1 |
On Sunday 25 February 2007 14:05, JC D wrote: |
2 |
> Hi! |
3 |
> |
4 |
> I think I have a serious problem with my partition table. I did all the |
5 |
> partitions with fdisk but may be not right, or somewhen screwed my |
6 |
> partition table up. I want to use a graphical tool like gparted also doing |
7 |
> some changes on my fat32 partition. But when I installed gparted I got an |
8 |
> empty partition table with nothing than |
9 |
> |
10 |
> unallocated 55.89 GB (my disk is an 60 GB Hitachi) |
11 |
> |
12 |
> nothing else!!! |
13 |
|
14 |
Did you write/save the partition table after your changes? Did you reboot |
15 |
thereafter? I make some suggestions below but they come with the health |
16 |
warning attached that I have no responsibility if it doesn't work. |
17 |
|
18 |
> Here is my outpu from |
19 |
> # fdisk -l |
20 |
> |
21 |
> fdisk -l |
22 |
> |
23 |
> Disk /dev/hda: 60.0 GB, 60011642880 bytes |
24 |
> 255 heads, 63 sectors/track, 7296 cylinders |
25 |
> Units = cylinders of 16065 * 512 = 8225280 bytes |
26 |
> |
27 |
> Device Boot Start End Blocks Id System |
28 |
> /dev/hda1 * 1 1913 15361888+ 7 HPFS/NTFS |
29 |
> Partition 1 does not end on cylinder boundary. |
30 |
|
31 |
This is in itself is not a problem. If you create/resize partitions based on |
32 |
MB you may or may not hit a clean cylinder boundary. Do your resizing in |
33 |
cylinders and the end result will coincide with their boundaries. |
34 |
|
35 |
> /dev/hda2 1914 4208 18427027+ c W95 FAT32 (LBA) |
36 |
> Partition 2 does not end on cylinder boundary. |
37 |
> /dev/hda3 5100 7296 17637480 f W95 Ext'd (LBA) |
38 |
> Partition 3 does not end on cylinder boundary. |
39 |
|
40 |
Look what has happened here, your extended partition starts at cyl 5100, while |
41 |
your first logical partition starts at cyl 4208. I'd say that neither are |
42 |
correct. I suggest that you resize your extended partition hda3 to start at |
43 |
cyl 4209, i.e. straight after the end of the previous primary partition |
44 |
(hda2). |
45 |
|
46 |
> /dev/hda4 4208 5099 7164990 83 Linux |
47 |
|
48 |
Here you need to change the beginning of this partition to be straight after |
49 |
the end of the previous primary partition and at the beginning of the |
50 |
extended partition. So, hda4 should start at cyl 4209. |
51 |
|
52 |
> /dev/hda5 * 5100 5108 68008+ 83 Linux |
53 |
|
54 |
This is good, but you do not need the boot flag. As a matter of fact it may |
55 |
confuse your WinXP bootloader and Linux does not need it. |
56 |
|
57 |
> /dev/hda6 5108 5173 521608+ 82 Linux swap / |
58 |
|
59 |
This is not good as it overlaps the end of the previous partition. hda6 |
60 |
should start at cyl 5109. |
61 |
|
62 |
> Solaris /dev/hda7 5173 7296 17047768+ 83 Linux |
63 |
> |
64 |
> Partition table entries are not in disk order |
65 |
|
66 |
That's because they were created at a chronological sequence which do not |
67 |
reflect their physical order on the disk. Not really important. |
68 |
|
69 |
> This seems wrong to me. Also parted says something with overlapping |
70 |
> partitions and its right I think. Is there any chance to repair this |
71 |
> without destroying the whole installation? I have backuped my installation |
72 |
> with rsync and also did save my mbr: |
73 |
> |
74 |
> dd if=/dev/hda of=/backup/mbr512.img bs=512 count=1 |
75 |
> |
76 |
> and |
77 |
> |
78 |
> dd if=/dev/hda of=/backup/mbr446.img bs=446 count=1 |
79 |
|
80 |
Since you have a backup the worst that can happen is to correct the partitions |
81 |
as suggested above and discover that your OS cannot read them! In that case |
82 |
reformat them and use your back up to restore the data. |
83 |
|
84 |
An alternative approach is to restore your partition table to a previous |
85 |
version and with it the previous boundaries of your partitions. After I was |
86 |
faced with a borked partition table too, I used testdisk to recover and |
87 |
restore previous partition table entries. |
88 |
|
89 |
http://www.cgsecurity.org/wiki/TestDisk |
90 |
|
91 |
HTH. |
92 |
-- |
93 |
Regards, |
94 |
Mick |