Good evening all!
So I have been trying to get Rancher/RancherOS upgraded all day and I have been running into some strange issues, especially with RancherOS.
I’ve upgraded the OS from v0.5.0 to 0.7.1 and upon rebooting, the rancher/os-docker container will no longer stay running. It appears to be in a constant state of restarting. This is what I can see when I run
sudo system-docker ps:
As you can see, the status is always showing
Restarting (1) N minutes ago
I’ve printed the logs for that container by running
sudo system-docker logs -t 10a35280b036
but all I see is this:
Just for clarification, I am running the debian console.
Does anyone know another way to debug this issue? At this point I am running out of things to try and I am still in the process of learning the ins and outs of Rancher.
Thanks for the help!
Alrighty, well I think I figured out why the os-docker container was in a tailspin.
It seems that the console was not updated along with the rest of the system when I ran the upgrade.
Now that I look back on ps output I attached I can clearly see the version on the debian console is 0.5.0-3
Anyways, I decided to change the console to ubuntu after looking at the logs again and getting suspicious of the “Starting Docker in context: console” line. Once I restarted with the latest console image, the os-docker started up fine and hasn’t stopped yet!
So my question now… Is this expected behavior after an upgrade?
Shouldn’t the console be upgraded to the same version as what you specify with the upgrade command?
Consoles (and system services in general) are upgraded by removing the existing container and creating a new one. All files not in persistent directories will be removed, which causes problems for users who expect installed packages and such to remain after an upgrade.
Previously a lot of code was written as shell scripts and tied to a particular console version. We’ve removed most of this in v0.8.0 so upgrade issues such as these shouldn’t be a problem going into the future.