Best practice needed for upgrading rancher 1.0.2 to 1.4.1

What is the best practice for upgrading from 1.0.2 to 1.4.1
Should we do all the steps 1.0.2 -> 1.1.x -> 1.2.x -> 1.3.x -> 1.4.x or straigth to 1.4.1?

I need the same details.
Can anyone suggest?

I just upgraded from v1.1.4 to v1.4.1 without intermediate steps. I made a point of upgrading all hosts (including the one running Rancher ‘server’) to RancherOS v0.8.1 and thus Docker v1.12.6 first.

To make things simpler I stopped all (bar one) of the stacks and also removed a few that I wanted to specifically wanted to check I could launch again afterward.

It went pretty well. I did hit an issue where it didn’t look like the new Infrastructure stacks had launched and I didn’t see the Upgrading banner across the top of the GUI. This was either browser caching related or because I’d previously disabled the community and official catalogues in Admin > Settings. Sorry I don’t know which as I enabled the catalogues and then immediately restarted my browser for the sake of speed.

After that things seemed to hang. It turned out stopping the stacks was a bad idea as the upgrade process couldn’t upgrade the Rancher load balancers it wanted to because I’d stopped them. I started them and everything was done a couple of minutes after that. I’d highly recommend just deleting all user stacks if you can get away with that.

I did see an issue with the Scheduler which doesn’t seem to like running on a VM, I just forced it to move. I’m still having this issue in an environment that is all VMs - haven’t troubleshooted it yet.

Hope this helps.

Thanks for your inputs.
I also have another Rancher running on 1.1.4 and I tried the upgrade v.1.4.1. Upgrade seems to fine and all containers also started. But none of the container was working due to connection timed out error. connection between container was not working and i suspect docker compose v2 and reverted back to 1.1.4 at last.

Any thoughts?

Did the upgrade complete? Were all the Infrastructure stacks OK?

Did you launch the problematic container fresh? I did have one communication issue between a Sonar and SQL container that were running through the upgrade. A restart of both did the trick.

Sorry I can’t be of more help.