Cutting my losses - node removal. TASMOTA

[Trying_to_learn]

I get that this node gives you some nice/neat features.

I have it installed but not directly being used. It is in a flow, but disabled.

I didn't think there were/are any implications of this.

But if I then delete the node from the flow and deploy things stop working.

All my comms to the device/s are directly through a MQTT node.

So how is this node not allowing things to talk to the devices if it isn't in the flow?

[knolleary]

Assuming this is a continuation of yesterday's thread, you disabled the flow node, but did you disable/remove the corresponding config node as well? The one that was logging errors?

[Trying_to_learn]

(Sorry...)

I determined that the tasmota node is the one causing the problem.

I just did it again and it is behaving itself.

Sincere apology.
I hope this helps reduce the problem/s - yes from the thread the other day.

update:

Ok, I stand corrected.

I tested things as I said and posted the reply.

Then a light I had turned off by MQTT turned itself back on.

I turned it off again and it (just now) turned itself back on.
I'll put back the two nodes I deleted and post an update.

[Trying_to_learn]

Put the two nodes back, deployed and the bulb is still turning back on after a few minutes.

Oh, but! (sorry) I monitor LWT and other messages.
If it gets a LWT it disables nodes (GUI).

When the bulb turns on, all those nodes are disabled for a moment then come back as enabled.
But they don't seem to be able to control their devices.
Tried clicking on the buttons and nothing is happening.
(Sorry, posting as things happen)
I also just got what I suspect was flooding of the MQTT broker.

The machine which is the MQTT broker is maxed out on load.
The bulb keeps turning on at random times. (nominal...)

I'm trying to look at the MQTT logs, but as the machine is maxed out, it is slow going from the CLI.

[Trying_to_learn]

Another observation:

Things are behaving themselves.... pretty well.
This machine is disconnected from the network. Things... well.... status quo.

I reconnect this machine and things go crazy.

This machine may be contributing to the problem.
Not 100% sure. But I did notice this.

Also I have a couple of scripts that monitor the CPU load of the MQTT broker.
That peaked and is staying high.
Soon there after another machine is monitoring if it can talk to the MQTT broker.
It signals no MQTT comms.

Something weird is happening.

While this is happening it is difficult to look at the MQTT log, as the machine is maxed out.
(RasPi 2A)

[Trying_to_learn]

Looking at: /var.log.mosquitto/mosquitto.log I get this:
(the end bit)

1637188840: New connection from 192.168.0.93 on port 1883.
1637188810: Client mqttjs_0ca6ebdf disconnected.
1637188810: New client connected from 192.168.0.93 as mqttjs_0ca6ebdf (c1, k60, u'').
1637188810: New connection from 192.168.0.93 on port 1883.
1637188780: Client mqttjs_1649aea3 disconnected.
1637188780: New client connected from 192.168.0.93 as mqttjs_1649aea3 (c1, k60, u'').
1637188780: New connection from 192.168.0.93 on port 1883.
1637188750: Client mqttjs_acd0b3db disconnected.
1637188750: New client connected from 192.168.0.93 as mqttjs_acd0b3db (c1, k60, u'').
1637188750: New connection from 192.168.0.93 on port 1883.
1637188720: Client mqttjs_521c3558 disconnected.
1637188720: New client connected from 192.168.0.93 as mqttjs_521c3558 (c1, k60, u'').
1637188720: New connection from 192.168.0.93 on port 1883.
1637188714: New client connected from 192.168.0.26 as BULB-2 (c1, k30, u'DVES_USER').
1637188714: New connection from 192.168.0.26 on port 1883.
1637188714: Socket error on client mqttjs_984fb2b0, disconnecting.
1637188714: New client connected from 192.168.0.93 as mqttjs_984fb2b0 (c1, k60, u'').
1637188714: Socket error on client BULB-2, disconnecting.
1637188714: Client BULB-2 has exceeded timeout, disconnecting.
1637188690: New connection from 192.168.0.93 on port 1883.
1637188660: Client mqttjs_85307ba8 disconnected.
1637188660: New client connected from 192.168.0.93 as mqttjs_85307ba8 (c1, k60, u'').
1637188660: New connection from 192.168.0.93 on port 1883.
1637188630: Client mqttjs_57e0958b disconnected.
1637188630: New client connected from 192.168.0.93 as mqttjs_57e0958b (c1, k60, u'').
1637188630: New connection from 192.168.0.93 on port 1883.
1637188603: Client mqttjs_d64d6caf disconnected.
1637188603: New client connected from 192.168.0.93 as mqttjs_d64d6caf (c1, k60, u'').
1637188600: New connection from 192.168.0.93 on port 1883.
1637188570: Client mqttjs_224d7718 disconnected.
1637188569: New client connected from 192.168.0.93 as mqttjs_224d7718 (c1, k60, u'').
1637188569: New connection from 192.168.0.93 on port 1883.
1637188540: Client mqttjs_ae6b2234 disconnected.
1637188540: New client connected from 192.168.0.93 as mqttjs_ae6b2234 (c1, k60, u'').
:

Which is fine.

Does that help anyone who can explain or help me determine what is going on?

Oh, small overview:
That is the main MQTT broker.

192.168.0.93 is a RasPiZ(W) which monitors the broker.
It polls (or should poll) it every 30-ish seconds.
Not flooding.