> Maybe an option to scrub... something that says "work on bitflips for > bad blocks", or "work on bitflips for bad blocks in this file"
I've suggested this, too. But in retrospect, there's no way to detect whether a bad block is indeed due a bitflip or not. So each checksum error, ZFS might just spent several hours on. You would work some idiot detection into ZFS by having it sum the values of each byte in the filesystem block and store it in a 32bit value, and when scrubbing with bitflip correction, see if the difference in sum didn't deviate by more than -/+ 127. BUt that'd incur more computing power required on writes and actually the availability of a 32bit field in the metadata block referring to the FS block. -mg _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
