Molly Gilliver wrote:
> In message <jo1thu$6nm$(E-Mail Removed)>, Paul <(E-Mail Removed)> writes
> []
>> I wouldn't have made those two logicals in the first place. That
>> would put me in a hard spot, when it comes to managing space, and facing
>> the situation you're in right now. From bitter experience, I know if
>> I make logicals now, they'll only be a roadblock later, to easy
>> solutions.
>
> Care to elaborate on that? Although I only have two partitions (one for
> the OS and software, the other for my data) on this my main machine, I'm
> curious to know your reasons for disliking multiple partitions.
When you need to move logicals, it isn't very convenient. Your best choice,
might be to make the fourth partition the extended one, and put the logicals in
there. But during your planning phase, you'd better get the sizes right,
or you're in for hours of fun. For example, if you needed to make your
third primary larger, you might need to shrink your logicals, shift them
to the right, shrink the Extended and move its left edge to the right,
until you have an unallocated gap suitable for making the third primary
partition larger. Which is a whole lot of work, with a whole lot of risk.
>
>> Sometimes you have no choice. In the old days, I had particular reasons
>> for having a computer with 20 partitions. But I just don't do stuff like
>> that any more. A lot of the old capacity barriers that caused solutions
>> like that, are gone.
>
> I agree the reasons you _had_ to have many partitions are mostly gone
> (mainly OS and/or motherboard limitations), but I think some people just
> like it as something logical. (Also, possibly, having one for large
> and/or fast-access files - such as video files - might still have
> advantages in some situations, though a physically separate disc would
> be better for most of those.)
> []
I had a certain backup tool, with size limits on backup. And the "20 partition"
machine, allowed me to beat their silly limit, which had no logical reasoning
in the first place. It was just an arbitrary limit in some commercial backup
software I'd bought. Rather than let the bastards win, I "fixed it". And at
the time, 20 partitions was the max allowed. I would have continued with the
silly idea, and added more partitions, if I could have.
>> ftp://ftp.symantec.com/public/englis...s/PTEDIT32.zip
>>
>>
>> Example output from that program, showing a Dell disk.
>>
>> http://www.goodells.net/dellrestore/files/dell-tbl.gif
> []
> That puzzles me. It seems to show a disc with three partitions (well,
> four, but the fourth one is all zeros - I presume that's its way of
> marking the end of the list). But although the first three have
> different "Starting" "Cyl Head Sector" numbers, they all seem to have
> the same "Ending" ones. I'd have expected each one to end at the sector
> before the next one starts?!?
>
> (I also see partition 1 starts at Cyl 0 Head 1 Sector 1, and has 63
> sectors before it, whereas the others all start on a Head 0; I presume
> the first head [0] of 63 sectors is where the partition table itself,
> and/or boot sector, resides, or something like that?)
Once you get past a certain capacity point, the CHS is bogus. Disks are
actually controlled by LBA (logical block address), the numbers down
the right hand side of the PTEDIT32 display. But the CHS is still used
by a lot of OSes, for decision making. For example, Disk Management in
WinXP, plans the offsets and sizes of partitions, quantized to "S". So
if there are 63 sectors per track, then everything on the right hand
side ends up divisible by 63. And that screws up two things. Efficient
operations on an SSD. And efficient operations on a 4KB/sector current
generation hard drive. If the number was 64, the "world would have been
a happy place".
At one time, disks had small capacities, and the CHS was physical. You
might have had say eight heads, some number of fixed sectors per track,
and the cylinder count was real. You would specify operations in terms
of particular C, H, and S values. But once disks got large enough, the
allocation for those fields (field width) ended up too small. So IDE drives
and the BIOS, added support for simple logical block addressing, a single
number specifying what sector you wanted. That's similar to how SCSI works,
which had LBA from the start (a much more reasonable design, but
with added complexity in the controller board strapped to the drive).
Real disks, are actually variable geometry. The disk is "zoned", meaning
the number of sectors per track, varies across the disk. To convert an
LBA, into actual internal geometry on the disk, would need to take into
consideration, how the zones work. A good drive design, numbers all the
sectors as well, so after seek is complete, and the head is on track,
the controller can read the sector headers, and verify its in the right
place to get the LBA numbered sector the user specified. IBM is an example
of a company, that stopped doing that (no double check via sector headers).
*******
That Dell disk has three primary partitions. It's possible PTEDIT32 also
has a notation for logical, but since I just don't use extended/logical
partitions here, I don't know what it does in that situation. Logicals
live in an extended envelope, and so one primary partition would have
a partition type field indicating extended. The GUI on that tool, doesn't
look like it has room to display logicals at all. Just the primary partitions.
PTEDIT32 and Partition Magic in general, are pretty "brittle" when it comes
to "fake CHS geometry" info, alignment to multiples of S, and so on. If
I were to present a Windows 7 disk, with alignment to 1MB chunks instead of
to 63 sectors, the tools would likely error out. Whereas, when PTEDIT32 and
the OS disagree on the fake CHS shorthand, you get warnings every time
those tools are started. Annoying, but not the end of the world. The thing
is, if you "let Partition Magic fix the problem", the problem only comes
back again later, after you make some change with Disk Management.
And if I switch over to one of the 500 Linux distros, in an attempt to do
maintenance work, the tools there have switched to Windows 7 style alignment.
Instead of providing "hobbyist style controls" that could do just
about anything, they instead provided "dumbed down controls", which
means I can't do any serious work with things like GParted, from Linux.
Summary: Things are a mess... The switch from CHS, should have started
a lot sooner. As in, completely ignoring CHS once the disks no longer
used CHS in a practical way. And it's mainly an issue, if you're trying
to maintain your older equipment.
Paul