Na página mdadm man:
When relocating the first few stripes on a RAID5 or RAID6, it is not possible to keep the data on disk completely consistent and crash-proof. To provide the required safety, mdadm disables writes to the array while this "critical section" is reshaped, and takes a backup of the data that is in that section. For grows, this backup may be stored in any spare devices that the array has, however it can also be stored in a separate file specified with the --backup-file option, and is required to be specified for shrinks, RAID level changes and layout changes. If this option is used, and the system does crash during the critical period, the same file must be passed to --assemble to restore the backup and reassemble the array. When shrinking rather than growing the array, the reshape is done from the end towards the beginning, so the "critical section" is at the end of the reshape.
Basicamente, ao crescer uma matriz RAID-6, há um breve período (normalmente alguns segundos) durante o qual não há redundância nos dispositivos RAID. Essa redundância pode ser fornecida por um arquivo armazenado em um dispositivo que não faz parte do array.