On Mon, 3 Mar 2008, me wrote: > I'm sure people using no redundancy (e.g. future OSX users) would > appreciate it, saving some grief if the bad blocks are indeed just > single bit flips.
In case people have somehow forgotten, most other filesystems in common use do not checksum data blocks. In spite of this, we rarely hear users wailing about single bit flips in their files. Instead we usually hear about people who find whole chunks of their file missing or overwritten, or find that the hard disk does not spin up at all any more. As we move toward solid state storage, the typical error cases will surely differ. Since ZFS is smart and is able to perform tasks in the background, one possibility to consider is to use otherwise unused storage space to store "weak" ditto copies or even forward error correction data. However, rather than explicitly writing these blocks during normal I/O, they could be created by a background task, and reused for other purposes when required. In this way, otherwise unused disk blocks would be taken advantage of in a similar way that otherwise unused memory is used to cache filesystem data. If the filesystem becomes very full, then there would be less protection but if the filesystem has plenty of free space then there would be lots of protection. Bob ====================================== Bob Friesenhahn [EMAIL PROTECTED], http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/ _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
