Check_MK agent crashing on a SQL cluster

Hi,

we’re facing a strange problem in our Check_MK. In one of our SQL nodes the Check_MK agent crashes quite regularly and the only way to bring it up again is restarting the service.

We already tried

Hi @Infonet,

Can ou provide Check MK server version and agent version and SQL version ? The other nodes work ok with the same config ? do that node have extra load compared to the others ?

Hi @tavanez,

thanks for the answer. The Check_MK server and agent version is 1.6.0p6.
As of SQL we are working with a SQL Server 2012, I’ll post a capture with the components and their versions:
image

In this particular host we had problems with all the agent versions we tried to install (1.6.0p6 & 1.2.8p1).
All the other hosts are OK with the same configuration and no, this node does not have more load compared to the others.

Best regards

1 Like

Maybe there is a job or some sort of output that :
a) Never ends;
b) returns something strange.

try to run locally the check-mk-agent.exe test during that period and see if there is anything strange in the output.

Hello,

thanks for the answer. Could you please let us know where we can find said test? We were unable to find anything in the installation path.
Also, once launched, what are we supposed to see/look for?

Kind regards

Go to your agent installation directory on the command line.
What is also important is what plugins and local checks are activated on your system.
Depending on the agent version you are using the folders are different.

I only write it for agent 1.6 as example

C:\Program Files (x86)\checkmk\service\check_mk_agent.exe test

If you use some plugins or local checks you can see inside the folders

C:\ProgramData\checkmk\agent\plugins
C:\ProgramData\checkmk\agent\local\

It is possible like @tavanez already said that there is a script blocking your agent execution.

1 Like

Hi, for the agent version is :
First cluster node OK - [agent] Version: 1.6.0p6, OS: windows, execution time 7.6 sec
Second cluster node CRIT - [agent] Empty output from agent at TCP port 6556 CRIT , Got no information from host, execution time 4.8 sec
[agent] Version: 1.6.0p6,

We also have the cluster log that you said to export if necessary

Davide

That is no information to my question.
How does the agent behave if you run it locally in test mode and what plugins do you use?

2 Likes