Did you notice if it crashed after a certain event? Perhaps after the disarm event? Ive read something that rapid messages could cause the current pubsubclient MQTT library to crash, hence potentially crashing after the disarm message…?
It seems to crash without any events - it degrades. First of all the motion updates stop, next you can Arm but it does not respond to disarm on either the web UI or Home assistant (which is sending MQTT disarm messages).
I just ran another test. This time it worked for about 30 mins, reporting sensors etc. At that point it stops responding to MQTT. Web interface: Zone names and Alarm Status still working, Arm and Disarm still working at present. BTW it is reporting Uptime: 00:00:07.06
When it crashes the Powerlink connection fails but can be re-established. However no zones show in either Alarm Status or Get Zone Names.
Hmm, so the Wemos was up for 30 minutes but only showing uptime of 7min? I wonder if there is a memory leak in one of the recent library releases or something…
Hi, just ran another test - same basic pattern but the clock is still ticking on after the MQTT has stopped.
I am wondering if the page had not refreshed as I clicked back to it.
And does your mqtt server give any indication or logging about what happened? Does the TCP connection die or does the connection stay and messages stop? Does it give a reason, such as timeout?
Seems like I may need to setup an mqtt server again so I can test myself… It seems to be entirely based on library versions as there is no code change.
Been watching the MQTT on a monitor as well as the Server log which reports - Client ESP8266Client has exceeded timeout, disconnecting. The monitor just goes quiet for all the topics.
I can access the web menus (sometimes they take 40-60 seconds to respond). I can’t tell if the TCP connection to the MQTT server is maintained. The Free heap varies from around 24000 down to 23000 and then rises again.
I have tried various triggers (sensors, arm/disarm) but there does not seem to be a pattern.
Hi I have been looking at ways to capture the battery status as a sensor in Home Assistant. Would it be possible to separate the problem report into individual mqtt messages with topic ‘alarm/zone#’ and payload of ‘low battery’?
Hi, I left it running for the last 12 hrs. It is just the mqtt part that fails. I can still access the web page.
Can you try the latest firmware on Github please? Discovered some funky timing in the MQTT library that could cause disconnections in some circumstances so gave it a bit more processing time to maintain a connection… Fingers crossed!
The status messages over MQTT was a nice added feature!
Trying the new firmware as of now. It works , all settings seem to be correct, but funny enough the password I set for the web admin page no longer works!?
Will do this weekend
I expect turning it off and on again (or changing the password again) would fix the admin page, though dont know why that would have happened.
Keep an eye on the MQTT status please, to confirm it doesnt die after a few hours, that was the problem from before.
Yes, I’ll try that. I’m using it on a daily basis so if MQTT halts for some reason I will notice. Thank you.
Looking good - running fine for 12 hrs
Less disconnections aswell as not crashing? Hopefully so!
None so far - 15 hrs
@Peter2 That must indeed be a very very old model as it has some quite weird features that certainly aren’t common on Powermax plus models. I can’t guarantee it will work but will say you have a few options:
I have done some testing of a new RS232 adapter that might make board connection very simple, one small adapter straight onto the pins that you can see under the power cable, and connecting straight onto the Wemos. This would be super simple and I reckon has a 50:50 chance of working.
You can also buy a new panel and not replace the zone sensors. A new panel has a good chance of being compatible with old sensors and might cost less than 50 GBP or 70 USD, especially if you have a look on eBay where there are plenty floating around. It would require some tinkering to get all set up but does offer you another option to getting it running without totally changing your system (the above kit would almost certainly be compatible with the new panel unless it was also super super old). That’s my impression anyway!
Having the login issue since latest upgrade I took the chance to reflash to a new NodeMCU unit with a smaller footprint that fits better in my panel. I’ve had in running for 24+ hours, no problems… But. Today I noticed it had disconnected from MQTT broker and would not reconnect until soft rebooted.
I will keep an eye on it to see if that was just a hickup.
Curious! Bloody MQTT!
You can always do a full reset of the device with the 5 quick power cycles as per github instructions, that should have fixed your first device. For info now, but potentially useful!