checks after they were passive for about 2 hours for all 400 Linux based hosts. They were reporting fine with passive checks but stop functioning with Zabbix active checks.

The queue shows all the checks building up and sitting in there marked 'red' after 10 minutes. However the host side is seeing the active checks and is dutifully sending back to the server.

  1745:20080910:203436 Active check [system.cpu.util[,wait,avg1]] is not supported. Disabled.
  1745:20080910:203441 Info from server: Processed 4 Failed 0 Total 4 Seconds spent 0.007925
  1745:20080910:203441 OK
  1745:20080910:203606 Info from server: Processed 1 Failed 0 Total 1 Seconds spent 0.000529
  1745:20080910:203606 OK
  1745:20080910:203611 Info from server: Processed 3 Failed 0 Total 3 Seconds spent 0.001247
  1745:20080910:203611 OK
  1745:20080910:203736 Info from server: Processed 1 Failed 0 Total 1 Seconds spent 0.000922
  1745:20080910:203736 OK
  1745:20080910:203741 Info from server: Processed 3 Failed 0 Total 3 Seconds spent 0.001380
  1745:20080910:203741 OK
  1745:20080910:203906 Info from server: Processed 1 Failed 0 Total 1 Seconds spent 0.001061
  1745:20080910:203906 OK
  1745:20080910:203911 Info from server: Processed 3 Failed 0 Total 3 Seconds spent 0.008165
  1745:20080910:203911 OK
  1745:20080910:204036 Info from server: Processed 1 Failed 0 Total 1 Seconds spent 0.000907
  1745:20080910:204036 OK
  1745:20080910:204041 Info from server: Processed 3 Failed 0 Total 3 Seconds spent 0.001377
  1745:20080910:204041 OK
  1745:20080910:204206 Info from server: Processed 1 Failed 0 Total 1 Seconds spent 0.000831
  1745:20080910:204206 OK
  1745:20080910:204211 Info from server: Processed 3 Failed 0 Total 3 Seconds spent 0.001287
  1745:20080910:204212 OK

They aren't being processed by the server. Hostname is set in the zabbix_agentd.conf file and I am able to telnet to zabbix's server IP port 10051 from the host.

[Palmertree]

If I am not mistaken, I believe the active checks also use the trapper.c code to listen for connections from the agents doing active checks. I wonder if this might be the same or related issue that I was getting with trapper items in thread: http://www.zabbix.com/forum/showthread.php?t=10469

Not sure but sound like similar issues.

[disgruntleddutch]

Problem is that I got these with 1.5.3 as well.

Hard to describe what is going on but basically I change a few checks to active and they immediately show up in the queue under Administration. As time progresses they move right until they hit the 10 minute plus. Queue details reveal that they are checks I just turned to passive. Next check time is correct but nothing is happening.

I know the host gets it because I see it in the logs show up on the host side but no hosts, none, are reporting their active checks.

I got this in 1.5.3 and still getting it in 1.5.4 and now 1.5.5 beta :-(. I should also mention that when this happens, CPU usage goes through the roof for mysql.