Comment 0 for bug 2023983

During an automatic btrfs balance (triggered through an btrfs maintenance timer) there happens a Enospace error.
The next automatic "btrfs balance" step leads afterwards to an kernel error (BTRFS: Transaction aborted (error -28)) and a readonly filesystem.
The system must be rebooted to get a writable filesystem. A scrub doesn't show any erros.
After canceling the paused balance a new manual full balance could be successfully executed.
In the attached syslog file there are the extract of the balance process with the kernel stack trace of a 5.15.0-72-generic kernel.
I've seen on other systems (with former kernels of Ubuntu 22.04 LTS) these problems with readonly filesystems after btrfs balancing.