I posted the same problem an May, 31.
Now it’s back…
I’m running hass.io on a raspberry.pi
‘Suddenly’ the file system changes in to read only.
I can describe the circumstances.
I was logging in to hass.io with a ‘distant’ machine, using a zerotier connection.
I edited configuration.yaml, tested the file and rebooted the raspberry pi.
It works fine: the changes I made are reflected in the dashboard.
Then I edited automations.yaml.
If I try to save the changes, I get the horrible Errno 30 - read only file system
I can log in to hass.io by ssh.
I can find the /etc/fstab, but I can’t edit it… because the file system is read only.
By the way this is the fstab-file
core-ssh:/etc# cat fstab
/dev/cdrom /media/cdrom iso9660 noauto,ro 0 0
/dev/usbdisk /media/usb vfat noauto,ro 0 0
For now, I shall do what I did last time… just make a new SD-card and start all over again…
All Linux operating systems default to read only on the file system when it detects any errors. IE Bad Blocks. It’s a way to protect itself. I don’t run a Pi but i’m sure some form of fsck on the sdcard should correct the issue.
Also, the SDcard make/speed is critical on a Pi.
Go to your console and type dmesg, you will probably find a lot of SDcard IO error messages. I had the same thing happening with a Beaglebone Black running zigbee2mqtt with log level info off an SD card. SD card was ready for the trash (it was a Kingston 4GB one, which is good quality normally).
Yep I think it will happen sooner or later with a pi running with a sd card. Maybe it’s time to upgrade to a NUC or equivalent with docker.
You may face such problem if suddenly you got the partition full, so the system will be mounted as read only. This is one of the cause.
Perhaps the database has grown to big. Just to log in and try
df -h /
to understand whether it’s the real cause.
Did you ever figure out what caused the issue? I’m on a NUC with M2 SSD and I am experience a similar issue.
First time happened a couple weeks ago. Yesterday I realized I cannot reach the UI and had a look today (connected a monitor). I saw ton of logs on the screen from systemd indicating that the file system was in read-only. Upon reboot, HASSIO did not start successfully. It went till starting the Docker container and then immediately rebooted.
I wonder if the M2.SSD is the issue and/or the HASS IO filesystem/setup. The NUC is approx. two year old and ran Ubuntu before without issues.
However, the incident also lines up with a new release (11/03).