My Systems:
PROD unit runs on RPi4B with Z-Wave devices, OH3.4.2 and Tailscale.
TEST unit runs on RPi4B, no devices yet, OH4.1.2, but struggles with Tailscale setup.
Tried a fresh install on TEST using openhabian 1.9c, but no Tailscale from openhabian-config.
TEST system therefor now a clone of the PROD system upgraded to OH 4.1.2 (w/OH3 Tailscale).
Problem:
Both systems now have the same Username due to the cloning. This causes obviously a Tailscale conflict showing a “Duplicate node key” message.
I believe (but not sure) that my systems need different Usernames, but I can’t find how/where to change that except at first time setup.
Anyone a hint?
Ultimately this is not an openHAB problem. But a quick search for “tailscale duplicate node key” in your search engine of choice quickly gives the cause of and solution to your problem.
You need to completely uninstall Tailscale on one of the machine and reinstall so it gets assigned a new IP address.
Thanks a million.
True but not quite true.
If openhabian 1.9c and its openhabian-config item 3B (setup Tailscale) had installed this would not have happened.
So maybe 1.9c can use a quick overhaul?
The two issues I have come across are Tailscale not installing and Frontail time not showing summer savings.
Make it a good one
Bjorn
But it did install Tailscale. You wouldn’t have received the error if it didn’t. The error was caused by cloning one machine to another machine. Because it was cloned, the Tailscale configuration from the first machine also got cloned to the second machine which includes the IP address assigned to that machine. The error occurs when trying to connect to the tailnet because you can’t have two machines with the same IP address on the same network at the same time.;
Cloning a machine to another and running both at the same time is not something openHABian supports nor is it something openHABian claims to support and it’s not a very common procedure. The supported approach is to install from scratch and restore a backup instead of cloning the machine.
If someone wanted to submit a PR to handle this use case I’m sure it would be accepted, but I doubt any current maintainer would implement it.
As for Frontail, that’s just shows what’s in the openHAB logs. It doesn’t do anything with time itself. That is something OH related. In all likelihood your regional settings are not correct in the operating system, or for Java. There are several threads that go through all the places to look and things to fix there. Here’s one: TimeZone DST - Log shows 1 hour behind
Again, my fresh openhabian 1.9c installation worked fine. But, config item 30/3B did not install Tailscale. It just returned a few error messages. My suspicion is a bad/old URL trying to contact Tailscale.
My workaround was to use a cloned Prod (OH3.4.2) SD card and config item 03 to arrive at an OH4.1.2 system for testing, which now works fine also with Tailscale.
So, my thinking simply is that the maintainer should verify the 30/3B finding and fix it at some point if found to be true.
Then please file an issue about the original failed install. Don’t mention the “Duplicate node key” error as that’s not relevant to the original failed install.
Installing tailscale via 3B works fine for latest openHABian. So no.
Also remember that openHABian is ever-changing. Always update on openhabian-config start first.
Just for the record, it’s not a maintainer’s job to fix stuff.
Maintainers just manage this process and are busy by nature, so wasting any of their time makes them unhappy.
It’s the user’s job to analyze issues properly to the extent that’s it’s undoubted a bug and if so then it’s the contributor’s job to fix it.
Filing a Github issue requires the user to believe it’s an openHABian bug at a minimum to reproduce the error and to provide a debug log there that clearly shows where and why openHABian is misbehaving, taken on a system that is in accordance with installation requirements (a cloned box for example is not as aren’t not-latest versions).
