I have a large (0.5MB) Node Red system that has been reliable to date. I have also been backing up the system regularly (about 4 times a day).
But one hour ago, Node Red displayed a dialogue screen to request updating nodes and that there would be no conflictions. I approved the update. After a few minutes of nothing, I realised that I had lost everything.
I did not bother looking into the Pi folders to see what had happened. I simply restored the last backup and accept that I have lost 2 hrs of work.
So my warning is that everyone should backup their flows before allowing Node Red to import updates. I'll look into seeing if I can turn off auto updates...
Perhaps I am missing some bit of knowledge in this area so would welcome any comments. But the purpose of this notes is to warn people to not become complacent with Node Red's reliability and back up often!
Did you or someone else have the Editor for the same instance of Node-RED open by any chance? That sounds like the conflicting update message that you get when a different tab or browser has had the same Editor instance open and made a change. If that happens, Node-RED tries to merge the changes.
Usually this means you have been editing the flow in two browser tabs. The safest thing to do when this happens is to close that tab down and find the original. You will only lose any un deployed changes in the window you close.
Colin, I must take most of the blame. I accepted the "Merge" operation in one dialog box, then selected "Deploy" to update my active script.
So in hindsight there would have been two tasks competing to update the same flows.
This is why I accepted defeat and referred back to a backup.
I suggest that if there is to be a "Merge" of updated nodes, that the "Deploy" button is disabled, or a message to highlight the problem (if this logic is correct).