Guidance on host OS

I know the documentation “says” I can install using RancherOS, Ubuntu, Centos7 - however, I have an Asus-based tower and I cannot get Rancher to work consistently as a single stack install. Server has 500GB & 2TB & 3TB HDDs and 32GB ram and 6-core CPU. While I have no problem with any of the OS installs (other than RancherOS), I cannot get docker or rancher to install successfully on any distro. I am beginning to think there is a firmware or driver issue with my box.

On Centos7 I got it to install using 7.2 and the upstream latest version of docker, but it didn’t behave well with the NICs and I kept losing external access to port 8080.

On RancherOS - I couldn’t get rancher/server to activate - errors kept spilling to the console even when trying to use docker logs to see what was the problem.

On Ubuntu I get a core dump after the database is loaded and the server echos out the environment variables and ends with a coredump in the OpenJDK.

I’m not a newbie to any *nix distros, but this has me very frustrated. Centos 7.2 got me the furthest, but I spent two days trying to debug the firewall and IPs with no success.

Any suggestions on how to isolate the issues?

Thanks in advance for any suggestions …

Have you tried something like memtest86 to check the RAM? Lots of weird errors on multiple OSes sounds like it could be a hardware problem.

Just docker on its own has issues, before you even run rancher/server?

Thanks for the idea about memtest86 - it flagged a single problem. I also found out my BIOS was ancient, so I upgraded the bios and Ubuntu 15.10 Desktop is stable-ish - I get UI problems, but the terminal is great! I can run all kinds of docker stuff with no problem … until I try to run the rancher/server container. At this point I get the same issues. I see mysql started and a bunch of CATTLE environment variables displayed, but then I get a core dump. How can I see/retrieve the core dump?

–Roger