Mass communication loss overnight?

[Jemster]

I came down this morning to a communication failure on the EvoHome controller. When I cleared the message, most (not all) of my zones, including hot water, are now reading '--' for the temperature.

Took a look at the fault log and lots of communication loss at about 2:30 last night. A few have returned, most have not.

I had breakfast, took another look and 4 more actuators had restored communication so, having to go to work, I hoped the rest would return. Looking at the app, I can see they haven't - it also seems to have overrides indicated on the zones that are back up and running so I've reset these remotely (probably!) but see that one of these hasn't 'stuck' and has gone back up to 20.

I also note it's been calling for heat overnight. So when communication was lost, it seemed to think heat was required even though all the valves should've been shut down for the night

Nothing weird happened during the night to the best of my knowledge. Normal day, no electricals on to speak of overnight. Anyone seen this or got any thoughts on getting it back and running with minimal disruption? Don't want to have to sit and rebind everything as nothing has changed, but not sure how to get it to pick everything up again?

Guess I'll be working at home this afternoon to sort this mess out.

[Jemster]

No explanations, nothing.

Pulled the battery on the controller and put it back. Now, about 90 minutes later, all appears to have returned to normal. Although in between it did give me a variety of randomness, telling me zones had overrides (when they hadn't) that I had to manually cancel. I've a Loop Energy Monitor and I read somewhere that these also occupy the same frequency range for communication. I've unplugged it for now and will see what happens but it's been working alongside the EvoHome for the past month without any problems.

Not very impressed at having to kick this thing up the arse with a physical battery removal to be honest. What if that happened while I was unable to get to the house (e.g. on holiday). Although I had remote communication with the controller through the app there was nothing I could do to resolve the issue - a reboot option would be nice. At least with the Loop Energy Monitor connected I had the pleasure of watching my bank account empty

[dty]

I have Loop and Evohome (and ZWave) and they all operate in the same frequency range. Of these, the most likely to get disrupted is Evohome as it's not a very robust radio protocol (despite Honeywell's advertising) - most messages are just sent "fire and forget" with no acknowledgement. Having said that, the radios in these devices should be listening for other transmissions before they attempt to transmit, but there's still the chance that two things will try and begin transmission at the exact same time. The chances of a collision are therefore pretty slim, and I've certainly not observed any significant issues with having all these things chatting away at once.

[Jemster]

Good to know, I honestly have absolutely no idea what happened yesterday. The system has behaved itself since I restarted it in the afternoon. I haven't plugged Loop back in yet. Guess I'll do that tonight and hopefully everything will remain stable.

What gets me is that if it were a collision happening, it would happen once and the next time around everything would be ok - chances of repeat collision are surely minimal. So if I eliminate the Loop from my enquiries that leaves me with either something blocking communications or that EvoHome gave up all on its own.

If something physical was blocking the communications, then simply resetting the unit would not have caused the blocking to stop, so I think I can rule that out.

Leaves me with EvoHome sh*t itself. But judging from the lack of replies to this thread, this is not a common occurrence either. Well, not to the extent that I saw - I have 9 zones + HW. I lost 6 of them. 4 upstairs, 1 split up and down (2 bathrooms), the other being the room with the EvoHome controller itself. The Hot Water came and went a couple of times through the morning.

[dty]

I would agree with your assessment. I have 24 HR92, half a dozen assorted wall stats, etc., and I'm not aware of any significant problem ever. Sure, the odd thing might have been missed, but most of the system works on just repeating itself every 5-10 minutes anyway (this might be Honeywell's definition of "robust"!) so I wouldn't have noticed.

[Jemster]

UPDATE:

Everything has been working faultlessly since ‘The Incident’. I’ve been checking fault logs religiously and in the last 6 weeks I’ve had one momentary lost hot water sensor but nothing else.

This morning I arrived in at 2:30am to find the heating on. I looked at the controller and it was reporting a failure on a bedroom actuator. I check the fault log and it gave me...
Bedroom 1: lost communication with sensor, lost communication with actuator (1 hr92 zone)
Living Room: lost communication with sensor (3 hr92s and 1 dt92)
The main display was showing overrides for 5 zones! (Nobody had touched anything on the system) and loss of temperature reading for 2 of them, the remaining 4 zones were all normal.

I cleared all the overrides then went into the Living Room and saw that the DT92 was ‘off’ (no display). I popped a battery and put it back in and it re-synced.

By this stage it was 3am and I’d had enough. Went to bed. This morning, I see one of the cleared overrides came back on 20 mins after I cleared it and had been causing demand all night. (Thanks Honeywell, let me know where you’d like me to send your gas bill)

I’ve looked into it further with Domoticz and it would appear that the overrides happened, for the most part, at about 11:45 on a few zones. The DT92 disappeared offline at about 1:30am. I have noticed in both Incidents the DT92 has gone off... but given the times, I am not sure if this is a cause or an effect.

What is really puzzling me is where these overrides are coming from. They have the clock symbol in the UI so aren’t coming from the phone app, they obviously aren’t coming from multiple hr92s simultaneously so how are they getting set???

Twice in 6 weeks is not good after an initial 5 months of reliable operation and I’m unsure where to turn. I don’t know which bit of kit is faulty. There’s no obstructions, no weirdly placed objects, no changes in the house, just total random failure with total inability to manage the failure.

Got to be either the controller, or that DT92 that goes off on its own... maybe...

[Jemster]

I've exported all my data from Domoticz and here's a timeline of what happened...
At about 11pm all the radiators go back down to 10 degrees on their schedule and at this stage everything was running normally.

23:20 Bedroom 3 - Set point unexpectedly goes up to 20 degrees
23:25 TV Room - Set point unexpectedly jumps to 20.5 degrees
23:45 Living Room - Set point unexpectedly goes up to 20.5 degrees
23:50 Computer Room - Set point mysteriously goes up to 20 degrees
00:30 TV Room - Set point drops to 20 degrees (!?!??!)
00:50 Living Room - Set point drops to 20 degrees (!?! like... why???? I can kind of understand a random override but this indicates more than 1 override happened)
01:25 Bedroom 2 - Stopped reporting temperature (This is the one I got the fault message for)
01:30 TV Room - Stopped reporting temperature
01:40 Bedroom 1 - Set point goes up to 20 degrees

02:45 I arrived in and cleared faults and overrides...

02:45 Computer Room - Set point goes back to 10 degrees when I cleared override
Bedroom 1 - Set point goes back to 10 degrees when I cleared override
Living Room - Set point goes back to 10 degrees when I cleared override
Bedroom 3 - Set point goes back to 10 degrees when I cleared override
TV Room - Set point goes back to 10 degrees when I cleared override

02:50 Computer Room - Set point magically goes back to 20 degrees
Bedroom 2 - Restored temperature readings
02:55 Computer Room - Set point goes back to 10 degrees when I cleared override for a 2nd time
03:10 TV Room - Starts reporting temperature again (I did a battery reset of DT92 about 15 minutes earlier)

03:20 Bedroom 3 - Set point went back up to 20 degrees!
03:20 TV Room - Set point went back up to 20 degrees!
03:45 Bedroom 1 - Set point went back up to 20 degrees!


Right now, re-reading this, I am having trouble not over-reacting to what I think of EvoHome and 'fit for purpose'. This is not a simple case of interference. This is total system collapse. Bear in mind that while this was going on, 3 other zones Plus the hot water were functioning normally.

Please... Anybody... any ideas or theories??

[paulockenden]

Is your Domoticz setup exposed to the outside world? If so, is it password protected?

A rogue 'script kiddie' attack that probes forms could end up doing something like this.

Have a look in domoticz/var/domoticz.log and see if there's anything suspicious around that time.

(Actually the location of the logfile will depend on what you're running it on).

P.

[Jemster]

Is your Domoticz setup exposed to the outside world? If so, is it password protected?

A rogue 'script kiddie' attack that probes forms could end up doing something like this.

Have a look in domoticz/var/domoticz.log and see if there's anything suspicious around that time.

(Actually the location of the logfile will depend on what you're running it on).

P.

Interesting theory. It's certainly not intended to be exposed to the outside world. I have just tried using my external IP and port 8080 but no response so I'm thinking it is correctly closed off. I've also ran an external port scanning check tool and it's not finding anything open. Does the domoticz server support PnP to set up the port forwarding? I've not seen any evidence of this on my router.... but I don't check it all that often.

I'm on OS X and according to my LaunchAgent, my log is meant to be at /var/log/domoticz.log and of level "normal" but I don't appear to have a log file at that location.... I've scanned my system but can't find any domoticz.log file That would've been very useful to see.

Chicken and egg. If I kill Domoticz I'll not be able to see it happening in such detail

[Jemster]

Actually, I’ve just tried an override from Domoticz and that gives me the mobile phone override symbol on the EvoHome GUI. That’s not what I had, I had the clock symbol override that you get when changing it on the controller itself.