Can it be said that a locical drive is a virtual drive?

"Tom" wrote:

"Timothy Daniels" wrote:
"Tom" wrote:
"Timothy Daniels" wrote:
"R. C. White" wrote:
One big problem in discussing this is the several ambiguous
meanings of some terms, such as "drive", "partition" - and
"boot". :(

As many writers have commented, "We BOOT from the
SYSTEM partition and keep the operating SYSTEM files in
the BOOT volume." Microsoft didn't invent this terminology,
but continues to use it.

The boot process must begin in the System Partition, which
must be on the boot device (typically the master HD on the
primary IDE controller); this must be the Active (bootable) partition,
which means it must be a Primary Partition. This partition must
have the proper NT-style Boot Sector, and 3 files (NTLDR,
NTDETECT.COM and Boot.ini) must be in the Root of that partition. (In
some installations, a few other files, such as
NTBOOTDD.SYS, are required, but these are not typical.)

All the rest of Windows (all recent versions, at least) must be
in the Boot Folder in the Boot Volume. The Boot Folder is
named \Windows, by default. (Except in WinNT and Win2K,
where the default is \WinNT; in an upgrade installation, the
new Windows installation inherits the name of the earlier version, so
some WinXP boot folders are named \WinNT.)
This Boot Folder - and its many subfolders - hold the GB or
so of WinXP files.

The Boot Volume can be any volume, either a primary partition
or a logical drive, on any HD in the computer. (Maybe it could
also be on an external HD or a rewritable CD/DVD - or even a
very large USB flash drive, but I don't know about that.) The
Boot Volume may share the System Partition - and typically does.
There should be a separate Boot Volume for each installation
of Windows; Microsoft (and nearly all other gurus) strongly advise
against putting two Windows installations into a single volume.

[............]
Logical drives are not bootable. A dual-boot must start in the
System Partition (which must be a Primary Partition). But
C:\Boot.ini will point to the Boot Folder, which may very well be
in a logical drive. (In my own system, C: is a small FAT-formatted
primary partition on my IBM SCSI HD; my main WinXP is in
D:\Windows, an NTFS-formatted logical drive in the extended
partition on that SCSI HD; other Windows installations are
in F:, L:, and X:, all logical drives in extended partitions on my
two IDE HDs and all NTFS. My boot process starts with
C:\NTLDR and then branches to D:\Windows - or wherever.)

A logical drive can be a "Boot Volume" - but you can't boot from it.
:(


Yes. You are totally correct. Significant points are that the "boot
files"
are in the SYSTEM volume (partition), and the Operating System is in
the BOOT volume (partition), and the two volumes (partitions) need not
be the same or even on the same hard drive. The SYSTEM partition
must be a Primary partition and it must be flagged as "active". The
BOOT partition (containing the OS), can be a Primary partition or a
Logical Drive in an Extended partition.

*TimDaniels*

Note, that any operating system installed in multiple boot scenarios are
all "active", since the boot process to load them (in the
extended/logical) are on the Primary partition, which has to have the
bootloader at the beginning of it.


If by "boot loader" you mean ntldr, it must be in the "active" Primary
partition of the hard drive that is at the head of the BIOS's HD boot
sequence, but the OSes that it loads may be from any of other
Primary partitions and logical drives in the system, none of which
even have to be on the same hard drive as ntldr.

*TimDaniels*

That isn't what I mean. All bootloaders are on any given Primary drive; you
can have 3 OSes on 3 primary drives, and they will have their own
bootloaders, but then you'll need a boot manager to get into those OSes.
What I said was, that any primary/logical setup, whereas you have e.g. three
OSes installed on them (1st OS primary, 2 and 3 on the extend/logicals),
they are all by default "active". This because the bootloaders for all the
OSes reside on the "active" (Primary) partition. This is why one has a
choice to what OS they can boot into when they get to that option after the
BIOS post.


It's unclear, really, what you mean. I can understand my grammar OK,
but yours leaves me baffled. What I wrote is right out of the Microsoft
Windows XP Resource book, and I have verified what I wrote in the above
paragraph. If you have some issue with it, please state what you disagree
with. As for your "bootloaders", I must still assume that you mean ntldr,
which can load an OS from any partition on any hard drive in the system.
For those OSes that reside in a logical drive (i.e. a partition within an
Extended partition), there is no ntldr resident in the same partition. For
those OSes that reside in a Primary partition, there need not be an ntldr
there, either, if it is not an "active" partition, as the OS can be loaded by
an ntldr in another partition that *is* active. Put another way, the OS that
is loaded can be loaded by an ntldr that is anywhere in the system. All it
needs is a co-resident boot.ini file that points to the partition that does
contain the OS.

*TimDaniels*

"Timothy Daniels" wrote in message
...
"Tom" wrote:

"Timothy Daniels" wrote:
"Tom" wrote:
"Timothy Daniels" wrote:
"R. C. White" wrote:
One big problem in discussing this is the several ambiguous
meanings of some terms, such as "drive", "partition" - and
"boot". :(

As many writers have commented, "We BOOT from the
SYSTEM partition and keep the operating SYSTEM files in
the BOOT volume." Microsoft didn't invent this terminology,
but continues to use it.

The boot process must begin in the System Partition, which
must be on the boot device (typically the master HD on the
primary IDE controller); this must be the Active (bootable)
partition,
which means it must be a Primary Partition. This partition must
have the proper NT-style Boot Sector, and 3 files (NTLDR,
NTDETECT.COM and Boot.ini) must be in the Root of that partition.
(In some installations, a few other files, such as
NTBOOTDD.SYS, are required, but these are not typical.)

All the rest of Windows (all recent versions, at least) must be
in the Boot Folder in the Boot Volume. The Boot Folder is
named \Windows, by default. (Except in WinNT and Win2K,
where the default is \WinNT; in an upgrade installation, the
new Windows installation inherits the name of the earlier version, so
some WinXP boot folders are named \WinNT.)
This Boot Folder - and its many subfolders - hold the GB or
so of WinXP files.

The Boot Volume can be any volume, either a primary partition
or a logical drive, on any HD in the computer. (Maybe it could
also be on an external HD or a rewritable CD/DVD - or even a
very large USB flash drive, but I don't know about that.) The
Boot Volume may share the System Partition - and typically does.
There should be a separate Boot Volume for each installation
of Windows; Microsoft (and nearly all other gurus) strongly advise
against putting two Windows installations into a single volume.

[............]
Logical drives are not bootable. A dual-boot must start in the
System Partition (which must be a Primary Partition). But
C:\Boot.ini will point to the Boot Folder, which may very well be
in a logical drive. (In my own system, C: is a small FAT-formatted
primary partition on my IBM SCSI HD; my main WinXP is in
D:\Windows, an NTFS-formatted logical drive in the extended
partition on that SCSI HD; other Windows installations are
in F:, L:, and X:, all logical drives in extended partitions on my
two IDE HDs and all NTFS. My boot process starts with
C:\NTLDR and then branches to D:\Windows - or wherever.)

A logical drive can be a "Boot Volume" - but you can't boot from it.
:(


Yes. You are totally correct. Significant points are that the
"boot files"
are in the SYSTEM volume (partition), and the Operating System is in
the BOOT volume (partition), and the two volumes (partitions) need not
be the same or even on the same hard drive. The SYSTEM partition
must be a Primary partition and it must be flagged as "active". The
BOOT partition (containing the OS), can be a Primary partition or a
Logical Drive in an Extended partition.

*TimDaniels*

Note, that any operating system installed in multiple boot scenarios
are all "active", since the boot process to load them (in the
extended/logical) are on the Primary partition, which has to have the
bootloader at the beginning of it.


If by "boot loader" you mean ntldr, it must be in the "active"
Primary
partition of the hard drive that is at the head of the BIOS's HD boot
sequence, but the OSes that it loads may be from any of other
Primary partitions and logical drives in the system, none of which
even have to be on the same hard drive as ntldr.

*TimDaniels*

That isn't what I mean. All bootloaders are on any given Primary drive;
you can have 3 OSes on 3 primary drives, and they will have their own
bootloaders, but then you'll need a boot manager to get into those OSes.
What I said was, that any primary/logical setup, whereas you have e.g.
three OSes installed on them (1st OS primary, 2 and 3 on the
extend/logicals), they are all by default "active". This because the
bootloaders for all the OSes reside on the "active" (Primary) partition.
This is why one has a choice to what OS they can boot into when they get
to that option after the BIOS post.


It's unclear, really, what you mean. I can understand my grammar OK,
but yours leaves me baffled. What I wrote is right out of the Microsoft
Windows XP Resource book, and I have verified what I wrote in the above
paragraph. If you have some issue with it, please state what you disagree
with. As for your "bootloaders", I must still assume that you mean ntldr,
which can load an OS from any partition on any hard drive in the system.
For those OSes that reside in a logical drive (i.e. a partition within an
Extended partition), there is no ntldr resident in the same partition.
For
those OSes that reside in a Primary partition, there need not be an ntldr
there, either, if it is not an "active" partition, as the OS can be loaded
by
an ntldr in another partition that *is* active. Put another way, the OS
that
is loaded can be loaded by an ntldr that is anywhere in the system. All
it
needs is a co-resident boot.ini file that points to the partition that
does
contain the OS.


My explanation is perfectly fine, you have have a problem with semantics as
I see it. NTDLR and the lingo "bootloader" term are very much the same
thing, just different words. After that, you just wnet on and stated IOWs
what I said about other OSes residing on the primary partition, that has
extened/logical drives with other OSes on them. Since the (I'll use this
term for you) NTLDR in on the Primary along with the boot.ini, and NTDETECT,
it is the active partition, that can load all the other OSes from it (if
they are installed as such); hence the other OSes are also active as well
(not on their own primary partitions); does this explain it better? That is
the point I was making sans you stating a thread before that the NTLDR (or
bootloader) has to be on the primary partition, which I stated well before
these most recent posts.

Note, I never said, nor did I even hint, and the NTLDR can reside or resides
in an extended/logical drive(s), and where you felt the need to explain this
further, maybe is a hint that grammar isn't the issue, rather reading
comprehension, or asking for clairification before answering inline, while
ultimately showing misundertstandings. If you read way back, you will see my
statement regarding the bootloaders having to reside on the Primary
partition(s).

On Fri, 25 Mar 2005 20:11:00 -0500, "Tom" wrote:
"cquirke (MVP Windows shell/user)" wrote
On Fri, 25 Mar 2005 08:49:12 -0500, "Tom" wrote:

Logical drives are bootable, or else dual-boot scenarios would not be
possible.

False - the bulk of an OS's code may reside on a logical volume, but
standard system MBR code does not transfer control to logical volumes
within an extended partition. There has to be bridging boot code on a
primary partition to do this; an OS in a logical is booted from there.

As I noted to RC, read the whole post, and not take out context what I said.

I share RC's love of accurate semantics, especially in the area of
partitioning. If you think of English is a content description
language, then the value of accurate coding becomes apparent.

When I say bootable in this case, I mean the OS can be booted to from the
list after the BIOS post. I won't requeest any further that you should
include the whole reading the next time :-).

The "list after the BIOS POST" is already post-boot. You have already
booted from the active primary partition; all you are now doing is
selecting what to load next.

Logical drives are not bootable, and do not need to be bootable to
implement dual boot scenarios, including those that are based on the
C:\NTLDR - C:\Boot.ini axis that I believe you are referring to when
you mention the "list after the BIOS POST".

It's on that basis that I labeled your assertion as false :-)

Now let's see the rest of the post you are referring to... I'd still
say your error is considering the continuation of loading an OS as
"booting" it. An OS is "booted" when control first enters the OS code
from system code, and that's mandated by system code.

Standard system (MBR) code mandates the partition table limitations
mentioned earlier; 4 partitions, only primary partitions bootable.

You may not see anything on screen before you see the "list after the
BIOS POST", but by then, one of two things has already happened:

1) Transfer from system code to OS code (C:\NTLDR - C:\Boot.ini)
2) Transfer to non-standard system code (i.e. add-on boot manager)

It really goes about the semantic detail on what it means to "boot"
(as opposed to "load") an OS.



---------- ----- ---- --- -- - - - -
Gone to bloggery: http://cquirke.blogspot.com
---------- ----- ---- --- -- - - - -

On Fri, 25 Mar 2005 12:47:33 -0500, "Tom" wrote:
"cquirke (MVP Windows shell/user)" wrote
On Thu, 24 Mar 2005 04:05:03 -0800, BAR

FAT only addresses up to 4Gb of disk space [Windows XP, 95 and earlier
Windows versions only]

Detail:
- NT can support 4G FAT16 volumes with 64k clusters
- Win95SR2 thru WinME support only up to 2G FAT16 with 32k clusters

Actually, clarif on "earlier Windows versions" is needed...

Versions that support FAT32:
- XP
- Win2000
- WinME
- Win98xx
- Win95 SR2.x

Versions that do not support FAT32:
- NT 4.0 and earlier
- Win95 SP1
- Win95 original
- stand-alone versions of MS-DOS
- Win3.yuk and earlier

FAT32 supports HDs and volumes well over 32G ...
All OSs that support FAT32, support FAT32 32G.

It's just that XP's volume formatter is not only too lame to format
FAT32 volumes over 32G, but is even too lame to realize its
limitations and avoid trying. Instead, it grinds along until it hits
32G, then it falls over because the volume is "too big".

This is not a FAT32 issue; it is an XP quality failure problem.

While that may be your opinion of lameness,

No, I think most folks would consider it lame, to have code
destructively start an operation, tie up the machine for several
minutes, then fail without the ability to undo the mess.

using a partition size on FAT32 on anything bigger that 16gs, is
a waste anyway IMHO, besides the facts of the file size limits FAT32
has. In todays larger drives that are available, it is a waste to use it.

I'd disagree there, though like you I would not use FAT32 for one big
C: drive on modern large HDs. Even if I were to use NTFS, I'd want to
avoid setting up a modern large HD as one big C:

I consider that function made purposefully for the
sole reason of wasting space and performance.

If that were the case, it would back out as soon as it sees the space
it is asked to format is 32G, without starting to write to it (and
thus mess up anything that might have been there).

The way it, smells like really poor programming, rather than any
particular deliberate by-design limitation. Unless there's a
deliberate intention to imply FAT32 flakiness and FUD by creating the
impression this is a file system problem, rather than bad XP code?

NTFS - addresses up to 2,000Gb of disk space [Windows XP]

AFAIK 2TB is the max limit for FAT32 as well.

Correct for both filing systems, for drive support that is, but:

The maximum possible number of clusters on a volume using the FAT32 file
system is 268,435,445. With a maximum of 32 KB per cluster with space for
the file allocation table (FAT), this equates to a maximum disk size of
approximately 8 terabytes (TB).

NTFS can be 16 exabytes.

Thanks - right now, even 8T is beyond the shopping trolley, and
hopefully by the time HDs exceed that capacity, we'd have decent
maintenance tools for NTFS so that there'd be no need to retain FATxx.



---------- ----- ---- --- -- - - - -
Gone to bloggery: http://cquirke.blogspot.com
---------- ----- ---- --- -- - - - -

On Fri, 25 Mar 2005 14:55:48 -0500, "Anna" wrote:
"cquirke (MVP Windows shell/user)" wrote

It's just that XP's volume formatter is not only too lame to format
FAT32 volumes over 32G, but is even too lame to realize its
limitations and avoid (tr)ying. Instead, it starts to format the volume
(losing any data that was there), and grinds along until it hits 32G,
then it falls over because the volume is "too big".

This is not a FAT32 issue; it is an XP quality failure problem.

I'm aware, of course, as I'm sure most of us are, that XP will not format a
FAT32 volume 32 G.

What annoys me is where those who should know better, present this as
if it was a limitation of the FAT32 file system. It isn't.

But I have never run into the situation where, through XP's Disk
Management utility, it begins formatting a partition than 32 GB
in FAT32 and then "grinds along until it hits 32G, then it falls over
because the volume is "too big"."

I have...

In every instance I've encountered, XP will not provide a FAT32
formatting option if the partition is 32 G. And to the best of my
knowledge this same restriction is present using the XP installation
CD. Have you actually experienced the situation you describe?

Yes, I have. I haven't re-tested it in subsequent SP1 or SP2
incarnations of XP, so maybe it's fixed, and I can't recall the exact
method details - botching large chunks of HDs isn't the sort of thing
one wants to test on a regular basis :-)

It was early in my experience with XP (i.e. within the first month of
the original release). I might have bounced off initial attempts to
format volumes over 32G as FAT32 (e.g. after noting no FAT32 uption
offered via drop-down) and tried some less obvious UI approaches to
get the job done, such as typing commands via Start, Run. I found
something that started the process, and fails as described.

It was definately native XP tools that I was using, and it wasn't a
matter of deep hacks either. Any extra effort required was shallow
enough that I'd forgotten all about it until now!



---------- ----- ---- --- -- - - - -
Gone to bloggery: http://cquirke.blogspot.com
---------- ----- ---- --- -- - - - -

"cquirke (MVP Windows shell/user)" wrote:
The "list after the BIOS POST" is already post-boot. You have
already booted from the active primary partition; all you are
now doing is selecting what to load next.

[........]
... I'd still say your error is considering the continuation of
loading an OS as "booting" it. An OS is "booted" when
control first enters the OS code from system code, and that's
mandated by system code.

Welcome, home, brother! Accurate terminology!
Too many people nowadays use "boot" to mean "load".
If you can remember when a "boot instruction" was a single
machine instruction set on physical switches by hand and
a "boot loader" was contained on a single punched card,
you automatically know the difference. :-)

But what would you call ntldr? It's not in the C:\WINDOWS
folder and thus not part of the OS, and it's not part of a
logical drive - which can contain an OS but the OS has to be
loaded by an ntldr contained in another partition. I've seen
it called "boot manager". But by your definition, it's not
part of the "boot" process. Is it right to call it just a "loader"?

*TimDaniels*

"Leythos" wrtote:
Timothy Daniels wrote:

If you can remember when a "boot instruction" was a single
machine instruction set on physical switches by hand and
a "boot loader" was contained on a single punched card,
you automatically know the difference. :-)

Actually you had to load the boot strap code in on switches and then
select a starting address to run it, that was before punch cards


Actually, I'm embarrassed to admit that I can remember punched
cards! :-)

*TimDaniels*

On Sun, 27 Mar 2005 09:59:37 -0800, "Timothy Daniels"
"cquirke (MVP Windows shell/user)" wrote:

The "list after the BIOS POST" is already post-boot. You have
already booted from the active primary partition; all you are
now doing is selecting what to load next.

... I'd still say your error is considering the continuation of
loading an OS as "booting" it. An OS is "booted" when
control first enters the OS code from system code, and that's
mandated by system code.

Welcome, home, brother!

Hey thanks!

Too many people nowadays use "boot" to mean "load".

Or "download" to mean "install" ... gibber

If you can remember when a "boot instruction" was a single
machine instruction set on physical switches by hand and
a "boot loader" was contained on a single punched card,
you automatically know the difference. :-)

I can't - I was in a different faculty in those days, so while my
engineering flatmates were going "ERROR EABT heh heh" to each other, I
was headscratching though orgasmic chemistry.

But what would you call ntldr? It's not in the C:\WINDOWS
folder and thus not part of the OS

Oh, it's part of the OS, alright. The OS starts in the first sector
of the active primary partition - that's where the system hands off to
the OS's pre-filesystem code. NTLDR is to NT what IO.SYS is to DOS
and Win9x; it's the first code file to be loaded for the OS.

What makes it a bit confusing is that one can dip into NT this way,
and then immediately jump out again to load a different OS instead -
either DOS mode, a Win9x, or Recovery Console. Those OSs are loaded
in the same way; NTLDR loads a partition boot sector code image as if
that had been in effect as the partition code, and jumps into that.

But by your definition, it's not part of the "boot" process.
Is it right to call it just a "loader"?

No, it's part of NT (NT, Win2000, XP). It's just that the OS you
start loading isn't always the OS you finish loading ;-)

It's like when you F8 you way into Win9x's boot menu, and load DOS
mode (Command Prompt Only) or even MSDOS (Previous Version of MSDOS)
instead of Win95/98. The IO.SYS that starts the boot process for
these OSs is common to both DOS mode and Win9x, but can change gears
and load an older MSDOS instead. Mind you, if you do that in
Win95SR2, you'll struggle to get back!

In the interests of accuracy, I have to mention that Recovery Console
isn't an OS, in that it cannot host programs other than itself. It
feels a bit like the MS-DOS 4.00 incarnation of DOS Shell; one hopes
it is the precursor of better things to come in future OS versions.



---------- ----- ---- --- -- - - - -
Gone to bloggery: http://cquirke.blogspot.com
---------- ----- ---- --- -- - - - -

"cquirke (MVP Windows shell/user)" wrote:
"Timothy Daniels" wrote:

But what would you call ntldr? It's not in the C:\WINDOWS
folder and thus not part of the OS

Oh, it's part of the OS, alright. The OS starts in the first sector
of the active primary partition - that's where the system hands off to
the OS's pre-filesystem code. NTLDR is to NT what IO.SYS is to DOS
and Win9x; it's the first code file to be loaded for the OS.

What makes it a bit confusing is that one can dip into NT this way,
and then immediately jump out again to load a different OS instead -
either DOS mode, a Win9x, or Recovery Console. Those OSs are loaded
in the same way; NTLDR loads a partition boot sector code image as if
that had been in effect as the partition code, and jumps into that.

But by your definition, it's not part of the "boot" process.
Is it right to call it just a "loader"?

No, it's part of NT (NT, Win2000, XP). It's just that the OS you
start loading isn't always the OS you finish loading ;-)


It seems, then, that a better descriptive term for ntldr's function
is "load manager" rather than "boot manager".

BTW, is there a routine further along in the OS loading process
that is more complex, or is ntldr the final loader of the system?

*TimDaniels*

On Mon, 28 Mar 2005 22:24:05 -0800, "Timothy Daniels"
"cquirke (MVP Windows shell/user)" wrote:
"Timothy Daniels" wrote:

What makes it a bit confusing is that one can dip into NT this way,
and then immediately jump out again to load a different OS instead -
either DOS mode, a Win9x, or Recovery Console. Those OSs are loaded
in the same way; NTLDR loads a partition boot sector code image as if
that had been in effect as the partition code, and jumps into that.

But by your definition, it's not part of the "boot" process.
Is it right to call it just a "loader"?

No, it's part of NT (NT, Win2000, XP). It's just that the OS you
start loading isn't always the OS you finish loading ;-)

It seems, then, that a better descriptive term for ntldr's function
is "load manager" rather than "boot manager".

Yes; I like that terminology. It's the difference between walking the
high street looking at Ford vs. BMW vs. Mazda, and walking into Ford's
showroom and looking at Falcon vs. Mustang vs. Cortina.

BTW, is there a routine further along in the OS loading process
that is more complex, or is ntldr the final loader of the system?

It will read like the Old Testiment, i.e. X begats Y begats Z, etc.



---------- ----- ---- --- -- - - - -
Gone to bloggery: http://cquirke.blogspot.com
---------- ----- ---- --- -- - - - -