I tried in a private mode of my browser too, to check if there is caching error. But same thing.
I updated to 25.10.0.1 yesterday, I was on 25.10.0. But the GUI did work yesterday (as I have to unlock some datasets after each reboot) as well as this morning.
EIO is an unexpected error when retrieving dataset properties. libzfs in theory have ENOENT set if the dataset is inconsistent due to replication. Can you PM me a debug please?
For future reference: From what I understand from Andrew, it seems to be caused by a failing pool and then excessive I/O operations causing issues with the GUI.
It is back now, though my mirrored SSD pool is showing some errors after scrub.
At the library level this means we’ll show the property as having the value ”<IOERROR>” rather than raising a runtime error. It’s a toss up about what’s a better way of handling it here, but with pool throwing IO errors on reading properties there should be red alerts in other areas of truenas.