Andrew Lentvorski wrote: > Paul G. Allen wrote: >> No matter what safeguards are put in place, no file system is ever >> going to be 100% safe from corruption from a power interruption. > > Define "corruption". > > The ZFS stuff really is safe from filesystem corruption. The last thing > written to disk is the checksum for the transaction. If the checksum > doesn't hold, the transaction is considered erroneous and ignored. > > Yes, you can still lose new data in actively open files, but you will > not lose the integrity of the filesystem, ever. And you should never > lose an existing file that has already reached checksum complete. > > That is a architectural design decision, and failures in that get filed > as major bugs.
If only that were true. The fact is that file system corruption can occur AFTER all the lovely checksums, error corrections and what have you have written the data to the disk. Disk defects can take out whole swaths of a disk surface. Power pulses to the write heads at an inopportune time can cause erasure. Random magnetic domain flipping can change data. Error correction codes can only make up for so much damage. At some point you are going to have problems that cannot be corrected/compensated for and you WILL lose data. Thus the need for backups. Gus -- KPLUG-List@kernel-panic.org http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
