The database is stored in the Docker volume that you pass as a parameter to your docker run.
On the site’s command it’s explained just below:
So find the folder you set during your docker run, and empty the contents of that folder.
Then restart Gladys and you should end up on a fresh new instance!
Hello,
I bought exactly the same sensor that I should receive this week 
I’d like to help you with the tests; I’d like to be able to install a test instance in parallel with my production (on my mini PC with a Zigbee USB dongle) but I don’t want to break anything.
I thought I had seen a tutorial somewhere but I can’t find it anymore.
Hi @Nagromdark
Can you tell us a bit more about your current setup?
We can then better advise you on how to set up a test Gladys.
FYI I’m picking mine up today 
Hi @mutmut,
I found the tutorial and followed it: Tutoriel: Lancer une image Docker de test
I followed the tutorial and installed the test image by doing this:
sudo docker run -d \
--log-driver json-file \
--log-opt max-size=10m \
--cgroupns=host \
--restart=always \
--privileged \
--network=host \
--name gladys-zigbee2mqtt-tuya-me201wz-test \
-e NODE_ENV=production \
-e SERVER_PORT=8001 \
-e TZ=Europe/Paris \
-e SQLITE_FILE_PATH=/var/lib/gladysassistant/gladys-production.db \
-v /var/run/docker.sock:/var/run/docker.sock \
-v /var/lib/gladys-zigbee2mqtt-tuya-me201wz-test:/var/lib/gladysassistant \
-v /dev:/dev \
-v /run/udev:/run/udev:ro \
gladysassistant/gladys:zigbee2mqtt-tuya-me201wz
I made sure to change the port and the storage space.
The instance is accessible on port 8001, no problem. But where I started having issues was with the MQTT broker and Zigbee2MQTT that I wanted to connect to the test instance. I thought that both Gladys instances could be connected to these containers but when I enable it on one it disconnects the other, it seems..
So for now I’ve stopped the test instance to avoid risking my production
Here are my containers:
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
caffb1ddd7e6 koenkk/zigbee2mqtt:1.42.0 "docker-entrypoint.s…" 43 minutes ago Up 43 minutes gladys-z2m-zigbee2mqtt
e2b29c057af6 eclipse-mosquitto:2.0.15 "/docker-entrypoint.…" 44 minutes ago Up 43 minutes gladys-z2m-mqtt
f869709d2ec6 eclipse-mosquitto:2.0.15 "/docker-entrypoint.…" About an hour ago Up 48 minutes eclipse-mosquitto
55728634212e gladysassistant/gladys:zigbee2mqtt-tuya-me201wz "docker-entrypoint.s…" About an hour ago Exited (0) 45 minutes ago gladys-zigbee2mqtt-tuya-me201wz-test
22290cf834c4 nodered/node-red:3.1 "./entrypoint.sh" 2 weeks ago Up 2 weeks (healthy) 0.0.0.0:1881-\u003e1880/tcp, [::]:1881-\u003e1880/tcp gladys-node-red
a5e4134c21d6 gladysassistant/gladys:v4 "docker-entrypoint.s…" 2 weeks ago Up 2 weeks gladys
942fa3b70db8 containrrr/watchtower "/watchtower --clean…" 3 weeks ago Up 3 weeks (healthy) 8080/tcp watchtower
And one question in passing: I don’t understand the difference between gladys-z2m-mqtt and eclipse-mosquitto, it’s the same image but with two different names, if someone can enlighten me 
Thanks for your help
First, for your question :
Your container named gladys-z2m-mqtt is your MQTT instance which is based on the eclipse-mosquitto image, so it’s normal that they have the same size.
Regarding your procedure, everything is perfect up until the moment of adding MQTT to your test Gladys.
I think you enabled MQTT like this :
There are 2 containers that use the mosquitto image; that’s normal.
There is an MQTT container for zigbee2mqtt and another MQTT container for the MQTT integration.
These 2 integrations do not use the same container.
@pierre-gilles I was able to install the latest version of your build. And I can see the module in the Zigbee integration under added devices, with 6 features:
And when I go to the dashboard and add a « Devices » block, I can choose 4 of these features:
I confirm that the unit in ‹ m › and the names are handled correctly 
The two missing features (‹ Installation height › and ‹ Maximum liquid depth ›) are not a concern, I think, because they are not data that will vary: they are the parameters to configure in zigbee2mqtt to size the tank and for the ‹ Liquid level (percentage) › feature to be calculated by the module.
I think you shouldn’t try to retrieve them from the device definition in Gladys. (I didn’t list them in my initial request.)
Thanks so much @mutmut and @_Will_71, I managed to launch a test instance in parallel on the same PC and I’m using the same zigbee2mqtt and broker as production. It’s perfect — messages are arriving fine in both instances
I just have to wait for the sensor to test (hopefully tomorrow), but I see that @StephaneB has already done everything ^^
At least now I know how to install a test image, now I just need to set up a dev environment but I’d need another PC ^^
Thanks for testing @StephaneB
Ok, so I’ll remove them if they’re not useful — better not to clutter if it’s a parameter and not data!
[quote=« Nagromdark, post:28, topic:9495 »]
Great thanks @mutmut and @_Will_71, I managed to run a test instance
Hello,
as for me it works well too:
You need to set the maximum liquid height and the sensor’s positioning height in zigbee2mqtt, and that’s it.
I just installed it, but for those wondering this sensor is really very accurate — not bad! We’ll see over the next few days whether the readings are correct
I can’t wait to have it in production 
so I can refine my automatic watering scenes (water for a shorter time if the level becomes low) and secure the operation of my lift pump (disable it if the water level becomes critical, to prevent it from losing its prime).
I love all these little scenarios that Gladys lets me create! Thanks @pierre-gilles
It’s coming soon! I’m working on Gladys tomorrow
I received mine yesterday and what a nightmare for testing 
I took the solar version, I thought it might be coming from that but no.
I tried putting it under a chair to get the distance, without much success because I got negative or positive distances, probably because I didn’t manage to configure the module.
I then tested with a water carafe and that was better but I had a small rather odd difference.
And today I installed it on my tank, the values seem good but you have to be super precise in the setup and therefore understand it well 
And finally, I have a few issues getting the module to apply my new settings, I don’t know if it’s due to my z2m which is on 2.2.1 or my SLZB-06M antenna which is on ember.
Anyway, I have a bit of investigating to do, sorry for not bringing more positive feedback.
Isn’t it a holiday?
It’s fine, it’s been removed from the PR, I rebuilt a Docker image with the same tag!
I don’t take public holidays, I prefer to work so things move forward!
@mutmut, I have the mains-powered version but I think it’s the same. Be careful, it works with ultrasound from what I understand so you need to position it in the middle of the tank and not near the water inlet. It shouldn’t be close to the walls, you need a fairly large tank otherwise the signal will bounce off the walls. So a glass of water won’t work
For the configuration you need to measure the maximum height the water can reach and also the height at which your sensor is positioned. For example I have 0.83 and 1.05 for those values and the reported % is correct. If it doesn’t look correct you need to tweak those values until it does
A quick question about the device:
Do ultrasound waves pass through the wall, or does there need to be a hole in the wall where the device is placed?
You need a clear passage (not necessarily a hole) otherwise it gets stuck.
For my part, I had to secure the module inside the cap because there wasn’t enough space above my tank and so I didn’t need to drill (ok, I still made a 25mm diameter hole to pass the cable and install 4 screws)
Thank you @Nagromdark, that’s also what I do to get the best result. On my side I set 1.02m (module in the cap and not above) and 0.96m for my overflow. However, these are values I took from the outside because I don’t have room above my tank to access it (I only have 5 cm of clearance) and the lengths are not necessarily exact.
@pierre-gilles I’ve tested this new image, and it’s all good for me.
