Some files and no fragmentation

I have read that the pagefile.sys and registry hive files once they have
been defragged do not need it again. How can this be? When memory pages are
constanly being swapped in and out of an active pagefile.sys and registry
keys and entries being added and removed; well isn't that going to cause
fragmentation?

Bill

"Paul" wrote in message
...

[...]

I can tell you that my pagefile has not fragmented *ever*
in the last five years. That's because

1) WinXP C: partition is transferred file by file, off the
partition, and to another partition. This is a natural
way to defragment files. The WinXP partition is formatted.
Then I copy the files back. I manually copy pagefile and
hiberfile, to place them next to the fastest end of the disk.
I finish the recipe with a "fixboot C:" from recovery console.
Every file is "minty fresh". I do this operation from a
second OS.

2) The pagefile definition is not managed by Windows. I define
a fixed size (2GB on a 4GB x32 OS).
[...]

You have certainly helped Paul. But in your first explanation here you
say you copy the files to another partition and format then copy back. I
understand but then you say you copy hyberfil.sys and pagefile.sys "next to
the fastest end of the disk." How exactly do you accomplish that? I have
always heard the page or swap file should come first. Then you repair the
mbr.

Bill

Bill Cunningham wrote:
"Paul" wrote in message
...

[...]

I can tell you that my pagefile has not fragmented *ever*
in the last five years. That's because

1) WinXP C: partition is transferred file by file, off the
partition, and to another partition. This is a natural
way to defragment files. The WinXP partition is formatted.
Then I copy the files back. I manually copy pagefile and
hiberfile, to place them next to the fastest end of the disk.
I finish the recipe with a "fixboot C:" from recovery console.
Every file is "minty fresh". I do this operation from a
second OS.

2) The pagefile definition is not managed by Windows. I define
a fixed size (2GB on a 4GB x32 OS).
[...]

You have certainly helped Paul. But in your first explanation here you
say you copy the files to another partition and format then copy back. I
understand but then you say you copy hyberfil.sys and pagefile.sys "next to
the fastest end of the disk." How exactly do you accomplish that? I have
always heard the page or swap file should come first. Then you repair the
mbr.

Bill

I use Robocopy. The following recipe is for FAT32. Every time I
do a cleanup of the WinXP partition, I have a log file for copying
and pasting some of the commands.

fat32format L:
robocopy E:\ L:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v /log:robocopy_e_to_L2.log
fat32format E:
(restore name as "WINXP" on E
volumeid E: 492A-AC63
(copy pagefile, hiberfile manually from L: to E
robocopy L:\ E:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v /log:robocopy_L_to_e2.log
(fixboot C:, while booted into recovery console...)
Reboot to clean WINXP.

Robocopy skips the already-copied files, so the
pagefile and hiberfile are not copied twice.

When that recipe runs (except for the fixboot step), I use
Win2K as the running OS. And at that point, WinXP is E:
and the backup partition is L: .

The volumeid step keeps a consistent value over
the life of the OS. The value was assigned original
at installation time.

Paul

"Paul" wrote in message
...

I use Robocopy. The following recipe is for FAT32. Every time I
do a cleanup of the WinXP partition, I have a log file for copying
and pasting some of the commands.

fat32format L:
robocopy E:\ L:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_e_to_L2.log
fat32format E:
(restore name as "WINXP" on E
volumeid E: 492A-AC63
(copy pagefile, hiberfile manually from L: to E
robocopy L:\ E:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_L_to_e2.log
(fixboot C:, while booted into recovery console...)
Reboot to clean WINXP.

Robocopy skips the already-copied files, so the
pagefile and hiberfile are not copied twice.

When that recipe runs (except for the fixboot step), I use
Win2K as the running OS. And at that point, WinXP is E:
and the backup partition is L: .

The volumeid step keeps a consistent value over
the life of the OS. The value was assigned original
at installation time.

Ok I found some kind of "resource kit" on MS's website. It includes
robocopy. The thing is I don't quite understand how you know where to copy
the hibernation file and the pagefile.sys to the "beginning" of the
partition. Does robocopy do that for you?

Bill

Bill Cunningham wrote:
"Paul" wrote in message
...

I use Robocopy. The following recipe is for FAT32. Every time I
do a cleanup of the WinXP partition, I have a log file for copying
and pasting some of the commands.

fat32format L:
robocopy E:\ L:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_e_to_L2.log
fat32format E:
(restore name as "WINXP" on E
volumeid E: 492A-AC63
(copy pagefile, hiberfile manually from L: to E
robocopy L:\ E:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_L_to_e2.log
(fixboot C:, while booted into recovery console...)
Reboot to clean WINXP.

Robocopy skips the already-copied files, so the
pagefile and hiberfile are not copied twice.

When that recipe runs (except for the fixboot step), I use
Win2K as the running OS. And at that point, WinXP is E:
and the backup partition is L: .

The volumeid step keeps a consistent value over
the life of the OS. The value was assigned original
at installation time.

Ok I found some kind of "resource kit" on MS's website. It includes
robocopy. The thing is I don't quite understand how you know where to copy
the hibernation file and the pagefile.sys to the "beginning" of the
partition. Does robocopy do that for you?

Bill

If you format C: (in this case the fat32format E: operation),
the partition is clean. When you copy the pagefile, it ends
up on the left of the partition.

When you format a partition, it removed the partition boot sector,
and the "fixboot C:" done from recovery console, puts it back.
Otherwise, the minty fresh partition would not be able to boot.

Every time you format a partition, it is given a new VplumeID.
Using the sysinternals.com tool of the same name, allows
you to put things back as you found them.

Paul

| You have certainly helped Paul. But in your first explanation here you
| say you copy the files to another partition and format then copy back. I
| understand but then you say you copy hyberfil.sys and pagefile.sys "next
to
| the fastest end of the disk." How exactly do you accomplish that? I have
| always heard the page or swap file should come first. Then you repair the
| mbr.
|

I always just put the swap file on another
partition, with a fixed size. You may not need
it at all if you have a lot of RAM. I have a 4 GB
fixed swap file on my D drive, so that it won't
affect C drive. I did that through the System
applet UI -- no special tools or command line
necessary.

Sysinternals used to have a Registry defragmenter.
I don't know offhand whether it's still relevant or
still available. It reboots the system and then deals
with the Registry at reboot.

I've had some trouble with defragmenters freezing.
I don't know why. But I've just stopped using them
anyway. When I occasionally run the XP defrag to
analyze it comes back saying I don't need to run
drefrag.

I wonder what you do with your computer that you
have so much disk-related trouble. You've been asking
these questions for months now.

"Mayayana" wrote in message
...

I always just put the swap file on another
partition, with a fixed size. You may not need
it at all if you have a lot of RAM. I have a 4 GB
fixed swap file on my D drive, so that it won't
affect C drive. I did that through the System
applet UI -- no special tools or command line
necessary.

Yeah I know you can do that. Put the swap file on another parition.

Sysinternals used to have a Registry defragmenter.
I don't know offhand whether it's still relevant or
still available. It reboots the system and then deals
with the Registry at reboot.

I've had some trouble with defragmenters freezing.
I don't know why. But I've just stopped using them
anyway. When I occasionally run the XP defrag to
analyze it comes back saying I don't need to run
drefrag.

There are a couple of tools I ave used with success. Erunt and another
one I believe. The same group designed them. One is a registry defragmenter
and another compresses the registry.

I wonder what you do with your computer that you
have so much disk-related trouble. You've been asking
these questions for months now.

Ah. I was wondering if that wa going to come up. No I really don't have
a lot of disk trouble. If I do it's my own doing. I like filesystems and
their design. I use fat32 for a while then ntfs. I've been using ntfs now
for a while. If I want to remove things I usually save on a USB drive and
zero out the mbr. And sometimes the boot sectors on each volume but it's
hardly needed at that point.

Bill

In message , Bill Cunningham
writes:
[]
There are a couple of tools I ave used with success. Erunt and another
one I believe. The same group designed them. One is a registry defragmenter
and another compresses the registry.
[]
From what I remember, ERUNT's purpose is to back up the registry (as ERU
did under '9x) - not to either defragment or compress it. Of course, a
back up (or rather a restore) _could_ have a side effect of
defragmenting it, depending on the condition of the drive/partition to
which the restore takes place.
--
J. P. Gilliver. UMRA: 1960/1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Intelligence isn't complete without the full picture and the full picture is
all about doubt. Otherwise, you go the way of George Bush. - baroness Eliza
Manningham-Buller (former head of MI5), Radio Times 3-9 September 2011.

On 03 May 2015, "J. P. Gilliver (John)"
wrote in microsoft.public.windowsxp.general:

From what I remember, ERUNT's purpose is to back up the registry
(as ERU did under '9x) - not to either defragment or compress it.
Of course, a back up (or rather a restore) _could_ have a side
effect of defragmenting it, depending on the condition of the
drive/partition to which the restore takes place.

You're correct. ERUNT does not compress or degragment the registry.
The developer offers another utility, NTREGOPT, that can "optimize"
the registry:

http://www.larshederer.homepage.t-online.de/erunt/

"J. P. Gilliver (John)" wrote in message
...

[]
From what I remember, ERUNT's purpose is to back up the registry (as ERU
did under '9x) - not to either defragment or compress it. Of course, a
back up (or rather a restore) _could_ have a side effect of defragmenting
it, depending on the condition of the drive/partition to which the restore
takes place.

OK I stand corrected. I remember the name but never really used these
tools. But the link is certainly accurate. That's the tools. So what does
the "optimizer" do? What to optimize? And how would registtry backup help
with defragmentation. It might have the effect of compression too might it?
I don't quite know why one would want to backup their registry Hive files.

Bill

"Nil" wrote in message
...

You're correct. ERUNT does not compress or degragment the registry.
The developer offers another utility, NTREGOPT, that can "optimize"
the registry:

http://www.larshederer.homepage.t-online.de/erunt/

The ntregopt utility I ran today and it said the registry hive was
"reduced" by 6%. Now does that imply compression?

Bill

"Paul" wrote in message
...
Bill Cunningham wrote:
"Paul" wrote in message
...

I use Robocopy. The following recipe is for FAT32. Every time I
do a cleanup of the WinXP partition, I have a log file for copying
and pasting some of the commands.

fat32format L:
robocopy E:\ L:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_e_to_L2.log
fat32format E:
(restore name as "WINXP" on E
volumeid E: 492A-AC63
(copy pagefile, hiberfile manually from L: to E
robocopy L:\ E:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_L_to_e2.log
(fixboot C:, while booted into recovery console...)
Reboot to clean WINXP.

Robocopy skips the already-copied files, so the
pagefile and hiberfile are not copied twice.

When that recipe runs (except for the fixboot step), I use
Win2K as the running OS. And at that point, WinXP is E:
and the backup partition is L: .

The volumeid step keeps a consistent value over
the life of the OS. The value was assigned original
at installation time.

Ok I found some kind of "resource kit" on MS's website. It includes
robocopy. The thing is I don't quite understand how you know where to
copy the hibernation file and the pagefile.sys to the "beginning" of the
partition. Does robocopy do that for you?

Bill

If you format C: (in this case the fat32format E: operation),
the partition is clean. When you copy the pagefile, it ends
up on the left of the partition.

When you format a partition, it removed the partition boot sector,
and the "fixboot C:" done from recovery console, puts it back.
Otherwise, the minty fresh partition would not be able to boot.

Every time you format a partition, it is given a new VplumeID.
Using the sysinternals.com tool of the same name, allows
you to put things back as you found them.

Ok if I am undrestanding correctly you are basically "transferring" the
same OS from one parition to another. And getting it to operate on both
partitions. Also using another OS too as a side. Hum. Never thought of that.

Bill

"Paul" wrote in message
...
Bill Cunningham wrote:
"Paul" wrote in message
...

I use Robocopy. The following recipe is for FAT32. Every time I
do a cleanup of the WinXP partition, I have a log file for copying
and pasting some of the commands.

fat32format L:
robocopy E:\ L:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_e_to_L2.log
fat32format E:
(restore name as "WINXP" on E
volumeid E: 492A-AC63
(copy pagefile, hiberfile manually from L: to E
robocopy L:\ E:\ /mir /copy:datso /dcopy:t /r:3 /w:2 /zb /np /tee /v
/log:robocopy_L_to_e2.log
(fixboot C:, while booted into recovery console...)
Reboot to clean WINXP.

Robocopy skips the already-copied files, so the
pagefile and hiberfile are not copied twice.

When that recipe runs (except for the fixboot step), I use
Win2K as the running OS. And at that point, WinXP is E:
and the backup partition is L: .

The volumeid step keeps a consistent value over
the life of the OS. The value was assigned original
at installation time.

Ok I found some kind of "resource kit" on MS's website. It includes
robocopy. The thing is I don't quite understand how you know where to
copy the hibernation file and the pagefile.sys to the "beginning" of the
partition. Does robocopy do that for you?

Bill

If you format C: (in this case the fat32format E: operation),
the partition is clean. When you copy the pagefile, it ends
up on the left of the partition.

When you format a partition, it removed the partition boot sector,
and the "fixboot C:" done from recovery console, puts it back.
Otherwise, the minty fresh partition would not be able to boot.

Every time you format a partition, it is given a new VplumeID.
Using the sysinternals.com tool of the same name, allows
you to put things back as you found them.

Ok the /dcopy:t switch didn't work at all and /dcopy wan't mentioned in
the word file describing robocopy's functioning. I ran into error 32 and
limits to coping when trying to copy the pagefile.sys and then the log I set
up stopped. Robocopy kept going until I halted it. I wanted to see what was
going on. Of course the pagefile was in use.

Bill

On 04 May 2015, "Bill Cunningham" wrote in
microsoft.public.windowsxp.general:

http://www.larshederer.homepage.t-online.de/erunt/

The ntregopt utility I ran today and it said the registry hive was
"reduced" by 6%. Now does that imply compression?

Seems obvious to me that it does. Does it not to you?

"Nil" wrote in message
...
On 04 May 2015, "Bill Cunningham" wrote in
microsoft.public.windowsxp.general:

http://www.larshederer.homepage.t-online.de/erunt/

The ntregopt utility I ran today and it said the registry hive was
"reduced" by 6%. Now does that imply compression?

Seems obvious to me that it does. Does it not to you?

Not quite sure. I have tried to find some docs to this. But all I see
are possible switches to the command line prompt. Yes it certainly can be
compression. But some of these utilities that defrag say a similar thing.
Reduced by so much in fragmentation.

Bill