Node freezes after several hours

[idtres]

Hi, Im using a WemosD1, one output set to dmx another set to PixelStrips, and also a 3 strips for the status LED It works amazing, super fast, no delay or flickering, but after several hours the node stops sending artnet data and wstrips data. 1 status LED goes Red, and the other 2 stop responding, but at the same time, I can ping the device and the webpage responds

not sure what the RED led means, I saw the red led and still worked fine for 2 o 3 hours

Im sending data using VVVV, I slowed down the output to 10 frames per seconds per universe, still same problem.

Please let me know what other test I can try to debug this.

I tested sending 1 universe only, (dmx) and it worked just fine for 5 hours.

[querex007]

I guess this is a known issue of the project...

I found a quick work around: When the node freezes (and still responding to ping and web GUI), you can send 2 reboot CURL's, which makes the node operating again...

CURL: -X POST -H "Content-Type: application/json" -d "@rebootAmbilight.txt" http://192.168.x.x/ajax Body of rebootAmbilight.txt: {"reboot":1,"success":1}

As a reboot only takes 2 or 3 seconds, you can schedule this curl on your notebook / computer every 15 minutes.

[idtres]

It seems a good alternative Im running the code compiled with arduino, latest arduino+8266 setup ESP8266 , I just defined #define ONE_PORT and Im testing the DMX out untill now seems working just fine, If i try to define // #define STATUS_LED_MODE_WS2 it crash, so It seems somehow related to the status led I will try to remap only 1 simple LED as status led like con V1 version I will keep this post updated

Also the idea of restarting the node every 15 minutes seems fine with me, I would love to avoid that

[terrag42]

Is there a way to use one ESP8266 to do nothing but periodically issue curl commands to my other ESP units which are running?

[idtres]

I have been using the node with 2 dmx output receiving data from Resolume for almost 4 hours The only changes I made are feeding the watchdog on the main loop ( ESP.wdtFeed(); )and commenting all lines about neopixel for the status leds,,, for some reason ( well its more like a feeling, no tech reason ) there is something wrong with that. The node as DMX is way more stable than for Neopixels. Using 600 pixels on 1 output only, it freezes all the time in less than 15 minutes. Same node as dmx can last 2 hours with no problem. After disabling all the status neopixels declarations, I was able to run the node for 4 hours without any problem.

I wrote a script to log if the ping stops, but this did not work, may times the node did not reply ping,but still works just fine. I was about to write a vvvv patch to evaluate 2 nodes, one as input a second one as output, but I wont have any way to tell witch one fails, so I will be using dmx as artnet sender + a webcam + rgb fixture, to change color every 1 second, and evaluate using the webcam if the color is correct.

@mtongnz did you found any other tip regarding this ?

UPDATE Im using a microtik router, I left the node runing for 10 hours, still running, Just to be sure, I reset it from the web interface, the counter on the microtik did reset, same If I power cycle the node and it reconnects ( it takes only 3 sec ). Im doing the same test, but for 170 leds. Then I will test again with more and update this post

Update2 : Update1 fail. Node froze under DMX in just 1 hour, it recover, but still, fail