ESPHome - All devices periodically disconnect (see log details)

klogg · August 26, 2020, 8:44am

I have several ESPHome nodes that ALL disconnect at (virtually) the same time at random. They all reconnect after some seconds.

I have recently migrated from Ubuntu/HA to Proxmox. Could there be something amiss with my ESPHome Integration following the snapshot restore from one ‘system’ to another or is this signs of a more fundamental network issue?

Here are the logs for two devices showing (virtually) the same times.

esp32_01 log

[09:13:13][D][sensor:092]: 'esp32_01 WiFi Signal': Sending state -57.00000 dB with 0 decimals of accuracy
WARNING Disconnected from API: Timeout while waiting for message response!
INFO Connecting to esp32_01.local:6053 (192.168.1.207)
INFO Successfully connected to esp32_01.local
[09:13:34][D][api.connection:583]: Client 'Home Assistant 0.114.3 (192.168.1.25)' connected successfully!
[09:13:34][D][time:029]: Synchronized time: Wed Aug 26 09:13:34 2020
[09:13:37][D][api:067]: Disconnecting Home Assistant 0.114.3 (192.168.1.25)
[09:14:09][D][sensor:092]: 'esp32_01 Uptime': Sending state 2015.99097 s with 0 decimals of accuracy

esp32_01 log

[09:13:11][D][sensor:092]: 'Loft Humidity': Sending state 57.30000 % with 1 decimals of accuracy
WARNING Disconnected from API: Timeout while waiting for message response!
INFO Connecting to esp32_02.local:6053 (192.168.1.190)
INFO Successfully connected to esp32_02.local
[09:13:34][D][sensor:092]: 'esp32_02 Uptime': Sending state 1966.03198 s with 0 decimals of accuracy
[09:13:34][D][api:067]: Disconnecting Home Assistant 0.114.3 (192.168.1.25)
[09:13:35][D][api.connection:583]: Client 'Home Assistant 0.114.3 (192.168.1.25)' connected successfully!
[09:13:35][D][time:029]: Synchronized time: Wed Aug 26 09:13:35 2020
[09:13:35][D][xiaomi_ble:161]: Got Xiaomi LYWSDCGQ (58:2D:34:33:B1:85):
[09:13:35][D][xiaomi_ble:164]:   Temperature: 20.4°C
[09:13:35][D][sensor:092]: 'Upstairs Hall Temperature': Sending state 20.40000 °C with 1 decimals of accuracy
[09:13:36][D][api:067]: Disconnecting ESPHome v1.14.5 (192.168.1.25)
[09:13:59][D][xiaomi_ble:161]: Got Xiaomi LYWSDCGQ (58:2D:34:33:AB:4D):
[09:13:59][D][xiaomi_ble:164]:   Temperature: 20.5°C

For info, I have a Unifi network (the UniFi controller is on a Proxmox Container)

Thanks for any help.

tom_l · August 26, 2020, 9:33am

Something amiss with your network or the home assistant api. As others are not reporting it my money is on an issue with your network.

Can you set up a ping command on another machine on the network and ping, 1) your router, 2) home assistant?

Check the results around the time your ESPs drop off.

klogg · August 26, 2020, 8:56pm

Thanks, that’s a good idea.

I’ve been away all day today but as luck would have it there were no occurrences all day anyway (I have a persistent notification when it does).

I’ll give this a try tomorrow…

Ben.S · August 26, 2020, 9:55pm

I had an issue with this happening after I had installed room-assistant on my rpi that is also running hass. Turns out the on board Bluetooth was clashing with the wired Ethernet controller. I put a usb Bluetooth adapter in and all was good again.

Different problem to you, but effectively the same in the fact it’s certainly network related and not a general issue.

klogg · August 27, 2020, 1:28pm

Well it looks like HA is becoming unreachable…

Reply from 192.168.1.25: bytes=32 time<1ms TTL=64
Request timed out.
Request timed out.
Request timed out.
Request timed out.
Reply from 192.168.1.25: bytes=32 time<1ms TTL=64

I can’t guarantee this is at the same time because I am not sitting here waiting for it to happen but I have caught it twice now so I feel confident.

I’m not sure if this helps…

tom_l · August 27, 2020, 1:49pm

And your router was still reachable at that time?

klogg · August 27, 2020, 1:51pm

Yes it was.

klogg · August 27, 2020, 1:55pm

I don’t know if it helps but I have also noticed that my Android tablet seems to give this error at about the same time

Login attempt or request with invalid authentication from 192.168.1.223

laca75tn · August 27, 2020, 2:44pm

This is a known problem with Esphome. Some of the solutions include:

Use static IP addresses for your nodes
Use a fixed/lower band width in your router/wifi
Use a fixed channel in your wifi

klogg · August 27, 2020, 2:47pm

Thanks I’ve seen that but I think that relates to random disconnects of individual devices.
For me it is always EVERY device.

laca75tn · August 27, 2020, 2:50pm

I had all my devices disconnect at once several times per day. This fixed the problem.

klogg · August 27, 2020, 2:51pm

@tom_l Just for the record I found another NirSoft utility for Ping that gives timestamps (I think there is a NirSoft utility for almost anything!) and here is evidence that only my HA becomes unreachable.

HA Server

27/08/2020 15:45:10	192.168.1.25	0	64	Succeeded	
27/08/2020 15:45:11				Request Timeout	
27/08/2020 15:45:13				Request Timeout	
27/08/2020 15:45:15				Request Timeout	
27/08/2020 15:45:16				Request Timeout	
27/08/2020 15:45:18				Request Timeout	
27/08/2020 15:45:20				Request Timeout	
27/08/2020 15:45:23				Request Timeout	
27/08/2020 15:45:25				Request Timeout	
27/08/2020 15:45:27				Request Timeout	
27/08/2020 15:45:29	192.168.1.25	0	64	Succeeded

Router

27/08/2020 15:45:10	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:11	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:13	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:15	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:16	192.168.1.1	11	64	Succeeded	
27/08/2020 15:45:18	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:20	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:23	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:25	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:27	192.168.1.1	0	64	Succeeded	
27/08/2020 15:45:29	192.168.1.1	0	64	Succeeded

klogg · August 27, 2020, 2:55pm

Ah, ok. Thanks.
Do you know which thing fixed it?

The current docs say:

Some steps that can help with the issue:

If you’re using a hidden WiFi network, make sure to enable fast_connect mode in the WiFi configuration (also sometimes helps with non-hidden networks)
Give your ESP a static IP.
Set the power_save_mode to light in the wifi: config (only helps in some cases, in other it can make things worse). See Power Save Mode.

I am not using a hidden WiFi Network but I have set fast_connect anyway
power_save_mode seems like poor fix if it can make it worse.

I can try a static IP.

Also, did you fix the IP in ESPHome or in the router. (Does it matter?)

laca75tn · August 27, 2020, 4:09pm

Set the IP on the esphome node.

Setting the width of the wifi to 20 mhz and fixed channel instead of auto is what really made a difference for me. In auto mode, the router changes channel and I think that was the main issue for disconnecting all nodes at once

klogg · August 29, 2020, 10:08am

I wanted to thank you for your input.

I have another thread on a related topic but approaching it from a different angle. I posted a thankyou there and tried to copy it here but the forum won’t allow me to post another reply with the same content.

Here is a link to it but TL:DR, I took the nuclear option and rebuilt everything. Seems to be all ok now.