The disk cannot keep up with database updates. The temporary fix is to restart your node. But this is not for long.
How your disk is connected to this device?
Is it SMR (see PSA: Beware of HDD manufacturers submarining SMR technology in HDD's without any public mention)?