Docker: Occasional ping packet loss

Troubleshooting
1

CMK version: 2.2.0p23
OS version: Docker

Error message:

Output of “cmk --debug -vvn hostname”: (If it is a problem with checks or plugins)
dbg.txt (6.6 KB)

Hey!

I migrated from a x64 Ubuntu CCE to Docker CCE. I migrated with omd restore —reuse —kill which worked.

However, I immediately got occassional ping packet losts for all external hosts. Not for internal ones. Thats it. I dont understand whats going on. The LAN connection is stable and no other container has that issue

Initially it happened with Ping6 (dual stack active) but switching to v4 did not helped.

Any idea? I think it should work with docker and I would be more than happy to help finding out whats wrong.

2

I made some more tests:

I thought, this is surely NOT any CMK issue, because the ping4 and ping6 from inside the cmk container are failing occasionally, too!

So I opened up another docker container and I dont see any issue inside it?

I started two parallel pings:

Screenshot 2024-03-10 102501
Screenshot 2024-03-10 1025011852×868 84 KB

Screenshot 2024-03-10 102359
Screenshot 2024-03-10 1023591875×879 86.4 KB

Screenshot 2024-03-10 102258
Screenshot 2024-03-10 1022581884×873 99.4 KB

Screenshot 2024-03-10 102152
Screenshot 2024-03-10 1021521863×871 97.8 KB

The fails are always inside the cmk container! Both containers are configured the same way. Both has the same macvtap interface and both have a static IPv4.

It does not happen with IPv4 in my tests. IPv6 issue?

3

I also set up a completely new cmk container. Same issues. Container is set up as:

docker run
  -d
  --name='checkmkk_empty'
  --net='eth0'
  --ip='192.168.178.29'
  -e TZ="Europe/Berlin"
  -e HOST_OS="Unraid"
  -e HOST_HOSTNAME="NAS"
  -e HOST_CONTAINERNAME="checkmkk_empty"
  -e 'CMK_SITE_ID'='l22'
  -l net.unraid.docker.managed=dockerman
  -v '/mnt/ssd/appdata/checkmk_empty/sites':'/omd/sites':'rw'
  -v '/etc/localtime':'/etc/localtime':'ro'
  -v '/mnt/ssd/appdata/checkmk/backup':'/backup':'rw'
  --mac-address 02:00:00:00:00:29
  --tmpfs /opt/omd/sites/l22/tmp:uid=1000,gid=1000 'checkmk/check-mk-cloud:2.2.0p23'

It seems to be an v6 issue only currently.

Does anyone observed similar things?

4

It feels more and more like a general IPv6 docker issue. Docker + IPv6 are not the best friends as it seems. Found several issues.

How it CMK image built? which base image does it use?