Large-capacity XFS Exception Handling

Large-capacity XFS Exception Handling today, some data exceptions were found when I checked the alarm. When I logged on to the backup machine to view the File status, I found that the disk space was full and inode was faulty, therefore, all the backup processes are stopped and the problem is checked. First, if the disk space is full, try to delete some files and find that some files cannot be deleted, and the file status is abnormal. Then, copy all the scripts, back up system logs, back up scheduled tasks, and perform various backups. And then start to work. It cannot be deleted. Fortunately, a few of them are quite normal, and the space of GB is restored at last. However, since there are a lot of backups each time, all the backup plans should be stopped first. Take a look at the inode space and release it. Then, detach the disk and attach it. inode is normal, but the above exception still exists. Unable to proceed.
Later, I asked my colleagues in the system group for help. During the handling, I suddenly said that XFS-format partitions may encounter exceptions in the case of large capacity, I suddenly remembered that I had handled another similar machine this year, but the machine was prompted that the disk space was insufficient, but df and df-I both showed that there was no problem, there was a fault with insufficient Life and Death. At that time, an inode64 was added to/etc/fstab, therefore, we tried to add this parameter to the configuration file of this machine, unmount, mount, and finally succeed!
The problem still occurs when XFS supports large disk space. By default, inode files are stored in the first T, therefore, when the disk space usage is large, similar problems may occur. Therefore, we finally decided to add the inode64 parameter to all the machines that use XFS to avoid similar problems from happening again.