me wrote: >> All that said, I'm still occasionally tempted to bring it back. >> It may become more relevant with flash memory as a storage medium. >> > > How common would be single on-disk bit flips in 128K blocks? Disk > manufacturers quantized it as a 1 to 10 to the power of god knows what, > which practically means every few years or so. If this is just optimistic > marketing crap, wouldn't it be viable to have a bit flip checker as option > to the scrub mode (with tons of warnings, yes/no confirmation and > recommendation to do this in single user mode)? 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. >
Most enterprise class disks are rated at 1 uncorrectable read error for 10^15 bits(!) read. For a 1 TByte disk, that means you can expect an uncorrectable read error about once for every 175 times you read the entire disk. Contrast this to consumer class disks which are UER 1 in 10^14, or 17 times for a 1 TByte disk. I posted some of our measured field data a while back, http://blogs.sun.com/relling/entry/zfs_copies_and_data_protection -- richard _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
