Re: Speed Disk vs Defrag

Tech-Archive recommends: Speed Up your PC by fixing your registry

Bill in Co. wrote:

John John (MVP) wrote:

Bill in Co. wrote:


The cluster points to the next cluster for files that span clusters, the
FAT only points to the first cluster that a file uses, the clusters
point to the next cluster.

That is how it works with FAT/FAT32, but it doesn't work like that at all
with NTFS. On NTFS that information is all kept as attributes in the
MFT, the file system does not need to flip trough each individual
cluster to find the next one, it's all held in a Virtual Cluster Number
to Logical Cluster map (VCN-to-LCN) in the file's data attribute.

That sounds like a more sensible and robust approach, since you don't have
to go out and read all the disk clusters to find out where the next one is
(and pray that one of them isn't corrupt, breaking the chain).

That is not really how it works, (on FAT) the clusters are daisy-chained
in the FAT, the operating system doesn't open successive clusters to
find out where the next one is, it reads that information in the FAT.
The first cluster information for the file is found in the Directory
entry and the other cluster information along with the last cluster is
read from the FAT. In any case, if a cluster in the chain is corrupt
the whole file is usually unreadable or corrupt anyway, chkdsk can
locate and try to repair these bad clusters but most of the time the
recovered clusters are of much use and the corrupt files are not easily,
if at all repairable.

John


OK (and I probably should have known this).
But above, I thought you had said that "the file system (in XP) does not need to flip through each individual cluster to find the next one"

It was in reference to a post by Leytos who said that the FAT only pointed to the first cluster and who suggested that the "chain" was then retrieved from, or that each successive cluster "pointed" to the next cluster in the chain. My reply was in made in that context where I made the point that NTFS doesn't flip through clusters, I wasn't suggesting that FAT did, I was only commenting about NTFS which is the native NT file system.

John
.

Relevant Pages

  • Re: Has anyone ever implemented a directroy as a linked list ?
    ... A file offset is mapped to the cluster by means of FAT chain, ... But NTFS maps file offset to cluster by using a list of extents - contiguous ...
    (microsoft.public.vc.language)
  • Re: to FAT or not to FAT?
    ... > FAT over NTFS has been ignored. ... Cluster size. ... > systems far outperform DOS systems when 1/2 gig and larger partitions were ...
    (microsoft.public.win2000.file_system)
  • Re: Partition and format to FAT32 on an USB external hard drive
    ... The wasted space depends on the cluster size of the drive. ... In NTFS a 125 Gig partiton would be formnatted in NTFS by XP to 2 ... In NTFS that file would take 34 kilobytes of space. ... IN FAT 32 that same file would take 64 Kilobytes of drive space. ...
    (microsoft.public.windowsxp.hardware)
  • Re: FAT32 or NTFS?
    ... >>of a hugely bloated FAT and large cluster sizes, ... resilience --- As far as FAT partitions go, don't forget to wave goodbye to ... technologies imbedded in the file system ensure that data is only written on ... Which brings me to this - In an NTFS file ...
    (microsoft.public.windowsxp.general)
  • Re: Resizing FAT partitions
    ... full cluster, and are thus part of the volume, but unused. ... determining the max amount of space to partition on a given drive. ... each type of FAT with a fixed table used for floppies. ... not be able to copy to such small drives. ...
    (microsoft.public.win32.programmer.kernel)