I am sure that you have backups but I have found that it is an interesting experience to do a complete reinstall of all the software on a Pi. Having a spare Pi to experiment with seems a good time to have a go. Recording every step may also help in the future.
Hi Guys!
Right...re-visiting this on a different angle of attack..
Scenario:
Existing Pi running quite happily, blissfully unaware that it is soon to become obsolete...
"New Pi" now set up with all the latest versions of the base packages I need, blank slated, no flows, databases, etc, just the functional framework ready to begin doing something, blank software packages including node red are installed....
Now
What files do I need to copy across to get flows, configurations, context data, the whole shooting match etc including passwords to log in to node red as a user or admin etc, to get the new pi now on a later OS, up and spinning and ready to take over the load?
Regds
Ed
Basically everything in .node-red with the exception of the node_modules directory or directories.
However there could be static content in some other directory (check settings.js for httpStatic)
Also check that the new Pi has access to network services such as databases, mqtt brokers.
Once everything is on the new Pi, you can rebuild node_modules by:
cd .node-red
npm install
My understanding is that this will install all required none-core nodes. I have never verified this!
You can find various discussions in the forum about backup and restore. For example https://discourse.nodered.org/t/backup-and-restore-node-red/86856/14
Also, if you are using UIBUILDER, check its sub-folder, there is another node_modules folder there that you wouldn't need. To reinstate
cd ~/.node-red/uibuilder # or wherever you've changed it to
npm install
If you have done a local install of Node-RED, that is another place to check and do similar.
I keep a BASH script around that does everything for me.
Thanks Guys, that was a lot less painful than my previous attempt.. and it worked!
I am not using uibuilder as yet...Still on my todo list...
Now... To try and get historical data from influx v1.8 to influx.current... Any hints?
TIA
Ed
Which current?
I never bothered to upgrade because of the changes. The current free version is v3 Core - BUT, that has significant restrictions. From the docs:
Core is a free, open source, single-node engine optimized for real-time/recent data
I'm running 1.13.1
You can run v3 Enterprise self-hosted for non-commercial use. However, there is no upgrade path, you have to export your data and import it to v3E. You are restricted to using 2 CPU cores. Your old queries should, however, mostly work after migration.
PS: If I were starting today, I would not use InfluxDB. I would probably use PostgreSQL. Possibly with the TimescaleDB extension if needed.
I've settled on V2...
Seems adequate for what I need... I like the web interface for basic graphing too...
Cya
E
I advise against that. It is no longer being developed and the migration can be painful. V2 was regarded by many as a mistake. V3 reverts some of the changes made in V2, but is arguably not yet stable enough for production use, so I agree with @TotallyInformation, stick to v1.13 for the moment.
