[Lawrence Griffiths]

Yep reading 1x also I tested at “lux”:15728.4 still 2 bytes running v2.0.0 of box/evb

payload 0e0f03fe 0805f9ef 05ffff 0b0168

[Lawrence Griffiths]

John, I’ve seen with multitech end-nodes (mDots) join as far as AEP is concerned but end-node seems to timeout so has to re-join. What region are you in USA – EU? as in EU the join request on a fixed channel which has 0.1% duty cycle.

One useful thing is to have an MQTT in node connect to localhost port: 1883 and add the following to topic lora/+/joined and then add output to debug node.

where XX:XX is
{ “topic”: “lora/00:80:00:00:00:00:XX:XX/joined”, “payload”: “lora/00:80:00:00:00:00:XX:XX/joined – “, “qos”: 1, “retain”: false, “_topic”: “lora/00:80:00:00:00:00:XX:XX/joined”, “_msgid”: “93217baf.6cde88” }

[Lawrence Griffiths]

@wojtek Alas NO don’t know why the MAJOR omission?
Let’s hope we get an good answer.

What I’ve done for 1.1.2 is copy custom nodes to: /opt/node-red/node/nodes/
And had to edit /var/config/app/install/development/.config.json to include nodes into the pallet. I’m not sure this last step is necessary but it worked for me.

Lawrence

• This reply was modified 8 years, 3 months ago by Lawrence Griffiths.
• This reply was modified 8 years, 3 months ago by Lawrence Griffiths.
• This reply was modified 8 years, 3 months ago by Lawrence Griffiths.

[Lawrence Griffiths]

OK there are two ways to install NR custom nodes.

1. Old way pre npm
Just copy the node_name.js & node_name.html to nodes folder
cd /opt/node-red – then npm install any dependent node.js modules
Re-start NR from admin UI

2. new way with nodes that are available in npm
cd /opt/node-red
npm install node-red-contrib-“”module name”
Re-start NR from admin UI

• This reply was modified 8 years, 3 months ago by Lawrence Griffiths.

[Lawrence Griffiths]

Are you installing nodes manually or via npm?

If manual just drop files to /opt/node-red/nodes
And re-start node red from admin

BTW: Node-red log is /var/volatile/log/app/node-red.log

[Lawrence Griffiths]

Andrew & Jason thanks

[Lawrence Griffiths]

Brian, hum I’m running 50+ nodes same use case as you, to multiple REST end- points and 2 different mqtt brokers along with a lot of Function nodes.

Also I have two custom nodes of my own which are poorly coded. I have 3 mDots sending every min. I do sometimes have to reboot, may be every other week.

>>Might be better just having a custom process in java to connect the MQTT queues to my backend….
You could Bridge the AEP MQTT (mosquitto) broker to yours..

[Lawrence Griffiths]

In a non MT node-red installation you would edit settings.js to remove https
/var/config/app/install/development/settings.js

// From MT support
Looks like the /sbin/app.py script is writing the settings.js file.
There appears no place for a user settings.js file to incorporate functionality you are attempting.

So for now looks like you might have to edit /sbin/app.py which in turn will create the settings.js

NB. this won’t withstand a firmware update

[Lawrence Griffiths]

I think you need to edit setting.js and I can’t remember what folder they are in on the AEP and I have hunch they might be overwritten.

[Lawrence Griffiths]

Brandon thanks

[Lawrence Griffiths]

Brandon that’s more like it thanks.

[Lawrence Griffiths]

Brandon thanks why such a large payload for a check-in? Any plans to compress this?

[Lawrence Griffiths]

Jason
>> When the network server is shutdown all the database is saved to flash…
Just supports my request for some AEP documentation. Without it you just make assumptions to fill the VOID. I will test this and feedback

Why NR re-boot I have a custom NR node that uses long polling and on occasions it needs a NR restart. But if a AEP reboot preserves LoRa end-node session keys then having a separate NR re-start isn’t required.

Lawrence

[Lawrence Griffiths]

Jesse yep know that thanks but with a NR re-start you still have all your LoRa end nodes joined to network. A complete reboot requires the LoRa end-nodes to re-join which is unnecessary in that it uses up battery and duty cycle!

[Lawrence Griffiths]

Forgot From DeviceHQ can we have Remote Re-Start of Node-Red please

[Lawrence Griffiths]

Jeff I fired a list of features to Dan & Derek not sure they got through to you.

// AEP database node
We need pestilence on AEP file store not good enough. A time series one would be Fab but appreciate that would take some work. I see mLinux has SQLite why not AEP?

I’ve been following http://lokijs.org/ it’s pure JS nosql db and small size and have been considering creating a NR node for it.

// custom NR node preservation on Firmware update

// LoRaOUT node I want to track & guarantee the sending of downlink commands to mDots. It would save me having to develop a queue manager if the node provided 2 output events on recite of an input:
1. output queue_id or msg_id in msg.payload, msg.action=”queued” & msg.timeStamp
2. output queue/msg_id in msg.payload, msg.action=”sent” & msg.timeStamp
At the moment from AEP you have no way of knowing if msg been sent all you can do is monitor mqtt with a filter lora/+/down

// Protocol nodes Modbus Serial/TCP (some MT equipment have modbus support) there is npm module https://www.npmjs.com/package/jsmodbus

Bug fix please for Ticket #5065120

Lawrence

• This reply was modified 8 years, 6 months ago by Lawrence Griffiths.
• This reply was modified 8 years, 6 months ago by Lawrence Griffiths.

[Lawrence Griffiths]

Jason thanks for digging there needs to be some Dev. Docs for AEP.
But modifying the python script won’t withstand a firmware upgrade!

There is clearly a reason why the AEP team think uses shouldn’t go near settings.js but it’s a key part of Node-RED so I’m going to raise a ticket.

One thing I have’t tried do custom nodes survive a firmware upgrade?

[Lawrence Griffiths]

Jason edited /var/config/app/install/development/settings.js and a Node-RED restart overwrites settings.js as does a Save & Restart from admin UI.

The preserved/retained for upgrade version of settings.js must be stored some place else!

Thanks
Lawrence

[Lawrence Griffiths]

Jason that is a snippet from my settings.js file in opt/node-red/settings.js

[Lawrence Griffiths]

Robert yes you need to set
– Network Interfaces: eth0 to WAN
– WAN set eth0 to top of list

I use Ethernet for dev in the office and then cellular for client demos

Lawrence

[Lawrence Griffiths]

Andrew I’ve been pondering whether a Proxy would be the best approach. But feel this a big bit of work. Idea the AEP has a node that emulates a LoRaWAN end-node and proxies to the TTN network an behalf on the real AEP end-nodes.

This would allow AEP work with any LoRaWAN provider. And would allow a start of starts using multiple AEP to build a bigger networks using a LoRa backbone. As the only way to AEP to AEP is wired/cellular.

Lawrence

[Lawrence Griffiths]

Brandon I thought Class C was just continuous listening. I don’t see anything in LoRaWAN 1,0 spec about peer to peer comms? I thought that was just stnd loRa radio function.

Lawrence

[Lawrence Griffiths]

Brandon both antenna’s off

Andrew are you saying you have to return the defective mCard before a replacement is shipped? If so I will have a word with UK office as to get us some replacements now.

[Lawrence Griffiths]

Brandon run the same test my results

at+ping
-113,5.0
OK

at+rssi
-123, -123, -102, -115
OK

What should we be looking for here?
As it might help in the future when trying to diag issues.

BTW I can never get an mDot to join without an Antenna even at a few inches!

Lawrence

[Lawrence Griffiths]

Brandon thanks

[Author]

Posts