OK, the way HDs have been set up is extremely wasteful, one giant
track and sector. All else there off to the sunset forever.
Seems to me, in the old daze where space was supposedly at a premium,
that non-use would have been accounted for.

Well, as a result we have (at least i do not have to take off my
shoes to count) 4 (four) partitions.
Then these new(er) OSes go and WASTE one by eating a partition for
the very space-wasting boot sector...more space non-use.
And they go-ahead and byte all over attempts to (independently) make
4 partitions.

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.
And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a partition
and use a partition manager to select and boot where i want.

Comments?
--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

On Wed, 6 Oct 2021 19:19:38 -0700, Robert Baer wrote:

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.

IMO, there's no point of having two Windows versions where both are NT6+.
When one is NT6+, it would only make sense if the the other is NT 4.x-5.x, because there are things that are no longer available in NT6+.

And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a partition
and use a partition manager to select and boot where i want.

That would work if it's just in terms of copying/restoring files from
different drives or backup/archive.

Copying OS requires more effort. Especially for Windows OS. Linux OS on the other hand, surprisingly doesn't require as much effort as Windows OS, as
its boot manager is quite flexible already, and the OS doesn't bound itself
to a specific partition location & size like Windows does.

IOTW, for Linux there's a high chance that it may not require additional effort. But for Windows, there's a high chance that it will require
additional effort. ...to make the copied OS partition be bootable from the
new drive.
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/6/21 8:19 PM, Robert Baer wrote:

OK, the way HDs have been set up is extremely wasteful, one giant
track and sector.

Please define what you mean by "one giant track and sector". *ALL* of
the hard drives that I've used for the past 30 years, and I do mean
/all/, have had many tracks, with each track having many sectors therein.

All else there off to the sunset forever.

And yet we have many different partitioning schemes. So it seems to me
that the first thing used wasn't /forever/.

Seems to me, in the old daze where space was supposedly at a premium,
that non-use would have been accounted for.

The slack space at the end of the files verses the allocation that they consumed has been a form of contention for the entire time that I've
been using hard drives. That's where cluster sizes comes into play with
file systems.

Well, as a result we have (at least i do not have to take off my
shoes to count) 4 (four) partitions.

What is a partition?

- primary
- logical
- BSD disk slice
- Microsoft dynamic
- Logical Volume
- something else

Then these new(er) OSes go and WASTE one by eating a partition for
the very space-wasting boot sector...more space non-use.

I suppose that you're talking about Microsoft OSes. I can still install contemporary Linux in a single partition with the boot loader living in
the unused space of the first sector that is not included in a partition.

And they go-ahead and byte all over attempts to (independently) make
4 partitions.

I think your attempts are significantly different than my ... successes.

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.

DOS will almost certainly need it's own partition.

Linux will almost certainly need it's own partition.

Windows 7 and Windows 10 may be able to be installed in the same common partition. I've done so with Windows NT 4.0, Windows 2000, and Windows
XP in the past. -- I'm sure it's contrary to Microsoft's
recommendations. But that doesn't mean that it's not possible.

And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a partition
and use a partition manager to select and boot where i want.

/Partition/ managers and /boot/ managers are distinctly *different*
things. They frequently come as a bundle, but they are quite different things. The former is meant for altering partition tables. The latter
is meant to allow you to boot different OSs. Sometimes the latter only automates altering the active partition. Usually the latter is in and
of itself a micro OS that presents you with a menu to choose what OS you
want to chain load.

Comments?

There is a lot of minutia in what you are talking about. Some minutia
is compatible with other minutia, and some is incompatible.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/6/2021 10:19 PM, Robert Baer wrote:

  OK, the way HDs have been set up is extremely wasteful, one giant track and sector. All else there off to the sunset forever.
  Seems to me, in the old daze where space was supposedly at a premium, that non-use would have been accounted for.

  Well, as a result we have (at least i do not have to take off my shoes to count) 4 (four) partitions.
  Then these new(er) OSes go and WASTE one by eating a partition for the very space-wasting boot sector...more space non-use.
  And they go-ahead and byte all over attempts to (independently) make 4 partitions.

  So, for arguments sake, i would like to put on the hard drive DOS, Linux, Win7 and Win10.
  And since i am starting with 4 drives i would naively partition an empty drive into 4 partitions, and just copy each drive onto a partition and use a partition manager to select and boot where i want.

   Comments?

Multiboot is a "learn by doing" topic.

It takes time and effort. There are no free lunches served.

Spoon feeding instructions just isn't the same.

GPT partitioning allows a *lot* of partitions.
It's a linear scheme, that requires no Extended/Logical.
But we don't know a thing about your computer, so that
puts the topic off limits for the moment.

Legacy MSDOS partitioning, the partition table in sector zero
has room for four entries. For max partitions, you place three
primary partitions, and the last partition is an Extended. In
the Extended go the Logical partitions. The logical partitions
are "chained" in some way, so again, a large number is possible.

https://i.postimg.cc/0jMDV94v/Multiboot-Legacy-Partition.gif

In that example, I used spacer partitions to keep Windows "hemmed in".
This prevented the extra partitions from being used. But, this is
also a naughty practice, and prevents the automatic recovery on
Windows 10. What I did in that picture, was to please you, and
is not Best Practice.

Both schemes (GPT,MSDOS) can keep a multibooter happy.

When Windows does a "Clean Install", by booting the DVD, that
tends to mess up the boot materials. If Windows does "Upgrade Installs",
today, it doesn't have to mess up the Linux materials. For piece of mind,
I'd do Win7, then Win10, then Linux. But when 21H2 or whatever for Windows 10 comes along, there should not be an issue with Linux getting trashed.
At least on a GPT setup :-)

Macrium Reflect emergency boot CD, can do Windows boot repair. What it
cannot do, is "add files", if you did something dumb and deleted the
boot files. But what it does do, is it modifies GUIDs on partitions,
composes a new BCD menu file, and tips things upright.

Linux has a third-party boot repair, but for me, it's only worked maybe
once or twice out of all the attempts. For Linux boot repair, you'll
need your own wits to make it happen. By chrooting in and doing it (grub-install for example, or update-grub if you modified /etc/defaults/grub).

Good luck, Mr.Noob [I've only lost a couple disks full of data, doing this!!!]
[Mr.Backup is your friend when multibooting]

And there's always a Level 39 wizard out there, to put you in your place :-)

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

JJ wrote:

On Wed, 6 Oct 2021 19:19:38 -0700, Robert Baer wrote:

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.

IMO, there's no point of having two Windows versions where both are NT6+. When one is NT6+, it would only make sense if the the other is NT 4.x-5.x, because there are things that are no longer available in NT6+.

* Do not know what NT6+ etc are.
Presume Win7 and Win10 are both NT6.
So maybe put Linux, WinXP, Win7 and Win10 on.
Very messy, as i only have disty install WinXP SP2 and separate SP3
install disc.

And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a partition
and use a partition manager to select and boot where i want.

That would work if it's just in terms of copying/restoring files from different drives or backup/archive.

Copying OS requires more effort. Especially for Windows OS. Linux OS on the other hand, surprisingly doesn't require as much effort as Windows OS, as
its boot manager is quite flexible already, and the OS doesn't bound itself to a specific partition location & size like Windows does.

IOTW, for Linux there's a high chance that it may not require additional effort. But for Windows, there's a high chance that it will require additional effort. ...to make the copied OS partition be bootable from the new drive.

--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

Grant Taylor wrote:

On 10/6/21 8:19 PM, Robert Baer wrote:

OK, the way HDs have been set up is extremely wasteful, one giant
track and sector.

Please define what you mean by "one giant track and sector".  *ALL* of
the hard drives that I've used for the past 30 years, and I do mean
/all/, have had many tracks, with each track having many sectors therein.

All else there off to the sunset forever.

And yet we have many different partitioning schemes.  So it seems to me that the first thing used wasn't /forever/.

Seems to me, in the old daze where space was supposedly at a premium,
that non-use would have been accounted for.

The slack space at the end of the files verses the allocation that they consumed has been a form of contention for the entire time that I've
been using hard drives.  That's where cluster sizes comes into play with file systems.

Well, as a result we have (at least i do not have to take off my shoes
to count) 4 (four) partitions.

What is a partition?

 - primary
 - logical
 - BSD disk slice
 - Microsoft dynamic
 - Logical Volume
 - something else

Then these new(er) OSes go and WASTE one by eating a partition for the
very space-wasting boot sector...more space non-use.

I suppose that you're talking about Microsoft OSes.  I can still install contemporary Linux in a single partition with the boot loader living in
the unused space of the first sector that is not included in a partition.

* THAT is the sane way to go, and the why of my complaint.
There should be a way to fiddle an M$ install to get that (and have
the whole thing in the first of the 2 partitions).

And they go-ahead and byte all over attempts to (independently) make 4
partitions.

I think your attempts are significantly different than my ... successes.

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.

DOS will almost certainly need it's own partition.

Linux will almost certainly need it's own partition.

Windows 7 and Windows 10 may be able to be installed in the same common partition.  I've done so with Windows NT 4.0, Windows 2000, and Windows
XP in the past.  --  I'm sure it's contrary to Microsoft's recommendations.  But that doesn't mean that it's not possible.

And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a
partition and use a partition manager to select and boot where i want.

/Partition/ managers and /boot/ managers are distinctly *different* things.  They frequently come as a bundle, but they are quite different things.  The former is meant for altering partition tables.  The latter
is meant to allow you to boot different OSs.  Sometimes the latter only automates altering the active partition.  Usually the latter is in and
of itself a micro OS that presents you with a menu to choose what OS you want to chain load.

* That is what i am familiar with and works very well.
The scheme of copy to partition for an OS, then use of a boot manager avoids the hassle and hours of re-installing lottza programs.

Comments?

There is a lot of minutia in what you are talking about.  Some minutia
is compatible with other minutia, and some is incompatible.

--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

Grant Taylor wrote:

On 10/6/21 8:19 PM, Robert Baer wrote:

OK, the way HDs have been set up is extremely wasteful, one giant
track and sector.

Please define what you mean by "one giant track and sector".  *ALL* of
the hard drives that I've used for the past 30 years, and I do mean
/all/, have had many tracks, with each track having many sectors therein.

* The partition sector is the worst.
ONE sector is actually used, and (to exaggerate) ten thousands of
others get to be unused and unobtanium merely because they are on the
same cylinder.

All else there off to the sunset forever.

And yet we have many different partitioning schemes.  So it seems to me that the first thing used wasn't /forever/.

Seems to me, in the old daze where space was supposedly at a premium,
that non-use would have been accounted for.

The slack space at the end of the files verses the allocation that they consumed has been a form of contention for the entire time that I've
been using hard drives.  That's where cluster sizes comes into play with file systems.

* Poor solution, i think.
More aggressive and allocation messy, is to (behind the scene) have variable start location and variable file length.
A no-byte-goes-unused scheme.

Well, as a result we have (at least i do not have to take off my shoes
to count) 4 (four) partitions.

What is a partition?

 - primary
 - logical
 - BSD disk slice
 - Microsoft dynamic
 - Logical Volume
 - something else

Then these new(er) OSes go and WASTE one by eating a partition for the
very space-wasting boot sector...more space non-use.

I suppose that you're talking about Microsoft OSes.  I can still install contemporary Linux in a single partition with the boot loader living in
the unused space of the first sector that is not included in a partition.

And they go-ahead and byte all over attempts to (independently) make 4
partitions.

I think your attempts are significantly different than my ... successes.

So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.

DOS will almost certainly need it's own partition.

* A complete DOS environment will live happily in a WIN folder.

Linux will almost certainly need it's own partition.

Windows 7 and Windows 10 may be able to be installed in the same common partition.  I've done so with Windows NT 4.0, Windows 2000, and Windows
XP in the past.  --  I'm sure it's contrary to Microsoft's recommendations.  But that doesn't mean that it's not possible.

And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a
partition and use a partition manager to select and boot where i want.

/Partition/ managers and /boot/ managers are distinctly *different* things.  They frequently come as a bundle, but they are quite different things.  The former is meant for altering partition tables.  The latter
is meant to allow you to boot different OSs.  Sometimes the latter only automates altering the active partition.  Usually the latter is in and
of itself a micro OS that presents you with a menu to choose what OS you want to chain load.

* The boot manager allows a copy of a configured OS to me in a
partition, avoiding need to re-install a bunch of programs.
That can save daze for a given OS.
But the eat-2-partition-per-os Micro$uck model decreases number of
OSes that can live on a drive.

Comments?

There is a lot of minutia in what you are talking about.  Some minutia
is compatible with other minutia, and some is incompatible.

--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

Paul wrote:

On 10/6/2021 10:19 PM, Robert Baer wrote:

   OK, the way HDs have been set up is extremely wasteful, one giant
track and sector. All else there off to the sunset forever.
   Seems to me, in the old daze where space was supposedly at a
premium, that non-use would have been accounted for.

   Well, as a result we have (at least i do not have to take off my
shoes to count) 4 (four) partitions.
   Then these new(er) OSes go and WASTE one by eating a partition for
the very space-wasting boot sector...more space non-use.
   And they go-ahead and byte all over attempts to (independently)
make 4 partitions.

   So, for arguments sake, i would like to put on the hard drive DOS,
Linux, Win7 and Win10.
   And since i am starting with 4 drives i would naively partition an
empty drive into 4 partitions, and just copy each drive onto a
partition and use a partition manager to select and boot where i want.

    Comments?

Multiboot is a "learn by doing" topic.

It takes time and effort. There are no free lunches served.

Spoon feeding instructions just isn't the same.

GPT partitioning allows a *lot* of partitions.
It's a linear scheme, that requires no Extended/Logical.
But we don't know a thing about your computer, so that
puts the topic off limits for the moment.

Legacy MSDOS partitioning, the partition table in sector zero
has room for four entries. For max partitions, you place three
primary partitions, and the last partition is an Extended. In
the Extended go the Logical partitions. The logical partitions
are "chained" in some way, so again, a large number is possible.

   https://i.postimg.cc/0jMDV94v/Multiboot-Legacy-Partition.gif

In that example, I used spacer partitions to keep Windows "hemmed in".
This prevented the extra partitions from being used. But, this is
also a naughty practice, and prevents the automatic recovery on
Windows 10. What I did in that picture, was to please you, and
is not Best Practice.

Both schemes (GPT,MSDOS) can keep a multibooter happy.

When Windows does a "Clean Install", by booting the DVD, that
tends to mess up the boot materials. If Windows does "Upgrade Installs", today, it doesn't have to mess up the Linux materials. For piece of mind,
I'd do Win7, then Win10, then Linux. But when 21H2 or whatever for
Windows 10
comes along, there should not be an issue with Linux getting trashed.
At least on a GPT setup :-)

Macrium Reflect emergency boot CD, can do Windows boot repair. What it
cannot do, is "add files", if you did something dumb and deleted the
boot files. But what it does do, is it modifies GUIDs on partitions,
composes a new BCD menu file, and tips things upright.

Linux has a third-party boot repair, but for me, it's only worked maybe
once or twice out of all the attempts. For Linux boot repair, you'll
need your own wits to make it happen. By chrooting in and doing it (grub-install for example, or update-grub if you modified /etc/defaults/grub).

Good luck, Mr.Noob [I've only lost a couple disks full of data, doing this!!!]
                   [Mr.Backup is your friend when multibooting]

And there's always a Level 39 wizard out there, to put you in your place
:-)

   Paul

Grumble grumble trouble in River City.
Followed https://www.howtogeek.com/74335/how-to-dual-boot-windows-7-and-windows-8/
up to "Once you click the shrink button Windows will shrink your
drive. Once it has completed shrinking your partition, the Disk Manager
will display this new space with a black header meaning it is an empty partition, you now need to give it a file system. To format the drive
with a file system right click on the black space and select “New Simple Volume”.

So set the size then drive letter.
The format partition step of the wizard is where it goes bonkers.
After supposedly formatting, i get the wonderful killing message box:
"The format did not complete successfully OK".
Most definitely *NOT* ok.

DOA, cannot put anything in second partition.

So what program can i use to fix that?

Thanks, R. Baer
--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/8/2021 2:35 AM, Robert Baer wrote:

  Grumble grumble trouble in River City.
  Followed https://www.howtogeek.com/74335/how-to-dual-boot-windows-7-and-windows-8/
  up to "Once you click the shrink button Windows will shrink your drive. Once it has completed shrinking your partition, the Disk Manager will display this new space with a black header meaning it is an empty partition, you now need to give it a file system. To format the drive with a file system right click on the black space and select “New Simple Volume”.

  So set the size then drive letter.
  The format partition step of the wizard is where it goes bonkers.
  After supposedly formatting, i get the wonderful killing message box: "The format did not complete successfully     OK".
  Most definitely *NOT* ok.

  DOA, cannot put anything in second partition.

  So what program can i use to fix that?

    Thanks,  R. Baer

You can see in my example here, I tried to format a
partition to FAT32 when the partition was too big
for Microsofts rules to let it happen. Even though
FAT32 formats just fine up to 2.2TB in size using a
third-party tool.

https://i.postimg.cc/N0znXp3n/formatting-fail.gif

In Disk Management, you can make a new partition RAW,
then resort to command line "format" command, as in the picture.

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/7/21 11:05 PM, Robert Baer wrote:

* THAT is the sane way to go, and the why of my complaint.
There should be a way to fiddle an M$ install to get that (and have
the whole thing in the first of the 2 partitions).

As far as I understand, Microsoft has not supported installing Windows
in a single partition since Windows Vista.

* That is what i am familiar with and works very well.
The scheme of copy to partition for an OS, then use of a boot manager
avoids the hassle and hours of re-installing lottza programs.

I disagree.

Copying something inherently means that there is something to be copied.
Thus the something to be copied implies that time has been spent
installing programs. Thus you aren't avoiding the time installing programs.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/7/21 11:41 PM, Robert Baer wrote:

* The partition sector is the worst.
  ONE sector is actually used, and (to exaggerate) ten thousands of
others get to be unused and unobtanium merely because they are on the
same cylinder.

Other things have been using the otherwise unused space in the sector
for years.

Also, the standard / common sector sizes are either 512 bytes or 4096 bytes.

Different partitioning schemes may start on the next track or the next
sector. But there isn't a lot of space that goes unused. By my rough
math you're talking about wasting between 32 kB and 256 kB. I feel like
more space than that is wasted elsewhere in the disk / partition / file system.

* Poor solution, i think.
  More aggressive and allocation messy, is to (behind the scene) have variable start location and variable file length.

Different file systems deal with the unused space in clusters in
different ways. Look beyond FAT and NTFS.

  A no-byte-goes-unused scheme.

Anything that that gets high enough granularity to use all bytes is
going to have performance penalties.

* A complete DOS environment will live happily in a WIN folder.

Not really.

The last time I checked, no DOS kernel understands NTFS. You need new
enough versions of DOS to understand FAT-32. So, you can store the
files in a folder, but DOS won't be able to access, much less boot from,
files in a folder on anything other than FAT.

* The boot manager allows a copy of a configured OS to me in a
partition, avoiding need to re-install a bunch of programs.

Again, a boot manager doesn't avoid the need to install the programs.
You still have to install them at least once, if not once per OS.

The days of sharing installed programs between different installations
of Windows are long gone. Windows 3.x was the last version that you
could safely assume could share programs between installations. Since
then, there are considerably more files than just in the program's directories. And let's not forget the program's entries in the registry.

  That can save daze for a given OS.

No it can't.

  But the eat-2-partition-per-os Micro$uck model decreases number of
OSes that can live on a drive.

I don't think it's two partitions per Windows. I think you can have a
shared UEFI (boot) partition plus one partition per Windows installation.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/8/2021 12:20 PM, Grant Taylor wrote:

On 10/7/21 11:05 PM, Robert Baer wrote:

* THAT is the sane way to go, and the why of my complaint.
There should be a way to fiddle an M$ install to get that (and have the whole thing in the first of the 2 partitions).

As far as I understand, Microsoft has not supported installing Windows in a single partition since Windows Vista.

Win7 and Win10 here, are single-partition each.

https://i.postimg.cc/0jMDV94v/Multiboot-Legacy-Partition.gif

That's my demo install in a VM, showing it can be done.

It would be near impossible to do that on GPT, because
the installer has "weed-like" behavior then. To do it on
GPT, requires a sacrificial install and some cloning-over.
Which is cheating, but might still be do-able. The above
procedure was more straight-forward.

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/8/21 4:56 PM, Paul wrote:

It would be near impossible to do that on GPT, because the installer
has "weed-like" behavior then.

Ah. So the difference is how Windows installer works with PC/BIOS/DOS partition tables vs GPT partition tables. That makes sense. If memory serves, I think that GPT became the default with Windows Vista.

To do it on GPT, requires a sacrificial install and some cloning-over.
Which is cheating, but might still be do-able. The above procedure
was more straight-forward.

I don't know if it's cheating per se, or if it's just outsmarting
Windows and making it do what you want it to do.

Aside: I wonder what the fixboot / fixmbr commands would do after such
a clone.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/8/2021 7:50 PM, Grant Taylor wrote:

On 10/8/21 4:56 PM, Paul wrote:

It would be near impossible to do that on GPT, because the installer has "weed-like" behavior then.

Ah.  So the difference is how Windows installer works with PC/BIOS/DOS partition tables vs GPT partition tables.  That makes sense.  If memory serves, I think that GPT became the default with Windows Vista.

To do it on GPT, requires a sacrificial install and some cloning-over. Which is cheating, but might still be do-able. The above procedure was more straight-forward.

I don't know if it's cheating per se, or if it's just outsmarting Windows and making it do what you want it to do.

Aside:  I wonder what the fixboot / fixmbr commands would do after such a clone.

Fixboot/fixmbr are WinXP utilities.

They no longer exist in Vista+.

There are other utilities now.

https://support.microsoft.com/en-us/topic/use-bootrec-exe-in-the-windows-re-to-troubleshoot-startup-issues-902ebb04-daa3-4f90-579f-0fbf51f7dd5d

bootrec /FixMbr /FixBoot /RebuildBcd

https://docs.microsoft.com/en-us/windows-hardware/manufacture/desktop/bcdboot-command-line-options-techref-di

There was some other command that could put a selected
version of the boot loader back. And you could control
whether a WinXP one, or a Vista+ one was used. But I
never got that to work properly when I tested.

I just use a Macrium Reflect Free emergency boot CD and
its "boot repair", to stitch together things. I haven't tested
whether it can put an intelligent 440 byte boot loader on the
MBR or not. Normally, just a few GUIDs need to be corrected,
and the "boot repair" does that for me. What it should not be
doing, is putting any necessary files back. If missing files
prevent boot, it's not intended to fix that.

You can't be too careless with multiboots, because you
might inadvertently replace the Linux MBR boot loader with
the Windows one. That's why I'm not "fixmbr" all that often
these days :-) "Measure twice, cut once" :-)

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

Paul wrote:

On 10/8/2021 2:35 AM, Robert Baer wrote:

   Grumble grumble trouble in River City.
   Followed
https://www.howtogeek.com/74335/how-to-dual-boot-windows-7-and-windows-8/
   up to "Once you click the shrink button Windows will shrink your
drive. Once it has completed shrinking your partition, the Disk
Manager will display this new space with a black header meaning it is
an empty partition, you now need to give it a file system. To format
the drive with a file system right click on the black space and select
“New Simple Volume”.

   So set the size then drive letter.
   The format partition step of the wizard is where it goes bonkers.
   After supposedly formatting, i get the wonderful killing message box: >> "The format did not complete successfully     OK".
   Most definitely *NOT* ok.

   DOA, cannot put anything in second partition.

   So what program can i use to fix that?

     Thanks,  R. Baer

You can see in my example here, I tried to format a
partition to FAT32 when the partition was too big
for Microsofts rules to let it happen. Even though
FAT32 formats just fine up to 2.2TB in size using a
third-party tool.

   https://i.postimg.cc/N0znXp3n/formatting-fail.gif

In Disk Management, you can make a new partition RAW,
then resort to command line "format" command, as in the picture.

   Paul

Both exFAT and NTFS acted the same way; i get a new partition RAW by default.
Did not try CMD, so
..seems to want to work with format N: but slow as all heck.
So tried format N: /q and got the following messages:
Creating file system structures.
The first NTFS boot sector is unreadabe.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

So, maybe the SLOW way or NO WAY.

Thanks.
--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/10/2021 4:36 AM, Robert Baer wrote:

Paul wrote:

On 10/8/2021 2:35 AM, Robert Baer wrote:

   Grumble grumble trouble in River City.
   Followed
https://www.howtogeek.com/74335/how-to-dual-boot-windows-7-and-windows-8/ >>>    up to "Once you click the shrink button Windows will shrink your drive. Once it has completed shrinking your partition, the Disk Manager will display this new space with a black header meaning it is an empty partition, you now need to give it a file system. To format the drive with a file system right click on the black space and select “New Simple Volume”.

   So set the size then drive letter.
   The format partition step of the wizard is where it goes bonkers.
   After supposedly formatting, i get the wonderful killing message box: >>> "The format did not complete successfully     OK".
   Most definitely *NOT* ok.

   DOA, cannot put anything in second partition.

   So what program can i use to fix that?

     Thanks,  R. Baer

You can see in my example here, I tried to format a
partition to FAT32 when the partition was too big
for Microsofts rules to let it happen. Even though
FAT32 formats just fine up to 2.2TB in size using a
third-party tool.

    https://i.postimg.cc/N0znXp3n/formatting-fail.gif

In Disk Management, you can make a new partition RAW,
then resort to command line "format" command, as in the picture.

    Paul

  Both exFAT and NTFS acted the same way; i get a new partition RAW by default.
  Did not try CMD, so
..seems to want to work with format N: but slow as all heck.
  So tried format N: /q and got the following messages:
Creating file system structures.
The first NTFS boot sector is unreadabe.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

  So, maybe the SLOW way or NO WAY.

   Thanks.

Did you "Run as Administrator" for Command Prompt ?

You can't format with a limited account.
You have to be elevated.

https://i.postimg.cc/bN15jj7G/W7-elevate.gif

*******

If that is not the case, and you were elevated

whoami /user /priv

perhaps the problem is the disk drive is not working
correctly. Or, the partition you created is not the
correct size and is too small to be formatted with
the file system selected.

And stick with /q please. Doing read-verify for nothing,
is just self-deception. Either the damn thing is
writeable... or it is not. Don't delay finding out.
Use the /q, Luke.

Paul


--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/10/21 2:36 AM, Robert Baer wrote:

..seems to want to work with format N: but slow as all heck.

Did you let the slow format finish? Or did you stop it?

  So tried format N: /q and got the following messages:
Creating file system structures.
The first NTFS boot sector is unreadabe.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

If you stopped the slow / long format, that error would be unsurprising
to me.

  So, maybe the SLOW way or NO WAY.

My understanding is that the quick / short format will only run if there
is an existing file system (maybe of the same type). Something that
doesn't exist if you stopped the slow / long format before it finished.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/10/2021 12:38 PM, Grant Taylor wrote:

On 10/10/21 2:36 AM, Robert Baer wrote:

..seems to want to work with format N: but slow as all heck.

Did you let the slow format finish?  Or did you stop it?

   So tried format N: /q and got the following messages:
Creating file system structures.
The first NTFS boot sector is unreadabe.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

If you stopped the slow / long format, that error would be unsurprising to me.

   So, maybe the SLOW way or NO WAY.

My understanding is that the quick / short format will only run if there is an existing file system (maybe of the same type).  Something that doesn't exist if you stopped the slow / long format before it finished.

No.

Short format = write new $MFT, done

Long Format = write new $MFT, then read verify the entire NTFS, if any
blocks return CRC error, mark off bad clusters using $BAD
sparse metadata method.

The long format is mostly a read operation.

The short format does everything technically
needed on good media. You should not really
be resorting to "living with bad media" by
using the $BAD method. It's a false economy.

If your media is that bad, when auto-reallocation
hardware support cannot fix a CRC error, then that
disk is in trouble, and on its way out.

While it is possible to have high-fly errors on
a disk drive, and marking off the CRC errored sectors
is perfectly good, this sort of error type is uncommon.
I've only seen it once here, and just after
committing a cardinal sin of turning an old
drive upside down. That's not a good idea, but
because the drive was freezing right-side-up,
I had no choice but to try it.

*******

Back in 1985 or so, when the very first $1000+ HDD
came in (5MB and 10MB drives), if you interrupted the
format of those, they would brick on you. That's because
back then, "format" meant "low-level format" on soft
sectored drives. The magnetic format written, redefined
everything. It was truly, a nuke and pave time. There
was an index mark in hardware, to signal when to write
your soft sectors.

Modern drives, the servo wedges are "permanent". You're
not allowed to "pave" any more. You can only fill potholes
(read sector and write sector interface).

Formatting does not format, not any more. It
does not do low-level format.

Windows format, like anyones format, only has the
ability to read and write sectors. It's not really
formatting as we knew it.

You don't need a file system to exist, to write over it.

The partition can be raw. The first sector of the
partition can be all zeros (the file system type
declaration), and "format" can still write it.

If we could not do that, we could not bootstrap formatting
on brand new year 2021 disk drives from the factory.

I can do a Secure Erase on a modern drive, and partition
and format it, with no reliance on any special things
pre-existing on the disk. Since the servo wedges are permanent,
the disk is always able to orient itself and follow LBA
addressing. Everything is there and everything is locatable
by address. Even Secure Erase, only writes over the payload
section of each sector.

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/10/21 12:34 PM, Paul wrote:

No.

Suffice it to say that we've had different experience.

I've had many occasions when Microsoft tools would not do a /quick/
format because there was no file system. I was /forced/ to do the
normal / long format.

The normal / long format has always been allowed on any type of partition.

Note: normal / long is not the same as actually doing a more thorough test.

I am also mostly speaking to the format command. Microsoft may have
changed their stance in their GUI format utilities over the years.
--
Grant. . . .
unix || die
--- Synchronet 3.20a-Win32 NewsLink 1.114

Paul wrote:

On 10/10/2021 12:38 PM, Grant Taylor wrote:

On 10/10/21 2:36 AM, Robert Baer wrote:

..seems to want to work with format N: but slow as all heck.

Did you let the slow format finish?  Or did you stop it?

   So tried format N: /q and got the following messages:
Creating file system structures.
The first NTFS boot sector is unreadabe.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

If you stopped the slow / long format, that error would be
unsurprising to me.

   So, maybe the SLOW way or NO WAY.

My understanding is that the quick / short format will only run if
there is an existing file system (maybe of the same type).  Something
that doesn't exist if you stopped the slow / long format before it
finished.

No.

Short format = write new $MFT, done

Long Format  = write new $MFT, then read verify the entire NTFS, if any
               blocks return CRC error, mark off bad clusters using $BAD
               sparse metadata method.

The long format is mostly a read operation.

The short format does everything technically
needed on good media. You should not really
be resorting to "living with bad media" by
using the $BAD method. It's a false economy.

If your media is that bad, when auto-reallocation
hardware support cannot fix a CRC error, then that
disk is in trouble, and on its way out.

While it is possible to have high-fly errors on
a disk drive, and marking off the CRC errored sectors
is perfectly good, this sort of error type is uncommon.
I've only seen it once here, and just after
committing a cardinal sin of turning an old
drive upside down. That's not a good idea, but
because the drive was freezing right-side-up,
I had no choice but to try it.

*******

Back in 1985 or so, when the very first $1000+ HDD
came in (5MB and 10MB drives), if you interrupted the
format of those, they would brick on you. That's because
back then, "format" meant "low-level format" on soft
sectored drives. The magnetic format written, redefined
everything. It was truly, a nuke and pave time. There
was an index mark in hardware, to signal when to write
your soft sectors.

Modern drives, the servo wedges are "permanent". You're
not allowed to "pave" any more. You can only fill potholes
(read sector and write sector interface).

Formatting does not format, not any more. It
does not do low-level format.

Windows format, like anyones format, only has the
ability to read and write sectors. It's not really
formatting as we knew it.

You don't need a file system to exist, to write over it.

The partition can be raw. The first sector of the
partition can be all zeros (the file system type
declaration), and "format" can still write it.

If we could not do that, we could not bootstrap formatting
on brand new year 2021 disk drives from the factory.

I can do a Secure Erase on a modern drive, and partition
and format it, with no reliance on any special things
pre-existing on the disk. Since the servo wedges are permanent,
the disk is always able to orient itself and follow LBA
addressing. Everything is there and everything is locatable
by address. Even Secure Erase, only writes over the payload
section of each sector.

   Paul

Well, i let the slow, long-winded "normal" format run. Over an hour
or so later, i got responses:
The new file system is NTFS.
WARNING, ALL DATA ON NON-REMOVEABLE DISK DRIVE N: WILL BE LOST!
Proceed with format(Y/N)? y
Formatting 238470M
488386560 bad sectors were encountered during the format - thesse
sectors cannot be guaranteed to have been cleaned/.
Volume label (32 characters, ENTER for none)? NewN
Creating file system structures.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

Wonderful! DOA.
Seems impossible to format that second partition.

So i deleted all that bushwah and started with the HD as one
partition, simple volume, "quick" format .. all of which takes quite a
while.
That works.
So.. what the heck..Use Properties and run error checking, scan for
and attempt recovery of bad sectors.

331007 free clusters processed.
Number seems slow to change, and the "progress" bar is maybe one
tenth across (15 minutes so far).
..looks liker it will take hours, so signing off till done.
Will report results.
--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/10/2021 8:02 PM, Grant Taylor wrote:

On 10/10/21 12:34 PM, Paul wrote:

No.

Suffice it to say that we've had different experience.

I've had many occasions when Microsoft tools would not do a /quick/ format because there was no file system.  I was /forced/ to do the normal / long format.

The normal / long format has always been allowed on any type of partition.

Note:  normal / long is not the same as actually doing a more thorough test.

I am also mostly speaking to the format command.  Microsoft may have changed > their stance in their GUI format utilities over the years.

diskmgmt.msc GUI Disk Management with format capability
diskpart.exe Command Line disk management with format capability (equal to diskmgmt.msc)
format.com Command Line (older) formatting routine. Size/context sensitive.
For WinXP, requires ExFAT package installed, if desired to do ExFAT
dd.exe Third party "helper"
disktype.exe Cygwin port of disktype, so can run on Windows

The format routine may examine Sector 0 of the target partition.
Even though it doesn't have to.

Without looking at sector 0 of the partition, the partition table contains
CHS info, offset info, size info, which define the envelope of the
file system. Using that information, the format routine can compute
how many whole clusters will fit. Thus the file system is gauged
to fit within the limits of the envelope, so the file system
doesn't (fractionally) spill off the end.

If you experience a problem, the partition has been assigned a drive
letter (so mountvol knows about it), then you can try clearing the
first sector of the partition, then run the formatter.

Instructions:

http://www.chrysocome.net/dd

Download:

http://www.chrysocome.net/downloads/dd-0.6beta3.zip

In an Administrator Command Prompt:

dd --list # collect win namespace info

\\?\Device\Harddisk3\Partition0 # partition 0 is the whole drive
link to \\?\Device\Harddisk3\DR0
Fixed hard disk media. Block size = 512
size is 57671680000 bytes
\\?\Device\Harddisk3\Partition1 # partition 1 is the offset
link to \\?\Device\HarddiskVolume31 # to the first partition
Fixed hard disk media. Block size = 512
size is 57668534272 bytes

dd if=/dev/zero of=\\?\Device\Harddisk3\Partition1 bs=512 count=1

That will zero out the file system header, preventing the
formatter from "tripping over its own shoelaces".

*******

For formatting large FAT32 volumes (up to 2.2TB), the Ridgecrop
fat32formatter can do that. This solves the ~32GB limit Microsoft
has on fat32 partitions.

https://web.archive.org/web/20200424145132/http://www.ridgecrop.demon.co.uk/index.htm?fat32format.htm

https://web.archive.org/web/20200410224838if_/http://www.ridgecrop.demon.co.uk/download/fat32format.zip

*******

You can tell when "dd" doesn't have permission to do something,
as the info about the partition will take fewer lines.
Sometimes, for a destination partition (when "cloning"), you
can format the destination partition to break the handles to
it, and allow the "dd" command to run next.

\\?\Device\Harddisk2\Partition11 \___ Not writable
link to \\?\Device\HarddiskVolume11 /
\\?\Device\Harddisk2\Partition12 \
link to \\?\Device\HarddiskVolume12 \___ Can write this one
Fixed hard disk media. Block size = 512 /
size is 104078508032 bytes /

Here, "Not writable" does not mean write protected or permanently
off limits. By fiddling with it, you can change the status
and get on with it.

*******

You can learn about the "disktype" package on Linux, before
wasting the time installing Cygwin to get a port running on
your Windows setup. I've used this a lot, since I got it.
It's too bad the hosting of this isn't at all convenient.

disktype.exe
cygwin1.dll
cyggcc_s-1.dll

sudo disktype /dev/sda # show partition info

disktype mint.iso # show ISO partitions and overlays

disktype foo.bin # show that loopback mount you've been using
# like, look inside a 4GB casper-rw.

Disktype might not do Veritas Dynamic Disks, but it does most
other stuff that matters.

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/11/2021 1:34 AM, Robert Baer wrote:

  Well, i let the slow, long-winded "normal" format run. Over an hour or so later, i got responses:
The new file system is NTFS.
WARNING, ALL DATA ON NON-REMOVEABLE DISK DRIVE N: WILL BE LOST!
Proceed with format(Y/N)?   y
Formatting 238470M
488386560 bad sectors were encountered during the format - thesse sectors cannot be guaranteed to have been cleaned/.
Volume label (32 characters, ENTER for none)?  NewN
Creating file system structures.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

  Wonderful!  DOA.
  Seems impossible to format that second partition.

  So i deleted all that bushwah and started with the HD as one partition, simple volume,
"quick" format .. all of which takes quite a while.
  That works.
  So.. what the heck..Use Properties and run error checking, scan for and attempt recovery of bad sectors.

  331007 free clusters processed.
  Number seems slow to change, and the "progress" bar is maybe one tenth across (15 minutes so far).
..looks liker it will take hours, so signing off till done.
  Will report results.

http://hdtune.com/files/hdtune_255.exe # 628KB

Install that, and use the Health tab to collect SMART data
about the drive in question. Use the menu near the top,
to select the drive you are trying to format.

I do two things with Hard Drives:

1) Read transfer curve, with the left-most tab.
This is a sensitive test for platter surface conditions.

The more modern the Windows OS, the harder it is to get the
OS to shut the fuck up, while you're trying to run this test.
Win2K was the best at this job. Win10 has the irritating properly,
that the instant you start the bench running, you look in Task
Manager, and Windows 10, on purpose, will fork a maintenance
task of its own.

2) Health tab, check "Reallocated" raw data column. Value should
be zero. If you see a finite number, like 123, that means
you're getting low on spare sectors. One of my drives here,
the remaining health claimed to be 5500 sectors, and you could
watch the "Reallocated" raw value grow from one day to the next
as the drive was used.

It might take five to ten minutes to do that much work.

*******

3) You can use the right-most tab to scan the surface for errors,
but you've already done that effectively several times now,
and waiting another two hours would be a bore. You know something
is wrong with this drive, then it's a question of how much time
to spend and so on.

The HDTune window misbehaves in some display managers, so if
you see some weirdness, "this is normal" for it :-)

Paul


--- Synchronet 3.20a-Win32 NewsLink 1.114

Paul wrote:

On 10/11/2021 1:34 AM, Robert Baer wrote:

   Well, i let the slow, long-winded "normal" format run. Over an hour
or so later, i got responses:
The new file system is NTFS.
WARNING, ALL DATA ON NON-REMOVEABLE DISK DRIVE N: WILL BE LOST!
Proceed with format(Y/N)?   y
Formatting 238470M
488386560 bad sectors were encountered during the format - thesse
sectors cannot be guaranteed to have been cleaned/.
Volume label (32 characters, ENTER for none)?  NewN
Creating file system structures.
All NTFS boot sectors are unwriteable. Cannot continue.
Format failed.

   Wonderful!  DOA.
   Seems impossible to format that second partition.

   So i deleted all that bushwah and started with the HD as one
partition, simple volume,
   "quick" format .. all of which takes quite a while.
   That works.
   So.. what the heck..Use Properties and run error checking, scan for
and attempt recovery of bad sectors.

   331007 free clusters processed.
   Number seems slow to change, and the "progress" bar is maybe one
tenth across (15 minutes so far).
..looks liker it will take hours, so signing off till done.
   Will report results.

http://hdtune.com/files/hdtune_255.exe            # 628KB

Install that, and use the Health tab to collect SMART data
about the drive in question. Use the menu near the top,
to select the drive you are trying to format.

I do two things with Hard Drives:

1) Read transfer curve, with the left-most tab.
   This is a sensitive test for platter surface conditions.

   The more modern the Windows OS, the harder it is to get the
   OS to shut the fuck up, while you're trying to run this test.
   Win2K was the best at this job. Win10 has the irritating properly,
   that the instant you start the bench running, you look in Task
   Manager, and Windows 10, on purpose, will fork a maintenance
   task of its own.

2) Health tab, check "Reallocated" raw data column. Value should
   be zero. If you see a finite number, like 123, that means
   you're getting low on spare sectors. One of my drives here,
   the remaining health claimed to be 5500 sectors, and you could
   watch the "Reallocated" raw value grow from one day to the next
   as the drive was used.

It might take five to ten minutes to do that much work.

*******

3) You can use the right-most tab to scan the surface for errors,
   but you've already done that effectively several times now,
   and waiting another two hours would be a bore. You know something
   is wrong with this drive, then it's a question of how much time
   to spend and so on.

The HDTune window misbehaves in some display managers, so if
you see some weirdness, "this is normal" for it :-)

  Paul

Checking disk is still running, maybe 1/4 "done" so i would say 2-3
daze (19,xxx,xxx free clusters processed so far).
Maybe i should let it run to bitter end for a report from the process?

We are talking about a brand new drive.
Send it back?
--
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

--- Synchronet 3.20a-Win32 NewsLink 1.114

On 10/11/2021 2:53 PM, Robert Baer wrote:

  Checking disk is still running, maybe 1/4 "done" so i would say 2-3 daze (19,xxx,xxx free clusters processed so far).
  Maybe i should let it run to bitter end for a report from the process?

  We are talking about a brand new drive.
  Send it back?

You have to do sufficient debugging, to determine where
the flaw is.

On the Typing Machine (now dead), one of the SATA ports
failed a couple years ago. And I had to be pretty careful
with it, to not use that port on any other drives. The
evidence now that I've had several failures, is it was
the motherboard at fault. But at the time, I had to
retire the drive and "mark off" the port as a
"do not use", just to be safe.

This is one reason, I keep a PCIe to SATA card here,
for testing, to eliminate a Southbridge SATA port as being
a problem. Sometimes it's the data cable that got kinked,
and the controlled impedance has a "bump" in it.

*******

If it is a new drive and not a refurb (check power on hours in SMART),
you could try writing it from end to end, then read-verify it.

dd or ddrescue could do that.

Even Windows "diskpart.exe" can do it. At least
part of the operation. The "Clean All" command will
write zeros over an entire hard drive.

After the write, you do a read verify, then check SMART
for any bad news.

Administrator Command Prompt:

diskpart
list disk
select disk 1
list partitions # The new drive should have a minimal number of partitions
clean all # takes two hours
exit

While it is running, you can Start : Run : perfmon.msc
and select the disk write performance counter and plot
the transfer speed as the writes happen. That gives some
idea whether the operation is abnormally slow or not.

There are plenty of things you could try.

*******

ddrescue is addressed a bit here.

https://www.data-medics.com/forum/how-to-clone-a-hard-drive-with-bad-sectors-using-ddrescue-t133.html

And here at the bottom. But the instructions are rather old.
And today, you don't need to compile anything. Just install
the package in some Linux distro like Ubuntu LiveDVD.

https://www.cgsecurity.org/wiki/Damaged_Hard_Disk

For a straight cloning operation, it would go like this.

Ubuntu, Terminal (example of copying /dev/sda to /dev/sdc)

sudo apt install gddrescue

man ddrescue # Manual page

sudo ddrescue -n /dev/sda /dev/sdc rescued.log # First pass...

gedit rescued.log # examine to see errors
# if you are curious

sudo ddrescue -r 1 /dev/sda /dev/sdc rescued.log # Second pass tries to
# fix the errored bits,
# It reads "rescued.log".

If you want to read-verify /dev/sda, you'd do it like this.

sudo ddrescue -n /dev/sda /dev/null rescued.log # First pass...

gedit rescued.log # Examine log for /dev/sda CRC errors

Paul
--- Synchronet 3.20a-Win32 NewsLink 1.114