Agent exited with code 2: CRIT -- docker -- raw version -- bug report

Troubleshooting
, ,
1

Hi,

i am running checkMK as docker: check-mk-raw:2.0.0-latest which is

2.0.0p11 (CRE) right now.

I am using “Individual program call instead of agent access”

Command line to execute is:

ssh -i /omd/sites/cmk/.ssh/id_rsa -T root@$HOSTADDRESS$

While on cli shell this command will work:

ssh -i /omd/sites/cmk/.ssh/id_rsa -T root@10.1.2.3 | head

<<<check_mk>>>
Version: 2.1.0i1
AgentOS: linux
Hostname: storage1.local
AgentDirectory: /etc/check_mk
DataDirectory: /var/lib/check_mk_agent
SpoolDirectory: /var/lib/check_mk_agent/spool
PluginsDirectory: /usr/lib/check_mk_agent/plugins
LocalDirectory: /usr/lib/check_mk_agent/local
<<<labels:sep(0)>>>

inside the UI, it will fail with:

Agent exited with code 2:

Another server, with same OS ( archlinux ) will work fine in cli and UI ( but

<<<check_mk>>>
Version: 2.0.0p11
AgentOS: linux
Hostname: 
AgentDirectory: /etc/check_mk
DataDirectory: /var/lib/check_mk_agent
SpoolDirectory: /var/lib/check_mk_agent/spool
PluginsDirectory: /usr/lib/check_mk_agent/plugins
LocalDirectory: /usr/lib/check_mk_agent/local
<<<df>>>

( while for some reason the “Hostname” will be shown empty here )

I tried to get more information about whats it’s issue by executing:

sudo -u cmk /opt/omd/versions/2.0.0p11.cre/bin/python3 /opt/omd/sites/cmk/bin/cmk --debug -nv storage1.local

from inside of the container. But just got:

Check_MK can be used only as site user.

During running the tests the docker container will run:

cmk       130799 83.6  0.5  99240 85912 ?        S    07:53   0:02 python3 /omd/sites/cmk/bin/check_mk --automation diag-host storage1.local agent 10.1.2.3  6556 1 5 5
cmk       130817  0.0  0.0   2392   700 ?        S    07:53   0:00 /bin/sh -c ssh -p 2211 -i /omd/sites/cmk/.ssh/id_rsa  -T root@10.1.2.3
cmk       130820  3.0  0.0  13072  5876 ?        S    07:53   0:00 ssh -p 2211 -i /omd/sites/cmk/.ssh/id_rsa  -T root@10.1.2.3

running the ssh command manually ( with " | head") from inside of the docker container will work and output:

<<<check_mk>>>
Version: 2.1.0i1
AgentOS: linux
Hostname: storage1.local
AgentDirectory: /etc/check_mk
DataDirectory: /var/lib/check_mk_agent
SpoolDirectory: /var/lib/check_mk_agent/spool
PluginsDirectory: /usr/lib/check_mk_agent/plugins
LocalDirectory: /usr/lib/check_mk_agent/local
<<<labels:sep(0)>>>

So as we see:

Version: 2.1.0i1 ( not working )

vs.

Version: 2.0.0p11 ( working )

So in case you run into the same issue, now you know that at least version 2.1.0i1 that i got yesterday from the official github repository:

https://raw.githubusercontent.com/tribe29/checkmk/master/agents/check_mk_agent.linux

is buggy. Choose another version and it will work :slight_smile:

Greetings
Oliver

2

As this is “only” the dev version and some command inside this causes the not 0 exit code. I also saw this as i was playing around with the agent. If you checked this, because you needed the cgroupv2 checks for containers then you can also port this to the older agent. I needed to do this for my Proxmox 7 system. This also uses cgroupv2 for all lxc containers.

3

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.