Member Avatar for Ant70

The disk in my laptop has become inaccessible after I changed the cluster size.

The problem

When I try to boot up, I get the laptop's proprietary screen as normal followed by a few seconds of the Windows XP start up screen followed by a brief flash of a blue screen, which begins: "A problem has been detected and Windows has been shut down to pretect your computer. If this is the first time..."

When I put the disk in a caddy, Windows XP reports that the drive "is not accessible. The file or directory is corrupt and unreadable."

What happened

I used Partition Magic (version 8.0) to change the cluster size (from 512kB to 2kB), running it from disk rather than the desktop. PM reported the process complete but also reported error #728 (with no other information). The laptop then wouldn't start up.

If I run PM from disk, it correctly reports the disk size as well as used and unused space. It also reports the following errors.

Error #1657: MFT mirror data position mismatch
Error #501: Cross-linked files were found
Error #745: [No message]

What I've tried

I've run fixmbr and fixboot in Recovery Console (from disk). When I type "dir" at the C:\ command prompt in Recovery Console, it gives: "An error occurred during directory enumeration."

Can anyone advise me how I can get this disk up and running again, or at least recover data from it? Help would be much appreciated!

The disk was running Windows XP (SP3). The file system is NTFS.

  • 3 Contributors
  • 5 Replies
  • 368 Views
  • 6 Days Discussion Span
  • Latest Post Latest Post by rubberman

Dani AI

Generated

Short, practical summary and next steps (safe order of operations):

The drive shows classic NTFS metadata corruption after an in-place cluster-size change. The highest-priority rule now is: stop modifying the original disk and make a full sector‑by‑sector image to another healthy drive. Working from that image preserves recovery options and is what data-recovery pros always do.

Suggested recovery workflow (do these in order)

  1. Image the disk (do not write to it). Use a sector copier that produces a log so you can retry/resume on read errors (examples: GNU ddrescue, Clonezilla, or a Windows raw-copy tool). Example ddrescue syntax (Linux):
    ddrescue -f -n /dev/sdX /path/to/image.img /path/to/image.log
  2. Work only on the image. Try non‑destructive metadata repair first: TestDisk can scan for NTFS boot sector/partition issues and attempt to rebuild the NTFS structure from the MFT/mirror. If that fails, use file-recovery tools that operate on images (R‑Studio, GetDataBack, DMDE) to recover files and folders.
  3. If metadata is too damaged, use file-carvers (PhotoRec, etc.) on the image to salvage file contents — filenames and folders may be lost but most common formats can be recovered.
  4. If the drive has physical problems or the above fails, consider a professional recovery service.

Important cautions and notes

  • Do not run repair utilities on the original disk (especially anything that writes automatically) until an image exists. Automated repairs (including chkdsk or similar) can overwrite metadata and reduce recoverability.
  • ’s pointer toward partition-recovery utilities is on the right track — use those on an image, not the source disk. is also correct that a full OS reinstall is often required to restore a bootable system; that’s a separate step after data is safe.
  • Expect that the system partition may need reformat/reinstall even if data is recovered. Prioritize data extraction before attempting to rebuild the OS.

If an image has already been made, note it and proceed with TestDisk/DMDE on the copy; if not, imaging is the immediate next action.

Recommended Answers

All 5 Replies

Member Avatar for Ant70

Anyone?

Member Avatar for caperjack

guess not ,someone come now and correct me ,and help you at the same time
im no expert but im sure its because [i dont thing changing cluster size on a active hdd is somethink that shoud/can be done ] you change the cluster size, is somethin that reccomended ,but i could be wrong ,and maybe i should do some re-search on it to see .
quick research sows people doing it so i quess its possible ,mind you everything i read is people having problems with it

Edited by caperjack
Member Avatar for Ant70

Thanks for posting, caperjack. I'll take a look into the findpart utility mentioned in the thread you linked to, and have a more careful read of the thread.

Yes, I recall reading a forum thread in which someone reported that they'd been told by Powerquest (makers of PM) that the tool shouldn't be used on the system partition. That's not good news for anyone with an OEM system installation on a low cluster size.

Member Avatar for caperjack

i thought there might be a performance issue as well so i googled and found a lot of info of course ,like this link about performance issue ,as it says with the newer bigger drive of today it hardly an issue anymore
slower

Member Avatar for rubberman

If you change the cluster size on an hdd, then you will need to reinstall the os and everything else! Not a good idea if you want to keep your system intact! I think you are fubar until you re-install the OS and all of your data and applications (have backups, do we?)...

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.