Following the Node-RED update and the RFlink issues, I decided to replace my 433 MHz plugs with Zigbee2MQTT ones.
Tonight I notice the terrace lights won’t turn on; a glance at the dashboard and I see that if I set the temperature sensor graphs of my refrigerators to « Last hour », they return no values, and when I go to http://gladys.local:8080/#/map to view the map nothing is found — in short Zigbee2MQTT has crashed, about 3 hours ago since the motion detector shows the last value received was 3 hours ago.
As I’m not near my Raspberry Pi I connect via ssh and run sudo reboot.
A few seconds later everything goes back to normal.
Later I return to the dashboard and find that Zigbee2MQTT has crashed again, about 2 hours ago this time, again indicated by the motion detector.
Another sudo reboot and once again everything restarts.
I also noticed that when adding a Zigbee2MQTT device you also have to reboot Gladys.
However, I noticed in Settings > Tasks that the Gladys backup showed 53%.
Since then I can’t even find that Task.
Can the backup cause part of gladys to crash? I’m on an SD card that’s not exactly new.
I haven’t noticed any new crashes since, apart from the loss of connection of the 2 free Sonoff temperature sensors received with orders from Domadoo and that I installed today.
Also, I have an infrastructure that’s starting to become substantial:
Either with docker logs in the CLI (you run docker ps beforehand to find the container name)
Or maybe in the interface, I think I see a « Logs » tab on your screenshot
Ah, indeed, while the backup is running (the first part of the backup), Gladys is temporarily unable to ingest new data because the database is locked.
Normally backups are between midnight and 2:00 AM, is that what you observed?
Normally this part is very quick, but if you’re on a somewhat old SD card, it’s possible it could take a few seconds/minutes.
It’s the SONOFF, I’ll need to flash it, because I’m at 36 devices and I still have 4 to install, the tutorial given by Domo-blog is for Windows, at my place everything runs on Linux even the phone so I need to look for the procedure for Linux.
For the 2nd blockage it was after midnight, for the first it was more around 9:30 PM, is it possible that the backup was done earlier than expected, because my backup is 107 MB, I’m on ADSL 512 — if it’s just me, my wife and my daughter in the house it takes about 30 minutes to upload those 107 MB.
Today there are 30 people in the house, with almost as many phones, tablets and computers so more than 2 hours to upload those 107 MB would not surprise me.
No, that on the other hand isn’t possible — it’s really always between midnight and 2 AM (if you’re indeed running the latest version of Gladys)
In my opinion you need to look elsewhere; it doesn’t seem to resemble a backup problem, especially since during the backup upload Gladys is not blocked — the freeze is really just for the DB copy step, and it’s relatively quick and not dependent on your network.
