On Wed, Mar 14, 2001 at 09:42:42AM -0500, D-Man wrote:
...
> It recognized the partitions in the same way whether they were hidden
> or not. They showed up as E: and F:, but other than showing up in "My
> Computer", they did nothing else. I think the properties listed it as
> 0 bytes or something. It wasn't quite correct, but it made sense
> because windows doesn't understand ext2.
I fear you hit here upon a different problem. As I said I read up on
partition types *after* being bugged by serious disk problems. For
one, my father and I had spurious ghost drives in Windows98, and
Windows insisted upon formatting them. I thought I knew what I was
doing (there _was_ an unformatted Windows partition there too) but I
didn't and ended up with a corrupted ext2 partition:( I think what bit
me then is showing up here again, don't know, just guessing, or
hoping, for pray god let win2k/winnt not try to make sense out of
linux partitions!
What bit me was the lousy support for bigdisks in Windows combined
with using Linux fdisk to create the partition table (I had to as
Windows fdisk didn't allow me to create more then one primary windows
partition and I needed that to be able to install a kids-only Windows)
I created an extended partition --an extended partition is an entry in
a partition table pointing to another (nested) partition table-- of
type 5 not knowing that now and again Windows will treat all info in
there as if it sits on a 8Gig tops drive, meaning all cylinders are
below the 1024 cylinder limit. That extended partition covered most
of the disk well above the 1024 cylinder limit, but Windows prefered
to interpreted it as if were all below that limit. So, horror horror,
Windows thought that the sector containing the nested partition table
was somewhere in the lower end of the disc, unfortunately at a place
in use, so it found info there (that is that extended partition table
wasn't zero-ed out:) and by shere bad luck found a description of a
Windows partition, better call it a ghost partition, which ofcourse
didn't excists in reality but was overlayed upon other partitions.
Changing the extended partition-type to 85 (extended Linux) solved it
as now, not knowing what kind of partition it is, Windows doesn't try
to interpreted the location info of this extended partition anymore.
Similar problems are reported on MS-sites, then related to partition
types E and F.
disclaimer: the above is me own thinking, likely to be flawed,
but he, it worked for me:)
Having read some more I now tent to think that it all bowls down to
Linux fdisk not fully adhering to Windows expectations of partition
table format, i.e. when a partition exceeds the 1024 cyl limit, both
the starting and ending fields for head/cyl/sector in the partition
table entry should be maxed out forcing Windows to use the start and
length fields instead.
-- groetjes, carel
