[allen2]

I saw 2 partitions on the same server having the same strange problem:
- explorer and dir report there is low or no space available
- the total data amount when selected use less than the half of the total space (and yes i have tried some of suggestions: http://blogs.msdn.co...free-space.aspx and i have the rights on all files even those in "System Volume Information" ).
- chkdsk report the space "correctly" it does show about the missing space as free in its report at the end.
- lads i used to create a report on the alternate datastream on this report also the space "correctly" and the total size of the alternate datastream file isn't even about the tenth of the missing space.
So if anyone has another idea except the ones needing to umount the partitions (i didn't do the chkdsk /F because of this).
- I also tried to fill the available space with 7zip but it was only able to fill the space seen by windows explorer
- also windows explorer doesn't report the total space correctly it different from the size seen in disk manager.

This post has been edited by allen2: 29 March 2011 - 02:14 PM

[cluberti]

Interesting indeed, although that would sound more like MFT issues than an actual NTFS issue. It might be worth booting from WinPE and doing a chkdsk and defragmenting offline, then rebooting. It might also be worthwhile to increase the MFT space, as MFT fragmentation can be an issue in misreporting data (and apps that read raw disk data would likely be correct, whereas the shell would be wrong). The default value for NtfsMftZoneReservation is 1, and for most systems 2 should be sufficient. Only use 3 or 4 if you plan to store LOTS of files on a volume - otherwise, you'll waste space on the disk and limit the number of files that can be safely stored there. Note that this does work better on a freshly formatted volume, but changing it, rebooting, defragmenting offline, booting into Windows, rebooting, and then defragmenting offline again (online if XP or higher is fine, as defrag can do this online in XP or higher) should make headway into this problem.

If this is an NT4 or 2000 system, however, you're going to have to offline the drive to fix this.

[allen2]

Finally i did the chkdsk /f and it solved the problem after 2 passes (and reboot between).

[allen2]

I will post the real cause and solution of this wierd problem:
- the chkdsk worked only a few days then the problem was there again.
- i decided to backup, format and restore the partition affected to solve this.
In the restoration process, after restoring all datas and all was fine, i restored FSRM quotas and configured shadow copies (to have a few versions of files there). And you know what the problem reappeared (in its low effect: i.e. the partition total size is reported wrongly in explorer and free space isn't much affected).
As i just restored quotas, i saw there was a quota set on the root folder and the quota amount was the same as the total space reported by explorer. So i removed the quota, and the problem wen away.
And after some more research, i found those MS KB: http://support.microsoft.com/kb/944830 and http://support.microsoft.com/kb/977522.
The first clearly show that there was a bug in the ntfs.sys driver and that with many shadow copies the reported space would decrease twice fast. And chkdsk /F would only "fix" the ntfs $bitmap as it would only update it.
I though i'd let anyone encountering the same behavior know the root cause and its dumb solution.