(Locally) how to apply an updated subflow over multiple machines?

General
1

Yeah, I probably have asked before.

I have just had a lot of fun retrospectively rolling out an update of a subflow I wrote over my machines.

Is there an easier way?

I had to copy/export the new code, go to each machine, edit the subflow and paste in the new code and hook it up.

Not the worst but a bit tedious.

Keeping things LOCAL...

Is there an easier way?
(Oh and keeping it simple.)
If it involves too much stuff, it may not be easy for me.

2

(I nearly asked again, but as I found the original post)

Sorry, but it is sort of becoming a bit more frequent in happening and I would prefer to keep things local.

3

I haven't done this, but I guess you could package it as a module which can then be installed into each machine and updated via npm as usual. Put in a github repository rather than publishing to npm of course since, presumably, it will only be of use to you. Packaging a Subflow as a module : Node-RED

1 Like
4

You could do via. the following programmatically

  1. Get global flows from target.
  2. Get global flows from source.
  3. Isolate the subflow from the source global - using subflow name.
  4. Replace the subflow in target global - using sub flow name . Replace it by subflow from #3.
  5. Send back global to target

1,2,5 you use NR APIs.
3,4 function nodes.

5

Thanks Colin.

But reading that link I fear you are overestimating my skills.

I'm not sure what is going on with me, but I do not feel confident enough to want to go down that path. Yeah, ok, so that means what I want to do can't be done.

Such is life.
But thanks.

6

Hi Sandeep.

Again, I think you are giving me too much credit for intelligence.

What you wrote is (basically) a foreign language to me.
All I am seeing is words. Their meaning is unknown.

7

Packaging my first nodes seemed intimidating at first (still new, I just learn as I go), but it really wasn't bad, and definitely worth learning.

8

Thanks. When I have the energy (just sustained other high priority things) would you mind helping me get my head around it?

9

Yeah, absolutely. Just send a message when you got time.

10

Andrew, have you considered moving away from subflows?

As you have multiple node-reds, you could elect one (the important one) to be a "master" (with ONE COPY of your important nodes & functions) on a regular flow TAB.

Then using HTTP or some other transport, call the remote code.

Example...

NODE-RED1 - TAB "shared functions" (Master Flows)

HTTP IN "/calculate-colour" → <some nodes> → HTTP Response

NODE-RED2

<some nodes> → HTTP Request to "/calculate-colour" → <some nodes, use result>

NODE-RED3

<some nodes> → HTTP Request to "/calculate-colour" → <some nodes, use result>

NODE-RED4

<some nodes> → HTTP Request to "/calculate-colour" → <some nodes, use result>

This way, all you code is in one place (on NODE-RED1) - no need to update subflows in multiple places.

11

I'm not sure that will always work.

If that machine dies.... I have multiple points of failure because of their dependency on it.

Not sure that is a good strategy. Still thinking.

12

That's true. But if there is a critical part that "must always be 'up"' then it's a viable option. Since if there is a critical element to your setup, the failure of that is more critical than some globally shared functionality.

13

And believe me, I am having the devil's own trying to work out where this line is....

I get that I may never REALLY NEED this stuff, but my curiosity is painful at best.

I am (just now) having a big discussion about push buttons and what they say.
Do they say/display the state or do they show what will happen when they are pressed.

(Sorry. Off topic)

But it is annoying with this mind set.

1 Like
14

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.