agent via proxy (agent.ping.nodata(300)}=1) problem in 1.8

[elvar]

Hello,

I'm having a problem where a Linux(Gentoo) agent(1.8) keeps getting reported offline using agent.ping.nodata(300)}=1. I know the host is not actually down as I'm on the host doing a constant manual ping to the Zabbix server box and it hasn't skipped a beat. Anyone else experiencing issues with this in 1.8 for your agents being monitored by a proxy?

In my scenario the proxy is on the same Gentoo linux box as the agent that's being reported unavailable.

Any help would be greatly appreciated.

[elvar]

When I switch this particular Gentoo box to be monitored via agent directly by the Zabbix server I do not experience these same problems. I only get the false nodata reports when I monitor it by proxy.

[untergeek]

I had a similar problem in that items monitored by Trapper (i.e. sent by zabbix_sender) were backing up or not making it through. My time stamps for latest data were off by minutes, which was enough to give false positives from the nodata(300).

I wonder it they're related...

[fascinatedcow]

same issue here

I have this issue also. Randomly I get a torrent (5-10 hosts at a time) of unreachable alerts from the agent.ping.nodata trigger. They clear within about 30 seconds and I can see that from the history no data is actually missing; i.e. there is a value of 1 for the agent.ping item every minute. This would lead me to believe that for some reason the data is backing up on the proxy. Anyone know a way to make it deliver data faster to the node?

[richlv]

add item zabbix[proxy,"proxy name",lastaccess] and monitor it with fuzzytime trigger expression. if that fires, you'll know that it's the proxy which is not sending in the data for some reason.

you can't look at timestamps afterwards, as proxy is sending in data with timestamps according to whenever it got the data, not when it was sent to the server

[iced-tux]

Same issue here w/o proxy

Hello all,
we are having the same non determinable problem.
We have looked in the available logs but found nothing.
These interrupts seem to appear at random.

We also looked at things like open_file restrictions and network buffer sizes.

Any hint would be greatly appreciated.

Greetings
iced-tux

[elvar]

Hello all,
we are having the same non determinable problem.
We have looked in the available logs but found nothing.
These interrupts seem to appear at random.

We also looked at things like open_file restrictions and network buffer sizes.

Any hint would be greatly appreciated.

Greetings
iced-tux

I've solved all my problems with this by having a properly tuned MySQL DB as well as having the proper amount of StartTrappers and StartPollers. Search the forums for tuning MySQL and you will find some good threads.

[fascinatedcow]

This is also now solved for me, but it wasn't the configuration of mysql that was causing the problem, it was the compiler that was used to build it! See http://www.zabbix.com/forum/showthre...9371#post69371 for more details. Now everything is completely fine.