Hi, I’m new to Kubernetes and Rancher. When I want to set up a local development/learning environment, I failed to make Rancher run in single-node mode using docker. After some painful testing, I figured out it might be related to openSUSE Tumbleweed.
I create 3 identical VMs, each having 4 cores, 8GB of RAM, 120GB of storage, 2 NICs (one for VMnet8, DHCP, for accessing the internet; one for VMNet2, static configured, for internal communication).
I installed Ubuntu 20.04LTS, openSUSE Tumbleweed, openSUSE 15.3, performed upgrade (apt update && apt full-upgrade -y, zypper dup -y, zypper up -y), and installed docker. The docker on ubuntu is installed by following official documents, the docker on openSUSE is installed by following the wiki (Docker - openSUSE Wiki). All dockers can run the hello-world with the non-root account.
Then I deployed the Rancher using command: docker run -d --restart=unless-stopped -p 80:80 -p 443:443 --privileged rancher/rancher:latest
The result is: Ubuntu 20.03LTS and openSUSE Leap 15.3 have no issues, but the container on Tumbleweed keeps exiting and restarting. It starts, and waiting for server on 127.0.0.1:6443, and then failed with a FATAL, sometimes it’s k3s exit with code 1, sometimes it’s just a connection refused. And the web UI cannot be accessed. Here is the full log:
(the log containing URLs to 127.0.0.1 and considered as a link, and new user can only have 2 links in the post, so I can’t post the log here)
During the waiting process, I can see there is a k3s process using about 100%~200% CPU, but then exits and the container failed.
To give more details about the system configurations, here is the kernel version and docker version:
Ubuntu 20.04: Kernel 5.13.0-37-generic, Docker version 20.10.14, build a224086
openSUSE Tumbleweed: Kernel 5.16.15-1-default, Docker version 20.10.12-ce, build 459d0dfbbb51
openSUSE Leap 15.3: Kernel 5.3.18-150300.59.54-default, Docker version 20.10.12-ce, build 459d0dfbbb51