Jump to content

Deleting MFT file names


Grinderedge

Recommended Posts

First I would like to greet everyone on these forums and express my gratitude to the Piriform team for their excellent work ;)

 

Like many users I'm looking to delete those pesky filenames entries from the MFT. I am trying to prepare a pair of hard disks in RAID 0, using Diskeeper to reserve enough MFT records to avoid fragmentation as much as possible.

 

Diskeeper does this by writing files on the filesystem, 1 file for each MFT entry the user has specified. I usually set that to 5 million entries. Everything goes fine, MFT grows as expected in one block, files get deleted and fragmentation is kept to a minimum.

 

But then if I use Recuva it will choke trying to show 5 million entries... Nothing really strange so far... next step is to try to get rid of all those name records in the MFT...

 

I came up with an interesting page people should read before pestering Piriform. Be aware that all the tools discussed in those pages aren't benign, they ALL could/shall convert your hard disk to pure energy without any messing around... not that any of those solutions will actually work...

 

DLS Reports

 

Is this option contemplated for a future release of CCleaner or Recuva?

Link to comment
Share on other sites

  • Moderators

As it says in your link, this topic has been discussed many times before.

 

As you've used Diskeeper (with which I am not familiar) to reserve MFT records you've accepted that the MFT doesn't shrink in size as files are deleted, otherwise there would be no point it running it. That's point one.

 

Point two is that you can't 'delete' any file name in the MFT, or erase it, or wipe it. File names (anything on the disk in fact) can only be overwritten. I understand that NTFS will clear out the entire file name field (no no no, not clear out, overwrite - possibly with zeroes) before writing the new filename, so no trace of the old name exists.

 

Presumably Diskeeper uses some harmless file name to create those entries, so what advantage would it be to 'delete' them? You would only be replacing one set of invented names with another. I don't know what length files Diskeeper creates: you could I suppose replace them with zero length files if they're not already, but that would be a lot of effort for very little reward.

 

If you're thinking of reducing Recuva's time to scan, or getting it to scan at all, then you're probably going to be disappointed. In my use of Recuva it has always done a full scan before showing the results, so if the Show Zero Length Files option is unticked it makes no difference to the scan, just filters the results shown.

 

Your final question is really invalid as it can't be done, removal that is. You'll just have to live with 5m file names.

Link to comment
Share on other sites

Point 1: That was my idea, I want it to be as big as to hold 5 million files (current usage is around 3.5 m files on my current system, with tiny and not so tiny MFT chunks all over the fs). I don't want it to shrink or anything, just grow big enough to cover 5 mil. entries.

 

Point 2: I could delete the names with Restoration. Recuva didn't find any name. I just don't understand how Restoration did it exactly and anyway it did blow half the filesystem up in the process. So I guess you are right, can't be done, at least in a civilized way...

 

Point 3: Diskeeper makes hidden dirs and 0 lenght files for padding the MFT. Pretty harmless, and all that gets delete afterwards. Just names remain in MFT. And yeah, I was looking for a way to skip all those names altogether in Recuva so it won't choke and drop me back to the desktop. I stand corrected and grateful for your precisions ^^

 

I'll abuse some more of your time with this matter thaugh. What means do I have to force 5 million entries in MFT, without having to create a file for each entry? I am aware that I can change MFT space reservation with fsutil, but as I copy files on the fs MFT grows in fragments all around the disk... Is there a way to make it grow in 1 block without fragmentation?

Link to comment
Share on other sites

  • Moderators
What means do I have to force 5 million entries in MFT, without having to create a file for each entry?

None that I know of, but I have never even contemplated creating an MFT of that size.

 

Is there a way to make (the MFT) grow in 1 block without fragmentation?

I guess that's in excess of 5 gb. In XP the MFT should be contiguous within the MFT zone, as long as the zone is large enough in the first place and the disk never becomes full enough to force data into the MFT zone. In Vista I believe (from memory) that the MFT is allocated in 200 mb chunks which can go anywhere. M/S says that this is no great problem, but no doubt there are ways of making them contiguous if you wish. There's some M/S info on the Vista MFT which should be easy to find on Google. I have however no experience on a system of your size.

 

P.S. Further to Recuva always doing a full scan, it doesn't appear to scan for undeleted files unless you ask it to. I doubt if this is much help though.

Link to comment
Share on other sites

  • 3 weeks later...

I've used a manual method that seems to work - you simply select the number of files somewaht similar to the number of deleted files in the MFT.

 

Open a CMD prompt and enter:

 

For /L %i in (1,1,5000) do fsutil file createnew recuva%i.tmp 1024

 

change the 5000 for the number of files that you want, the recuva to whatever name you want (xyz.tmp) and 1024 to the size of the file that you want in bytes. Make sure to switch to the directory you want to create the temporary files in - such as c:\temp.

 

I did the above with 80,000 files and Recuva hasn't found any entries for typically deleted files such as pictures, etc.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.