On Apr 04 16:16, 1998, Torbjorn Mohn of 2:211/37 wrote:
>> As I told Toriq in a separate message - THERE IS NO Scandisk feature
>> built in to NT.  You must use "chkdsk".  You can run chkdsk from a
>> normal command prompt within NT or you can schedule chkdsk to run each
>> time upon bootup by placing "/SOS" after the normal boot line in
>> boot.ini.
TM> Running chkdsk every time you boot NT is normally not neccesary, is it? 
It 
TM> only makes the boot up time about 5 minutes longer? :)
>> When run from a command prompt within NT, use /f as a parameter if you
>> wish to repair any errors.  NT will not repair errors on a drive that's
>> in use or the boot drive.  It will, however, ask you if you wish to
>> schedule the repair on the next bootup.
TM> Or, run the Disk Administrator, It can all be done from within there.
>> And...NT is not "self fixing".  Cluster errors, orphaned long file
>> names, etc. must be fixed manually by using chkdsk.
TM> Well... may be we are not speaking about the same thing then?
TM> First of all, I just ran a little check here on my BBS computer (NT 4.0 
TM> Server SP3). It has been running for about 4 months continously now, 
TM> without a reboot, and chkdsk did not find any errors at all! When I ran 
the 
TM> same computer (with the SAME HD under DOS 6.22/DV2.73/QEMM8.0), I had 
lost 
TM> cluster problems every other day! So it's my experiance that NTFS is much 
TM> more reliable and stable in this area than any of the FAT systems.
TM> I also looked up the Course Compendium from Microsoft: Microsoft 
Education 
TM> and Certification; Supporting Microsoft Windows NT 4.0 Core Technologies, 
TM> Student Workbook, and the section about the NTFS filesystem states in 
TM> Module 5, pages 152 and 153 the following:
TM> NTFS additional features:
TM> ....
TM> Transaction-based recoverability. NTFS has high reliability. It has a 
TM> recoverable file system that uses transaction logging to log all folder 
and 
TM> file updates automatically. This is used by Windows NT to redo or undo 
TM> operations that failed due to system failure, power loss, and so on.
TM> Bad-cluster mapping. NTFS supports the recovery technique of bad-cluster 
TM> remapping. If an error occurs because of a corrupt sector on tha hard 
disk, 
TM> NTFS allocates a new cluster to replace the cluster with the corrupt 
TM> sector. NTFS marks the original sector as corrupt. This is transperent to 
TM> any applications performing disk I/O.
TM> ....
TM> NTFS Implementation Considderations:
TM> Recoverability is designed into NTFS. Users do not have to run a disk 
TM> repair utility on an NTFS partition.
TM> ....
TM> This is what I meant when I stated that NTFS is "self fixing". Am I wrong 
TM> here? Or are we talking different things?
Misunderstanding ?? What you stated did not say self fixing. My understanding
is that NT works (maps / marks) around the coruptions. Transaction loging is 
more like keeping a record of changes as a backup in case of disaster.
If someone is making a Bank to Bank transfer of millions of bucks, and the 
power gets yanked while one bank has sent but before the receiving bank has 
received, you would wish you had a fall back. Transaction logging provides
that assurance. Sort of says. Oops! lets  go back to where we were before 
your power gave up on you !!
Chuma
--- Msged 4.00
---------------