> From: Charles Polisher [mailto:cpol...@surewest.net] > Sent: Wednesday, November 06, 2013 10:23 AM > > Sadly, I have direct experience of exactly this, a drive wrote data > while retracting the heads, a spiral of nonsense resulted. Also > reference Chris Seibenmann's blog for some ways in which ZFS can > become corrupt, also interesting his findings on SSD's transparently > de-duping redundant superblocks for you. Also check USENIX ;login: for > 'silent data corruption'. That you haven't encountered a problem is > not the same as the problem can't happen.
Oh - I didn't say corruption can't happen. I said thanks to journaling and correction on the fly, there's no point to fsck after system crash, because I have the belief that fsck happens automatically on the fly. (Apparently that's an implementation detail, which deserves fact check. But I haven't bothered.) I always use ZFS now. Even in ZFS, corruption can creep in, but thankfully, it will discover it when you try to access it (and if a good copy exists, it will get corrected too.) For this and several other features, I put ZFS underneath everything else in the world now. Windows servers and linux servers all virtualized and stored on ZFS remotely. Wish I could say the same for laptops. _______________________________________________ Tech mailing list Tech@lists.lopsa.org https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
