ZFS stores information about a file, including its physical location on the drive, as metadata. Like index nodes, metadata information is stored in ditto blocks which a separated as far from each other as possible. The number of metadata copies is a configurable attribute of the pool.
In addition, ZFS computes and stores checksums for all blocks. Thus anything that corrupts the data in the block is detectable, and ZFS can restore the block from a backup copy, if one exists.