Before you do this I would definitely run Norton from it's CD and see
whether it can solve the problem more easily. If it runs into severe
problems it will tell you.
Johannes
On 29.08.2003 12:01 Uhr, Dennis W. Manasco wrote
> Andrew --
>
> From what you've said I'm not certain _precisely_ what Norton is
> telling you, but I will comment with a precis of the problem as I see
> it.
>
> If you have bad blocks then the data in those blocks has been
> compromised. To put it more technically, if the sector check-sums
> that have been written for a block do not consistently match (over
> multiple reads) the calculated check-sums of those sectors' data then
> either the sectors are inconsistently readable and the data within
> them cannot be relied upon or the recorded check-sums are
> inconsistently readable and they cannot be trusted, which is
> essentially the same thing.
>
> This cannot be corrected without copying the data (in its uncorrupted
> form) from the suspect blocks onto other, more reliable, blocks.
> However, a repair program cannot know whether to trust the data it
> receives from the individual sectors or that from the check-sums, so
> it has to take the most conservative approach and assume that the
> data in the individual sectors is unreliable and that the problem is
> thus insoluble without replacement of the data from a backup. This is
> generally a very good approach.
>
> Statistically it is far more likely that the information in the bad
> sectors is compromised than that the check-sums are at fault. And,
> regardless, data cannot be recovered using the check-sums; it can
> only be checked for validity. There are programs that will repeatedly
> read a sector and attempt to write a consensus version to another
> file, but these programs are for very specific data-recovery
> solutions and would probably not be useful to you.
>
> If you have bad blocks in the System area (which involves quite a
> number of files in OS <=9.x and a tremendous number in OS X) then the
> system will be unreliable. How unreliable it is depends on the
> severity of the block degradation and the exact code that is
> affected, but it will eventually cause the system to be unusable.
>
> If your bad blocks are within user files then I believe that the
> problem can be solved by deleting the problem files, marking the
> sectors as bad and then reinstalling (a known good copy of) the
> questionable files. Unfortunately I am unfamiliar with any currently
> viable program that can mark a sector as bad without a reformat.
>
> The only solution I know for bad blocks within the System area is to
> reformat (thus marking the blocks as bad {ATA, which is on the iMac
> that I believe you have} or remapping the bad sectors from the
> spares{SCSI}) and then reinstalling the operating system.
>
> One way or another a reformat of the drive will probably be required.
>
> What I would personally do: Use Retrospect to backup your entire hard
> disk. Twice. Restore a few files to make sure your backup works. Boot
> from the OS X install disk and reformat the hard disk. Restore the
> entire disk using Dantz's live-restore instructions (n.b. the advice
> to upgrade your System to its previous state before restoring from
> backup).
>
> This may not be a perfect solution if your bad sectors are only bad
> _some_ of the time -- you may wind up restoring unreliable
> information. There are ways around that pitfall, but they rely on a
> greater degree of granularity in the analysis of Norton's errors than
> your question allows. The Dantz live-restore instructions _should_
> protect you from most restore problems.
>
>
> Hope some of this helps,
>
> -=-Dennis
>
>
> _______________________________________________
> Finale mailing list
> [EMAIL PROTECTED]
> http://lists.shsu.edu/mailman/listinfo/finale
--
http://www.musikmanufaktur.com
http://www.camerata-berolinensis.de
_______________________________________________
Finale mailing list
[EMAIL PROTECTED]
http://lists.shsu.edu/mailman/listinfo/finale
