-
-
Notifications
You must be signed in to change notification settings - Fork 1.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Aqara water leak sensor reports false leakage after update to latest version #12018
Aqara water leak sensor reports false leakage after update to latest version #12018
Comments
Hello, I had exactly the same problem and after several hours looking for a solution I found a very simple solution to solve this problem, just soak the detector in water so that it sends a leak alert, then just dry it and the status becomes dry in home assistant. Hoping to have helped you ;) |
@xekil thanks for sharing. I will try this out. Probably connecting the two metal pins with a paperclip will already do the trick. |
@groenmarsmannetje with pleasure |
I can confirm that this workaround also worked for me. I will leave the issue open for a while so that others who also experience this, can read this. |
Very annoying … after a couple of restarts the sensors are giving false reports again. |
@groenmarsmannetje could you provide the debug log when the incorrect reports are being send? See https://www.zigbee2mqtt.io/guide/usage/debug.html on how to enable debug logging. |
At the moment the sensors are acting normally again. But will do this next time when it occurs again. |
Having the same problem immediately after updating to 1.25 @Koenkk any idea why this issue now is present @1.25? I'm using z2m for years and having the waterleak sensors at least 2 years but never had this behaviour. Cheers and thanks, Norbert |
@McGiverGim I expect this is is caused by https://github.com/Koenkk/zigbee-herdsman-converters/blob/2351faa29151a1d4ef34a1012bd5159f57454537/lib/xiaomi.js#L215 , before Koenkk/zigbee-herdsman-converters#3994 we didn't parse any message from the basic Xiaomi cluster: https://github.com/Koenkk/zigbee-herdsman-converters/blob/6f422d10d1d10e41cdc5dae22f584c5a4a9d2345/devices/xiaomi.js#L904 |
I will look into it. I've one Aqara water leak so I can try to reproduce. |
Some of you, with the problem, can put the logs in debug mode and look for something like:
with the name of the Aqara device? |
Nobody can publish some log with the problem? In my case, I can't reproduce the problem. This is a log without water:
This is a log with water:
As you can see, the attribute |
Hi, sorry for the delay. I could but after trying the "hint" about sending a "false alert via shorting the two sense contacts" i now fixed it and since last night i did not have any alarms at all. tried today if normal/real alert would work - and it does. Please let me know if i can help any further. As i do not understand the background of this issue, i only hope that this quick fix is somehow the solution and the problem will not come back :-) Cheers Norbert |
I also had the same problem today with my two devices. The problem went away after the first false alarm without any workarounds. However, I wanted to reproduce the issue. The first device continued to send Here's my (filtered) debug log:
It seems noteworthy that the |
Maybe I have found a way to reproduce it. When I only restart the zigbee2mqtt addon the problem doesn’t occur, but when I reboot the host (I use Hassos) then after everything has been started the sensors are reporting leakage, but they automatically turn off after approx. a minute. So in that case the coordinator also had a power cycle. But there is a difference with the first time after installing the new version at that moment the state remained on until I did the manual workaround. Unfortunately I didn’t have debug logging on. |
@groenmarsmannetje If the problem is the code of the genBasic that was added, then it has nothing to do with a reboot of Z2M or HAOS. The code only processes messages sent by the device and the device does not send anything after a reboot because it is not aware of that.
Is better the debug on, but with info you can see too if there is some
Maybe it sends something after a battery replacement, but I tried to remove, wait, and add again the battery without luck. What I suspect is that internally, maybe the device sends a "true" until the first real "true", then it starts working right. I would like to have this in logs, maybe there is a difference that we can use, only to be sure before reverting. Someone at this thread have the problem or it is working as intended now? Some of you changed the battery or maybe repaired it before the problem? |
I am not sure how it works under the hood. But in case of reboot then the coordinator is also power cycled and I do not think this happens when I restart the addon. The memory/cache of stick, mqtt and HA will then also be flushed and reinitialized with wrong (on) state. While that should be state unknown or unavailable (or last known state). Just a thought… I will try to reproduce it tomorrow with debug on. |
The problem has not resurfaced so far. I tried a different angle: I have a daily backup of all my files and looked through the last 20 backups of the state database in
Note that the lines are not in chronological order. If I go further back, I can even see this value changing between 0 and 1, even though I didn't have a water leak back then either. For me it looks like attribute 100 does in fact not designate a water leak, but some other value. |
Strange. In my database the 100 is 0 (as expected). What seems wrong is this value to 1 and this explains the behavior seen here. I think that when z2m is restarted, it uses the value of the database until it receives a real message of the device. For this reason it starts to work when you put the device in water (or you press the button of the device). |
@McGiverGim It contains a |
Then I think we need to see if once this value is right (contains 0), the device fails or not at some moment. |
I'm still running 1.24 and my sensor has showed "water_leak" as null for quite some time (maybe already in 1.23. I did test once that putting the sensor in water changed water leak to true and drying the sensor turned it back to false. But restarting z2m seems to have made the value disappear again. |
This is a debug log with the records for one of the sensors when leak is on (see second line) during startup after reboot and automatically turns off 16 seconds later (see last line):
|
@groenmarsmannetje this log does not contain the The device sends two kind of messages:
Everybody in this thread, how is the state of this bug after some days? It continues happening? I wanted to reproduce before fixing it, but if it continues hurting I will simply revert the change in the |
@McGiverGim The problem has not re-occurred so far. I continue to have debug logging enabled and will provide a log file should this issue appear again. |
I also had this issue at one of my two sensors. After doing the workaround, the sensor no longer sent a false "water_leak: true" messages. Even after rebooting my pi or restarting z2m, no false leak warning appeared. |
I've the same situation as @ggottwald . after workarround the issue is gone.....and did only happen on one of my two sensors. you can imagine how fast i drove home from work when i realized the alarm in the office. later i realized that the remote update i triggered was linked to this..... |
This happened once before #1967 |
Can you get some debug logs of this sensor? Press the button on top to force a genBasic message. |
@McGiverGim Unfortunately, when I click on that sensor there is nothing in the log.
|
Yes, but I wanted one of the sensor with the problem. |
@poisondima seein my logs, the device sends itself a |
@McGiverGim that's exactly what I'm waiting for when the device sends me a request |
@McGiverGim I got the message:
|
I suppose you have not rebooted Z2M, right? Ok, this message demonstrates that something is wrong with the devices at some moment, because it sends the message with a "1" instead of a "0". |
I did a Z2M reboot on 2022-04-09 10:12:32
Took it out of the water:
|
PR pushed to ignore the field. If nobody has some better idea, is the only way at this moment to not receive this kind of false positives 😢 |
What happened?
I have 3 of these Aqare water leak sensors and after updating to version 1.25.0 they all give false leakage reports because it is still totally dry.
What did you expect to happen?
The state of these water leak sensors should be unaffected after update.
How to reproduce it (minimal and precise)
I only updated to new version of the Home Assistant zigbee2mqtt addon.
Zigbee2MQTT version
1.25.0
Adapter firmware version
20211217
Adapter
zzh
Debug log
No response
The text was updated successfully, but these errors were encountered: