Jump to content











Photo
- - - - -

Resize a Partition without moving its content


  • Please log in to reply
19 replies to this topic

#1 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 25 June 2011 - 10:21 PM

Its frequently required to move data between drives for better organization, which at times involves resizing, deleting and combining partitions. There is an old known problem for which I never managed to find a solution and simply avoided. Its easy to resize (make bigger) a partition to the right, but when trying to resize to the left, any partitioning program I used so far required to move all data to the left as well - to the beginning of the resized partition.

This makes very little sense to me. When you look at any Defragmenter's work, they can move data around the partition in any direction, driven by their optimization algorithm (different for each). Again, looking at disk structure in Degragmenter, the File System data (MFT / MFT Zone) can be at any place around the disk, not nesessarely at its beginning. So why Partitioning tools can't do the same? Leave MFT and all other data where it is (at least on a user demand for non-bootable partitions), or move merely MBR & MFT to the beginning of a resized partition, but leave data (which takes a lot more space than MFT) untouched.

Its hard to comprehend the logic of moving 500Gb around the same partition, putting a hard drive to totally useless severe stress, loosing time, and risking to loose data along the way as the drive heats up. The problem was somewhat addressed in this thread on Gparted Forum, but without any solution, citing no objective reasons for lack of such feature apart from "extra programming required". Given how many people refrain from repartitioning and can't use disk space efficiently due to this very reason - to avoid senselessly moving huge amount of data to the left, the extra effort required appears well justified.

Hence a question: is there any free & non-free partitioning tool that allows to extend a partition to the LEFT without moving its data to the LEFT (except possibly boot or MFT related data)?

#2 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13594 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 26 June 2011 - 09:46 AM

You seem like not having very clear the way a NTFS volume is formatted. :thumbsup:

Some info, JFYI:
First 16 sectors are the bootsector and metadata.
The $MFT is always (normally) starting on cluster 786432.
The $MFT reserved space size is proportional to size of the Volume. (normal is 12.5% of Volume size)
The $MFT Mirror is at half of the Volume.
First and only sector outside the Volume and inside partition space is the backup bootsector.
Small files (up to 1.5Kb) are stored directly inside the $MFT.

AFAIK the few programs that can reliably resize a NTFS partition do "whatever is needed" without "moving all data to the beginning of the partition".

Have you tried Partition Logic?

http://partitionlogic.org.uk/index.php
the 0.7 version is quite recent, maybe it can do the resizing "to the left"? :thumbsup:

:cheers:
Wonko

#3 paraglider

paraglider

    Gold Member

  • .script developer
  • 1707 posts
  • Location:NC,USA
  •  
    United States

Posted 26 June 2011 - 11:26 AM

I have used Paragon's Partition Manager to resize partitions many times. Never had any problems. Never worried about how long it takes.

#4 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 26 June 2011 - 05:15 PM

Hi Wonko,

Assuming you read the linked topic above, you understand the question. Its not a matter of a program being "new", but a matter of these programs not-including a somewhat different algorithm of resizing to the left than the algorithm they use to resize to the right. In particular, instead of using a separate algorithm, the programs just copy all disk content to the left when resizing a partition to the left, and then expand it to the right to new size using existing "right" algorithm. This is NOT, what many users look for. I don't need my data moved anywhere stressing the disk, except MFT & MBR if needed. More important, its NOT REQUIRED to move all data in this scenario, its just the existing programs I tried lack a separate algorithm to do it more efficiently without moving data (fiscal reasons).

I can mention several analogies to this situation (new Acrobat doesn't have MultiTab UI. Or Data Recovery soft doesn't recover & recreate only MFT to restore access to data on a disk (logically there is no need to copy data anywhere for that), they always want instead to COPY found data to a new disk and use existing MS algorithm to create MFT along the way - also fiscally motivated; etc.). Closer to the point, as I mentioned above, Defragmenters move data where they want, and place MFT where they want every time, yet the system works perfect each time after I run PerfectDisk to defrag. :)

Returning to my question, does anyone come across a Partition Manager that CAN expand a partition to the LEFT without moving all its data to the left? Only your PERSONAL experience with such tool matters, no abstractions pls. Since resizing a partition to the left is one of standard scenarios, someone hopefully will find enough motivation to write a program that will do it without senselessly moving all data.

#5 Icecube

Icecube

    Gold Member

  • Team Reboot
  • 1062 posts
  •  
    Belgium

Posted 26 June 2011 - 06:06 PM

I think the problem with resizing to the left, is that the program that resizes the partition needs to understand each filesystem very good.
Because, without moving the actual files, all metadata records (MFT for NTFS) needs to be modified (location of the files is relative to the start of the partition).

All NTFS defragmenter programs, use the Windows API to do their lowlevel defragmenting work, so Microsoft took care of the filesystem support. AFAIK, there isn't a Windows API for resizing filesystems.

#6 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 26 June 2011 - 06:54 PM

Possibly, this is also the reason, why Data Recovery soft demands to COPY found data to a different drive instead of merely recreating access references (MFT Records) to all found files (which entales no risk to loose anything - especially when a user doesn't value the lost data to worse a new drive to buy). The devs all rely on MS to do the job instead of them, but charge hefty fees for "data recovery", which looks less disturbing when compared to the price of a new HD they demand a user to buy for recovered data. I attempted once to discuss the issue on Data Recovery forums, and NONE developer (despite being all frequent their) commented on it with anything particularly relevant to the discussion.

A lot of scamming is always going on in programming field covered by "user interest or safety" talk, even when a user expresses what his interest actually is. Plus no-one (among Partition Manager devs in this case) wants to face with MS "patent rights" issues following NTFS reverse engineering.

#7 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13594 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 26 June 2011 - 07:01 PM

Maybe it's a conspiracy. :thumbsup:

:thumbsup:
Wonko

#8 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 26 June 2011 - 07:06 PM

No need for conspiracy, when actions are driven by common economic laws - all individuals behave in similar fasion without ever knowing each other, or copy actions of each other while understanding motifs behind them being in the same business. On the other hand, there is nothing unusual or uncommon about conspiracy - antitrust laws in every country is clear evidence conspiracies driven by common fiscal interests are frequent enough to prompt such legislation, while usually hidden (but not their consequences) from a common man (how you can tell that about Wonko?!). :smiling9:

I have used Paragon's Partition Manager to resize partitions many times. Never had any problems. Never worried about how long it takes.

I already shipped back to Seagate a couple of HDs on warranty for replacement, paying 3-times higher S&H for their mandatory carton box compare to postal prices. This is enough motif to worry, what a Partitioning tool actually does, when it comes to a full 2-3 Tb drive.

#9 paraglider

paraglider

    Gold Member

  • .script developer
  • 1707 posts
  • Location:NC,USA
  •  
    United States

Posted 26 June 2011 - 08:49 PM

Are you saying that resizing a partition resulted in your hard drive dying. Sounds unlikely to me. Hard drives die. Its a fact of life. They may last a week or they may last five years. No way to tell how long they are going to last.

There is evidence that 2TB hard drives are less reliable than 1TB. Probably even worse for 3TB.

#10 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 26 June 2011 - 09:42 PM

Resizing a partition to the left (say you deleted a previous to it partition to add some space) causes in most Partitioning Tools senseless moving all data to the left. Given terrabytes of data, of course it affects HD lifespan, especially given poor reliability record. Did you notice that most HDs don't die on day one (there are exceptions)? Lower quality materials, wider tolerances, and buggy firmware all contribute to poor reliability, but limited resource exhausting is often a factor contributing to failure.

#11 paraglider

paraglider

    Gold Member

  • .script developer
  • 1707 posts
  • Location:NC,USA
  •  
    United States

Posted 26 June 2011 - 09:49 PM

Of the 3 hard drives I have had fail in the last 3 years - one WD raptor, one SAS 15K Fujitsu, one SAS 15K Seagate - the first 2 failed after a year. The third was a replacement for the second and lasted a month.

#12 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 26 June 2011 - 09:59 PM

A "replacement" drive is always a refurb drive, someone else before you already returned to the authorized repair agent of the manufacturer. While drive repair may entail different procedures depending on cause of failure, it usually seldom removes the original cause of failure - pure material quality, residual FW bugs, higher than specked voltage requirements, etc. Hence its not surprising to expect a replacement drive to die soon after - depends on its usage pattern too (including re-partitioning style). If you replaced a drive, but seldom use it - depends on what's on it - it will remain responsive longer due to simply sitting in your PC most of the time without spinning or moving its heads.

Frequent cause of drive failure - errors when writing to Drive Service Zone. Such errors can't be corrected on a regular PC, since that zone isn't accessible. Despite each drive has multiple duplicates of records in that zone, due to FW and HW bugs errors in a bad drive's service zone gradually grow to the point when undamaged record is hard to find - resulting in ongoing "clicking" without reading the drive. Another cause of clicking - too low voltage drop to start spinning at boot time, failure common in case of multiple drives in a single PC. Or a drive head mechanism is damaged. Many other causes of failure exist, but one thing is clear - don't move huge data massive without real need to extend your HD service life.

#13 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13594 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 27 June 2011 - 06:50 AM

Despite each drive has multiple duplicates of records in that zone, due to FW and HW bugs errors in a bad drive's service zone gradually grow to the point when undamaged record is hard to find - resulting in ongoing "clicking" without reading the drive. Another cause of clicking - too high voltage required to start spinning, failure common in case of multiple drives in a single PC. Or a drive head mechanism is damaged. Many other causes of failure exist, but one thing is clear - don't move huge data massives without real need to extend your HD service life.


Wow. That's a nice explanation, containing a lot of new things I didn't know, but you forgot voodoo/black magic :cheers: which is probably a far more frequent cause. :blush:

And another cause, that is very frequent with "portable drives", in USB or Firewire or eSata enclosures, drops/falls:
http://www.msfn.org/...internet-noise/

For NO apparent reason, Rule Eleven:
Spoiler


:ph34r:
Wonko

#14 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 27 June 2011 - 06:26 PM

Wonko

I am aware of your character issues... :) But don't want to go that route, unless you make me to. :dubbio:

Prove, a single word from the above citation is wrong. On the other side, I can site feedback & advice of leading HD repair experts - the developers of hardware and software used worldwide to repair hard drives. Are you among them? :yahoo:

Of course, HDs are made to be used, and enterprise HD versions are designed to be intensively used 24/7 (yet they frequently fail too due to pure make quality). We are not talking about these, but about consumer HD versions prevalent among Reboot members.

#15 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13594 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 27 June 2011 - 07:01 PM

Prove, a single word from the above citation is wrong.


Why? :unsure: I don't get it. :dubbio:

I just posted to express my appreciation for your list and to add a possible additional cause, quite common BTW.

About :cheers: really, the issue, is, as I see it, that in a number of cases you don't really *know* what happened, (example):
  • you (unfortunately) have a failed drive
  • if you have been wise enough to backup properly you have no problems about data
  • then you try to re-condition/repair yourself the disk drive
  • BUT you can ONLY do some very basics tests as you miss the actual tools (or knowledge/specific disk or controller data, etc.) so, unless the fail is due to something VERY evident and anyway happened "on the outside" of the disk, you don't really know WHAT happened <- this is :wodoo:
  • since you already have the data safe, if you don't succeed, you just toss the stoopid failed drive and buy a new one, so the actual cause remains a mistery.<- this is :wodoo:
  • OR, for any reason, you desperately need some data on the failed drive, then, after having tried (with the above mentioned lack of proper tools/knowledge/data/etc.) some "common sense" fixes, if you don't succeed you need to go to a professional Data Recovery firm
  • let's say that you are lucky (TWO times, i.e. that you find a good data recovery firm AND that they manage *somehow* to recover your data): they can tell you WHATEVER they want about what was the cause and which steps they took to recover your data <- this is also :wodoo:
  • if they cannot get the data back, they can as well tell you WHATEVER they want about what the cause was and why it isn't possible to recover the data <- this is also :wodoo:

Rest assured that I have seen more than one case of Data Recovery firms telling the poor customer among the tallest tales I have ever heard :w00t:.

The known Seagate 7200.11 hiccup shows this clearly.

To be really picky :) , I can find something in your post that doesn't sound "right" to me :ph34r: , I would say "not enough current available to break bearing stiction at spinup" instead of "too high voltage required to start spinning", but it's not anything of relevance :).

On the other side, I can site feedback & advice of leading HD repair experts - the developers of hardware and software used worldwide to repair hard drives. Are you among them?

Not "leading" and not "developer", but maybe (and modestly) more "expert" than the average Joe. :yahoo:


:cheers:
Wonko

#16 sambul61

sambul61

    Gold Member

  • Advanced user
  • 1568 posts
  •  
    American Samoa

Posted 27 June 2011 - 07:46 PM

I was lucky enough to stay away from Data Recovery firms. But loosing a drive is not constrained only to data recovery. You loose investment expected to work at least 5 years, considerable compare to PC cost. And it tends to be repeated with every 3-d drive. When Seagate moved HD production to Vietnam and CN, their execs ripped the benefits of low cost manufacturing, while holding same retail prices. And their customers enjoyed all drawbacks of untrained and unexperienced workforce, and 2-d class production facilities with equipment unable to deliver required part tolerances. The same was true about HD components and raw materials they get from their local suppliers. Add on top of that, the most experienced product design team members were, as often happens, replaced with recent grads and contractors to cut on salaries at each merger and production transfer. You have a recipe for disaster, someone "in-the-know" will call "black magic".

How Seagate solves the dilemma? They offer SeaTools Diagnostic Tool that can't access the drive if it doesn't spin. And then refuse to provide warranty repairs, unless you submit a diagnostic code generated by that tool (NONE). If you are unusually insistent, you'll eventually be able to ship it back by overpaying on S&H (i.e. paying repair costs as S&H on warranty drive). But the drive you get back is a refurb (not even yours original) that won't work beyond 1-month residual resource limit. Want to repeat the same - go on... I don't think, such company execs failures should be quietly omitted in users feedback. Not everything is about data recovery. A user did suffer from overpaying for a low quality cheap product, and by electing recovery route stimulates the maker to keep doing even worse.

I mentioned drive failures to illustrate that under such circumstance smart HD lifespan resource management becomes a major issue. That would include developing Partition Management tools that do their elementary job without stressing the drive beyond believe totally unnecessary. Its also in line with energy saving GREEN trends. Hopefully, partitioning tool developers will join the discussion as invited. Even if they elect not to, they will still read it for sure along with plenty of their customers.

#17 Wonko the Sane

Wonko the Sane

    The Finder

  • Advanced user
  • 13594 posts
  • Location:The Outside of the Asylum (gate is closed)
  •  
    Italy

Posted 28 June 2011 - 07:42 AM

Well, if you use NTFS or similar journaling or semi-journaling filesystems you have anyway a very relevant number of write accesses to the HD even if you never "move" data around.

If you care for this you could remove the file access timestamping, that by itself will save quite a number of write operations (in the long run possibly many more than a "one time every x days/weeks" defrag or move, of course depending on the use that is made of the PC).
http://www.pctools.c...stry/detail/50/
http://cyberst0rm.bl...-timestamp.html

:smiling9:
Wonko

#18 o.k.

o.k.

    Newbie

  • Members
  • 22 posts
  •  
    Hungary

Posted 11 February 2014 - 08:47 AM

Hi!

 

I need some help.

 

I resized a partition from the left with acronis disk director, without moving the whole data. Later I tried to move it left with paragon partition master. When i noticed that it want to move the full content, and it takes too much time, I cancelled the operation. Now it seems all data lost. I tried to recover with active partition recovery. It finds two partitions, and shows the files for the first one, but when I recover the files, they are corrupt.

 

Thanks for reply.



#19 Wally Bass

Wally Bass
  • Members
  • 4 posts
  •  
    United States

Posted 30 October 2014 - 06:24 PM

With respect to resizing partitions on the left, versus on the right, I'd like to comment:

As of the new "BCD store" booting introduced in Vista, "specifying a partition" (in BCDedit, the utility which one uses to maintain the "BCD store") has become somewhat strange, in my view. The "old" method, as per entries in the XP BOOT.INI file for NTLDR, was to use an "ARC" specification like "multi(0)disk(0)rdisk(0)partition(3)", which effectively specified the partition by giving a "navigation path" to the partition's entry in the partition table (i.e, 'PARTITION(3)' is whatever is pointed to by the 3rd "primary" partition entry in the MBR partition table). "/dev/sda3" type specifications under Linux are similar in nature.

But, but with BCDedit (the tool for maintaining the "BCD store"), you do it by specifying something like "partition=D:". That is, you give the letter currently assigned to the partition, at the time BCDedit is run. Obiously, specifications like "D:" are context sensitive, so "D:" can't be (and isn't) what is actually stored in the "BCD store", and some tranlation is done by BCDedit. So what is it that is actually stored in the "BCD store"?

For MBR partitioned drives, what is actually stored as a "partition ID" has two components. The first component is the 4 byte "volume signature" of the containing disk, as found at offset 1B8h in the MBR. The second is the offset (8 bytes) of the partition, in bytes, from the beginning of the disk. The same scheme was/is used to identify partitions for drive letter assignment purposes, in "HKLM\SYSTEM\MountedDevices", as far back as Win 2K (NT has a similar, but different in detail, scheme). Let me call a reference based on these two components a "REG_ABS_PTNREF."

Why this is important (even if Microsoft never mentions it), is that the form of this reference determines the conditions under which a REG_ABS_PTNREF is invalidated. A REG_ABS_PTNREF which is invalidated can cause your system to be unbootable, and, in the case of Vista "HKLM\SYSTEM\MountedDevices" references, can result in the "Windows system drive" getting assigned a letter other than "C:" at boot time (which causes quite a large amount of subsequent trouble). (In Vista, if the "MountedDevices" references are invalid, the "boot partition" will "get "C:", forcing the system partition to get "D:". In most cases. Win 7 doesn't have this problem, because the boot partition doesn't get a letter by default under Win 7.)

So, basically, a very important thing to remember is that resizing a partition "on the right" does NOT invalidate any REG_ABS_PTNREF type references, whereas "risizing on the left" always invalidates any REG_ABS_PTNREF references to the partition that may exist. That is probably another reason that, in providing a resize function, Microsoft doesn't provide function to resize them "on the left."

Wally Bass
 



#20 C0m4

C0m4
  • Members
  • 1 posts
  •  
    Israel

Posted 14 November 2015 - 03:16 PM

I've just accomplished the task with EaseUs Partition Master 9.2.2 Free Edition. Took about 3 minutes to move a 500 gb NTFS partition to the "left" for 50 gb.






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users