Windows directory: $hf_mig$

[Bullwinkle]

I'm running windows xp sp2 an looking in the c:\windows directory, the first
sub directory is the on above $hf_mig$.

In looking at the contents of this directory I see all the windows updates
in sub-directories.

Do I need to keep these files? Or can I delete them without causing problems
in running windows?

Any help appreciated

Regards,

 

[Guest]

Hi Bullwinkle,

You should keep those files. Have a look here for a description of the
function of that folder:

Description of the contents of Windows XP Service Pack 2 and Windows Server
2003 software update packages
http://support.microsoft.com/default.aspx?scid=824994
Scroll down to the "Notes" section



Regards,

 

[cquirke (MVP Windows shell/user)]

On Tue, 10 Jul 2007 07:22:01 -0700, Patti MacLeod

You should keep those files. Have a look here for a description of the
function of that folder:

Description of the contents of Windows XP Service Pack 2 and Windows Server
2003 software update packages
http://support.microsoft.com/default.aspx?scid=824994
Scroll down to the "Notes" section

You can keep them, but not have them clogging up C:, if you have a
small C: volume that is intended to operate as a lean, fast "engine
room". As it is, for every 1 active code file, you can end up with 3
inactive forms of the same file (DLL cache, patch backup version,
patch new vserion). I can't see a 75% dead-weight overhead tolerated
anywhere else (75% RAM unused, 75% CPU cycles unused), can you?

That way, if you need to "undo" a patch, you can copy the material
back to C: before doing so.

If you follow this approach, you can free about 600M to over 1G on C:,
just by moving off the main $hf_mig$ subtree and the other $..$
subtrees that follow it.

That's before you look at "harder" bloat targets, like Installer,
ServicePackFiles, Downloaded Installations and SoftwareDistribution.
Of those, I generally only relocate ServicePackFiles.

However, if you do relocate these, it's important to keep layering the
relocated set with new stuff as it is spawned by updates etc. else the
subtree will no longer be complete and accurate.

 

[Bullwinkle]

Thank you Patti. I basically just want to know if this is wasted space. I
have plenty of room to keep them there.

Regards,

 

[Bullwinkle]

Thank you for the info.

Regards,

On Tue, 10 Jul 2007 07:22:01 -0700, Patti MacLeod


You can keep them, but not have them clogging up C:, if you have a
small C: volume that is intended to operate as a lean, fast "engine
room". As it is, for every 1 active code file, you can end up with 3
inactive forms of the same file (DLL cache, patch backup version,
patch new vserion). I can't see a 75% dead-weight overhead tolerated
anywhere else (75% RAM unused, 75% CPU cycles unused), can you?

That way, if you need to "undo" a patch, you can copy the material
back to C: before doing so.

If you follow this approach, you can free about 600M to over 1G on C:,
just by moving off the main $hf_mig$ subtree and the other $..$
subtrees that follow it.

That's before you look at "harder" bloat targets, like Installer,
ServicePackFiles, Downloaded Installations and SoftwareDistribution.
Of those, I generally only relocate ServicePackFiles.

However, if you do relocate these, it's important to keep layering the
relocated set with new stuff as it is spawned by updates etc. else the
subtree will no longer be complete and accurate.

 

[Guest]

You're welcome



Regards,

 

[LadyDungeness]

So, If I move the $hf_mig$ subtree & folders to my G drive, will
future updates know to add their folders there? Should I make a
redirect command or regkey someplace? Should I just add new $hf_mig$
subfolders as they accumulate?

Thanks.

`
Lady Dungeness
Crabby, but Delicious!
~~~~~~~~~~~~~~~~~~~~~~


On Wed, 11 Jul 2007 12:01:10 +0200, "cquirke (MVP Windows shell/user)"

|On Tue, 10 Jul 2007 07:22:01 -0700, Patti MacLeod
|
|>You should keep those files. Have a look here for a description of the
|>function of that folder:
|>
|>Description of the contents of Windows XP Service Pack 2 and Windows Server
|>2003 software update packages
|>http://support.microsoft.com/default.aspx?scid=824994
|>Scroll down to the "Notes" section
|
|You can keep them, but not have them clogging up C:, if you have a
|small C: volume that is intended to operate as a lean, fast "engine
|room". As it is, for every 1 active code file, you can end up with 3
|inactive forms of the same file (DLL cache, patch backup version,
|patch new vserion). I can't see a 75% dead-weight overhead tolerated
|anywhere else (75% RAM unused, 75% CPU cycles unused), can you?
|
|That way, if you need to "undo" a patch, you can copy the material
|back to C: before doing so.
|
|If you follow this approach, you can free about 600M to over 1G on C:,
|just by moving off the main $hf_mig$ subtree and the other $..$
|subtrees that follow it.
|
|That's before you look at "harder" bloat targets, like Installer,
|ServicePackFiles, Downloaded Installations and SoftwareDistribution.
|Of those, I generally only relocate ServicePackFiles.
|
|However, if you do relocate these, it's important to keep layering the
|relocated set with new stuff as it is spawned by updates etc. else the
|subtree will no longer be complete and accurate.
|
|
|
|>------------------------- ---- --- -- - - - -
| I'm on a ten-year lunch break
|>------------------------- ---- --- -- - - - -

 

[cquirke (MVP Windows shell/user)]

So, If I move the $hf_mig$ subtree & folders to my G drive, will
future updates know to add their folders there?

No. A new $hf_mig$ will be spawned in %WinDir% again.

Should I make a redirect command or regkey someplace?

AFAIK it's hard-coded, hence...

http://cquirke.spaces.live.com/blog/cns!C7DAB1E724AB8C23!302.entry

Should I just add new $hf_mig$
subfolders as they accumulate?

That's what I've been doing

--------------- ---- --- -- - - - -

Saws are too hard to use.
Be easier to use!