Linked by Thom Holwerda on Fri 11th Jan 2008 21:39 UTC, submitted by nicoladagostino
Thread beginning with comment 295671
To read all comments associated with this story, please click here.
To read all comments associated with this story, please click here.
Silent data corruption can only be healed if there's redundancy. That requires either ditto blocks for data or actual physical redundancy. Latter is unlikely to be seen in generic consumer machines, former probably just on certain filesystems created for that purpose (e.g. the document folder).
ZFS stores copies of written blocks in unused blocks and as such resolves the problem. There are of course CRC checks and such also. I remember having read that it stores actually three copies of a single block all spread out in the pool but those extraneous copies aren't counted as used disk space and so can be overwritten by data if space on the pool is getting small.






Member since:
2006-03-20
Silent data corruption can only be healed if there's redundancy. That requires either ditto blocks for data or actual physical redundancy. Latter is unlikely to be seen in generic consumer machines, former probably just on certain filesystems created for that purpose (e.g. the document folder).
(PS, I'm a ZFS on Solaris Express user with a ZFS mirror.)