I switched two months ago from Homey Pro 19 to Homey Pro Early 23. Immediately after restoring the backup, Homey started showing instability, manifesting as consecutive reboots lasting up to 2 hours. Afterward, it functions perfectly. I have tried resetting and disabling apps alternately, but I cannot pinpoint the root cause. Has anyone experienced similar issues? These reboot cycles can occur once a day or after 3–4 days of normal operation.
1 Like
Please contact Athom Support.
Next to reporting to support, there’s one thing you could do, reinstall the firmware.
- Create a backup
- Create a backup
- Perform “Software download”; this reinstalls de firmware from scratch
Howto: https://support.homey.app/hc/en-us/articles/7263435817116-Connecting-Homey-Pro-Early-2023-to-a-PC-or-Mac-with-USB
Hello, I did it, and reinstalled the firmware and had a factory reset. Unfortunately the problem didn’t stop. They asked me to test without reinstalling my back up. But considering that the problem can occur after a week, this would mean to have my home off for one week.
In such case, share more on your setup and the behavior your are experiencing, particularly :
- what color has the LED ring, when it happens
- which power-supply you use, have you tried also another USB-C cable
- Ethernet or wifi connected or both ?
- install Sysinternals app, add all compatible devices and track for memory, temperature and also under-voltage issues
- in case your smart home is bigger one (>20 apps, >200 devices) = how many apps and devices you have ? (you will find on the forum here some scripts, which can help you with some statistics)
hello @Sharkys, thanks for your support!
-
what color has the LED ring, when it happens: it turns to red for a while , then makes the standard process to reboot. After rebooting it turns to red again after 4 5 minutes, these cycles can continue for two hours, then it stops and everything is back it normality
-
which power-supply you use, have you tried also another USB-C cable: yes I’ve tried everything, with no result.
-
Ethernet or wifi connected or both ? Tried both, now just ethernet (at the really beginning I thought it was the connection)
-
install Sysinternals app, add all compatible devices and track for memory, temperature and also under-voltage issues: I’ve been using the Dashboard app for several days to track all this activity. I have to say that the temperature is always below 58°C, memory usage is around 300-350 MB, there are no undervoltage events, and the CPU load averages below 20% (of course, when it starts rebooting, it’s much higher, but not before the issue occurs). The problem is that nothing unusual seems to happen before the rebooting starts.
The annoying thing is that it’s almost impossible to break the rebooting loop. I’ve tried cutting the power for 5 minutes—this rarely works. Disabling apps generally only increases the time frame between reboots but doesn’t stop the loop entirely. -
in case your smart home is bigger one(>20 apps, >200 devices) = how many apps and devices you have ? (you will find on the forum here some scripts, which can help you with some statistics) I have 229 device (65 virtual diveces) and 26 app, 280 actvive flows (many based on homey script), I’ve also Homey Bridge, but I’ve now turned it off, I really do see any advantage. The main part of the divices are AQARA, SHELLY and KNX, plus 9 alexa’s and about 10 Philips Hue (1 of these is managed without philipls hub, since it is to far away).
If there were an issue with my backup, would it be useful, in your opinion, to uninstall some apps and reinstall them manually without using the backup?
That usually indicate crash of homey-pro container, resulting in restart. Athom support shall see the reasons in the log.
Please elaborate
Dashboard app ? I spoke about sysInternals | Homey - add 3 devices and share screenshots from them. Also share insights of uptime + memory available , ideally during crash. Also please share Homey Developer Tools and hide/obscure any ID (eg. cloud ID).
That’s why I’m curious on the available mem vs uptime insights.
That’s mid sized home, it should not cause any performance issues at all nor crashes.
You have added it satellite, or ? Did you just shut it down or removed correctly ? Anyway, I see no reference to the issue mentioned just to avoid future problems.
Have you tried also to do factory reset via USB with restore of backup after ?
Maybe a long shot, but is your ethernet adapter connected in the right way?
Adapter plugs directly into the power supply. USB cable from adapter to your Homey.
In that case he wouldn’t be able to reach Homey at all. I think it’s not related.
Yeah maybe the memory graphs of all app’s & Homey reveal something going on.
However, I can’t imagine memory issues will (be able) to cause these weird reboot loops at some point in time.
@Andrea_Marchisi
Your Homey doesn’t seem very busy, that’s good
Please specify all app names.
And can you show the disk usage graphs?
Preferably right after a succesful restart and after a few days.
This is very interesting!
It might be a “faulty” or misbehaving app.
Certain apps can have logs enabled, for troubleshoot purposes, but should be disabled because logs can grow huge sometimes.
It’s possible some app fills all partition space.
Not sure how Athom prevents partitions from getting full.
I think it’s odd how Athom appearantly can’t find any clues in a system diagnostic report.
From technical POV, that’s a good idea, to rule out the influences of all additional stuff.
But I can imagine you can’t miss Homey for a week .
However, in my opinion, a smart home should be able to function without network / Homey. Things break. Always.
I will miss certain automations, but I can toss Homey out the window at any time.
hello, yes it connected in the right way.
@Peter_Kawa @Sharkys , I’ll try to answer to all the questions:
It seems that Athom support doesn’t understand the issue. After reinstalling the firmware via USB-C link on the PC and performing another factory reset, the problem persists. Athom’s suggestion is to do another factory reset, avoid using my backup, and see if the problem occurs again. If it does, they say it’s a hardware issue.
here below you have the CPU load and the memeory usage during the last 7 days. Here, note that there is an anomaly: the use of at least 100 MB more memory than usual (during a long period of stability without events). However, in general, other cycles also behave similarly, with no noticeable CPU or memory spikes. The only observable issue is the drop after the restart and the high CPU usage during the subsequent reboots. In the same graph you can see other two events.
here the last 31 days. Eachtime you see a drop in memory and a pick in CPU, I had the reboot cycles.
here below some details of my system:
Here below the apps installed:
Here below memory and archive:
Here below the last 24 h. No reboot event, the drop at 9:30 it was me that I reboot the system. I unistalled IKEA app, reboot and reinstalled it with the three zigbee extenders.
With reference to HOMEY Bridge, I’ve unistalled it from the app and then unplugged it.
I hope to have answered to all your questions-
So far all looks fine, submit the Homey device insights (the one from sysinternals - free memory in MB together with uptime and CPU temperature) when the crash occurs with 6 hours resolution. Unfortunately when you share insights for longer duration, the peaks are ironed out, so we can just observe trends.
Also when the crash will occur again, you can try to install experimental version by enabling it dev. interface - Homey Developer Tools. I guess you have nothing to loose.
And indeed, it start looking like HW issue.
1 Like
Hello, around 4 PM it happened again and it contiuned for about 1 hour. When in the graph you see the straight bottom line means that I kept it unplugged for a while.
Looking in each single applications graph (CPU load and Memory) it seems that nothing happen just before the issue. Very often the issue happen between 4 and 8 PM. Never happend during the night. Three weeks ago, we spent one week out. The home remained empty and I connected from remote several times per day. No issue at all, till when we came back. I don’t know if was just a case.
Moderator note: Deleted one screenshot where homey hostname was visible. Hostname has the cloudId in it.
2 Likes
I"d request a (used) replacement unit to use as temporary main Homey, for you to test run your Homey empty (factory resetted).
Then it will show if it’s a faulty unit or not.
And,
when the replacement Homey keeps on going, it’s a simple conclusion right?
Imho Athom can’t seriously let you test-run your Homey, empty, for, possible scenario, a whole week.
Unfortunately the 2 week Xmas holidays started…
Still I’m very curious which apps you disabled resulting in Homey running longer until the reboot stuff occurs
What is very strange is not the crash itself rather then boot loop after… true is it’s also partially Athom design issue, as they resolve any major crash by reboot of the unit itself… eg. when container dies, they reboot whole system. If service fails upon restart, they reboot whole system…so eg. if Bluetooth service fails for whatever reason,it reboots unit…
Btw if you would like to troubleshoot, you can take USB backup, decompress partitions using 7ZIP and have a look on User partition on the logs yourself. (it’s not straight forward though) Depends how much you are skilled and mostly how much time you would invest. I would already go for Warranty replacement, if you have that option
How long do you think I need to keep it unplugged to completely clear the memory? If I can at least stop the loop, that would be a big step forward. Would 15 minutes be enough? I noticed that 5 minutes doesn’t seem to make much of a difference. This morning, I wrote to Athom to request a trial of the Homey Pro 2023 so I can run tests on mine. Let’s see what they say. Regarding your suggestion, I’m not deeply into IT. I use ChatGPT to support me with HomeyScript, and I could probably follow what you’re suggesting as well, but unfortunately, time is not something I have in abundance. Anyhow thanks @Sharkys and @Peter_Kawa . If I had to start from scratch it would be a “tragedy”.
