On Sat, Jun 22, 2013 at 6:48 AM, David G. Miller <dave@xxxxxxxxxxxxx> wrote: > Chiming in with some additional information that only *partially* > contradicts certain things that have been said in this thread. First off > though, the advice that drives are cheap and data is expensive is absolutely > correct. Do NOT let anything I say talk you out of making sure any critical > data on this drive is backed up. > > Given that tee-up, smartctl/smartd reports that the disk has an > "uncorrectable bad sector" when there is a read error from the drive for a > sector. The error is "uncorrectable" because the sector cannot be read. > Note that the detection of a bad read (or write) takes place at the physical > and drive firmware level when the CRC is checked. The only thing that the > drive has to work with is that there was an attempt to read a sector and > that read resulted in a CRC error. > > The bad sector is part of a file and only you, the user, can make a > determination as to whether the rest of the file is still good or if the bad > sector is throwing a CRC error but the file is still usable. That's also > why the error is "uncorrctable". The drive doesn't have enough information > to fix it and it can't silently remap the sector since it can't read the > data. If it did, you would end up with a file with a null sector somewhere > in it at the location that corresponds to the bad sector's data. > > Write errors the drive takes care of through the reallocation process > mentioned earlier in the thread (since data is being written, any existing > data is being replaced so the data can be written to a remapped sector). > Read errors the drive can only report the problem since the read error > implies that data cannot be retrieved. > > My advice: buy a new drive but run badblocks -w on the old drive once you > have your data safely off of it. You will probably find that the badbloocks > write test (-w) lets the drive see the bad sector being written to and then > remaps the bad sector and you end up with a drive that is now completely > usable again. Be absolutely sure you have your data off of the drive before > running badblocks -w. It will overwrite any data on the drive. > > I have "recovered" several drives by doing this. I've also had some that > threw errors all over the place. Those became targets. Thanks, Dave, for your very clarifying answer. Should I conclude from your words that I have already some corrupted files? If so, is there some way to identify them? Paul -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
