K
KevinGPO
Ok. A lot of these fabulous windows partitioning tools are paid-only. Plus,
most of them only recover files by copying them onto another hard disk. I
don't have a spare 120GB hard disk so most programs are not feasible.
I managed to restore my 10GB primary FAT 32 windows XP partition - using
Knoppix / TestDisk. However it was not able to find anything else.
While still in Knoppix, I used gpart and it found all my partitions. Wow!
Brilliant. Wrote the partitioning tables. Rebooted back into Knoppix/linux
and used:
cat /dev/hda5 | less
to double-check that it starts with the NTFS header. Brilliant. Both my 50GB
NTFS partitions seem to be displaying the correct information.
Now, the final test. Boot into windows XP. The two hard disks D: and E: are
there. But they are inaccessible. No disk size information and it says it
needs to be formated. Hmm...
gpart actually founded ALL the partitions as primary partitions. Both the
50GB NTFS partitions were actually logical partitions inside a extended
partition container. Could this be why windows can't read them because they
are logical partitions within an extended partition?
So I went back into Knoppix and used fdisk to manually delete the 2 NTFS
partitions. I created a extended partition. Then I created the logical
partitions. (after a couple hours of hard work playing around with the
offsets...) Somehow NTFS partition 1 start offset changed by 5 sectors
(512bytes x 5). While NTFS partition 2 start offset changed by 252sectors. I
find out that NTFS partition 1 ending offset ends 64 sectors before the
start of NTFS partition 2.
Anyway, I got my partitions back in Windows XP now. What am worried about is
whether NTFS partition 2 ending offset if correct... since I just used the
default value of the end of the hard disk. i remember Windows XP has some
funny thing about leaving 8MB space at the end of the hard disk. I don't
know.
What I want now is a program to verify and double-check my partitioning
tables correct and intact. I wished gpart could detect the extended
partition but it didn't. Could it be intelligent enough now to see it and
verify the logical partitions within?
most of them only recover files by copying them onto another hard disk. I
don't have a spare 120GB hard disk so most programs are not feasible.
I managed to restore my 10GB primary FAT 32 windows XP partition - using
Knoppix / TestDisk. However it was not able to find anything else.
While still in Knoppix, I used gpart and it found all my partitions. Wow!
Brilliant. Wrote the partitioning tables. Rebooted back into Knoppix/linux
and used:
cat /dev/hda5 | less
to double-check that it starts with the NTFS header. Brilliant. Both my 50GB
NTFS partitions seem to be displaying the correct information.
Now, the final test. Boot into windows XP. The two hard disks D: and E: are
there. But they are inaccessible. No disk size information and it says it
needs to be formated. Hmm...
gpart actually founded ALL the partitions as primary partitions. Both the
50GB NTFS partitions were actually logical partitions inside a extended
partition container. Could this be why windows can't read them because they
are logical partitions within an extended partition?
So I went back into Knoppix and used fdisk to manually delete the 2 NTFS
partitions. I created a extended partition. Then I created the logical
partitions. (after a couple hours of hard work playing around with the
offsets...) Somehow NTFS partition 1 start offset changed by 5 sectors
(512bytes x 5). While NTFS partition 2 start offset changed by 252sectors. I
find out that NTFS partition 1 ending offset ends 64 sectors before the
start of NTFS partition 2.
Anyway, I got my partitions back in Windows XP now. What am worried about is
whether NTFS partition 2 ending offset if correct... since I just used the
default value of the end of the hard disk. i remember Windows XP has some
funny thing about leaving 8MB space at the end of the hard disk. I don't
know.
What I want now is a program to verify and double-check my partitioning
tables correct and intact. I wished gpart could detect the extended
partition but it didn't. Could it be intelligent enough now to see it and
verify the logical partitions within?