Mail Archives: opendos/1999/03/14/20:41:29
If anyone knows of a better place to ask these questions, with relatively
quick, appropriate replies, let me know. I ask here just because I've
been subscribed for a couple of years, and I know there are some very
knowledgable people here. (And because I think DR-Dos's fdisk is to
blame; but I could be wrong.)
So...
Things are now looking even worse than I had feared.
I was finally able to get the fdisk option to write a standard dr-dos
master boot record. I did so, and all seemed to be going "fine." It
was now booting from my hard drive, albeit from the *new* bootable
partition, and still not recognizing two of my dos drives. But that was
progress, right?
Then I was able to get into Linux via the RH CD Rom (god, that
installation program is smart; I didn't even have to reinstall a single
byte...it let me create a boot disk, and put me right in). Here's where
the strangeness soaked in.
Cfdisk reported a fatal error--corrupt partition.
So then I rewrote the "original" MBR from dos's fdisk--which put it back
into the "no message, just hang" state, but still let me get into Linux
from my new wonderful more-beautiful-than-a-redhead boot disk.
I tried cfdisk again. Same problem. So I tried Linux's fdisk. And here
is what it reports.
NOTICE that it doesn't report anything for /dev/hda1, /dev/hda2,
/dev/hda6, or /dev/hda7 (the latter two are understandable considering the
contents of /dev/hda5--the first two are only understandable when you look
at my previous message and recall that Dos's fdisk has refused to
acknowledge their existence since before this mess started):
Disk /dev/hda: 128 heads, 63 sectors, 523 cylinders
Units = cylinders of 8064 * 512 bytes
Device Boot Start End Blocks Id System
/dev/hda3 * 1 166 669280+ 6 DOS 16-bit >=32M
Partition 3 has different physical/logical beginnings (non-Linux?):
phys=(133, 1, 1) logical=(0, 1, 1)
Partition 3 has different physical/logical endings:
phys=(298, 127, 63) logical=(165, 127, 63)
/dev/hda4 167 178 48384 5 Extended
Partition 4 has different physical/logical beginnings (non-Linux?):
phys=(299, 0, 1) logical=(166, 0, 1)
Partition 4 has different physical/logical endings:
phys=(310, 127, 63) logical=(177, 127, 63)
/dev/hda5 ? 349621 412506253551826+ 49 Unknown
It seems that the logical reportings might be correct (only the first
number differs in each case, and the logical numbers are ordered).
Unfortunately, there are only three "valid" partition entries with which
to work, so even if the logical numbers are correct they won't restore my
partition. Correct?
Fdisk did report that the errors (or some errors?) would be corrected with
the 'w'rite option. But I've learned not to trust software. :)
Is there any way to recover my partition table? Is there any
software--however expensive, at this point :)--that could reconstruct a
brand new partition table, without the loss of data?
Actually, it would seem easy enough to edit the disk directly, and recover
the starting points of the FAT tables on my dos drives, and thus recover
the data on those drives. It's been a while since I've looking at NTFS,
so I don't know if anything similar could be said for those.
Unfortunately, I don't have the original starting and ending positions
written down; all I have are the order of the positions, and the
approximated size as reported to two decimal places by Linux Cfdisk. If I
did have this information, would recreating the partition table work?
Could it still work now, knowing just the sizes (and the fact that
/dev/hda3 and /dev/hda4 *apparently* are OKAY, at least according to
Dr-Dos, which lets me read and write data, without any noticeable
problems).
Is it possible that all of this is caused by just ONE original byte being
wrong? Is it possible that just editing the partition table directly,
andmaking one or two changes, will make everything all right?
Sigh. And I always laughed at people for not backing up their stuff. If
I lose my primary hard drive, well, I'll have lost most of my reason for
living. Or something.
--
Michael Jones
- Raw text -