• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 698
  • Last Modified:

Do I need to worry about my HDD & if so, what HDD of the ones listed below do you recommend?

I have Windows 7 Pro 64-bit with 8 GB DDR3 RAM on a Gigabyte motherboard.  

I also have:
1. a 7200 RPM Seagate SATA2 500GB HDD (16 MB cache),
2. a 7200 RPM 500 GB Western Digital SATA2 Caviar Blue (16 MB), and
3. a 5400 RPM 1 TB Western Digital SATA2 Caviar Green (64 MB cache) hard drive.

HDD 1 above is he OS drive.  HDD 2 above is for downloads and special projects.  HDD 3 above is the backup drive.

Recently, I had a HDD failure coupled with a backup failure, losing a lot of files in the process.  It was a Seagate drive (see http://www.experts-exchange.com/Networking/Network_Management/Disaster_Recovery/Q_25056767.html).

I replaced the old Seagate with HDD 1 above, which had been the backup drive.  Ever since I got this hard drive, I have been able to hear sounds coming from it, like crunching ball bearings.  I used the Seagate software (Seatools) and ran each of the tests offered. The HDD passes all tests.

I then used WD Data Lifeguard tools to check the HDD 1.  Strangely, it lists the HDD as an IDE, rather than SATA.  The HDD passes the SMART status, but any other test it fails.  In fact, all HDDs fail the WD tests, but pass the Seagate tests (error received for each HDD:  "Cable Test::Write sector error!").

My problem is this:  Windows 7 backup of the OS HDD occasionally fails (once or twice per week) due to the disk being corrupt, with the Action Center suggesting I run Chkdsk.  And each time I restart my computer, the chkdsk runs, corrects errors and then the backup is fine for a few more days.

The specific error associated with the failed backup as listed in the Event Viewer is:

Source: Ntfs, Event ID: 55, Task Category: (2), The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume C:.

I have run chkdsk on the other HDDs, and chkdsk states that they are clean.

I also get the following error often:  Source: Disk, Event ID: 7, The device, \Device\Harddisk1\DR3, has a bad block.

I have included a couple really long images showing most of the errors listed in the Event Viewer.

My questions:
(1)      Do I need to get a new hard drive for the OS (HDD 1)?
(2)      If I do need to get a new HDD, would you recommend WD6400AARS, WD5001AALS, WD6401AALS or WD6400AARS (or any other) for the OS drive?
(3)      Lastly, what do I need to be able to transfer the image from HDD 1 to whatever new HDD I get (i.e., I do not want to be reinstalling Windows and all the software I have…it takes days to do so)?

Thank you,
edsager
2010-04-15-003840.jpg
2010-04-15-004042.png
0
edsager
Asked:
edsager
  • 14
  • 7
  • 4
  • +3
2 Solutions
 
BawerCommented:
You need to have a HDD i recommend you buy a Free fall sensor disks which are much more better than old disks,,,
0
 
edsagerAuthor Commented:
Bawer: This is a desktop computer.  How would free fall sensor disks help?

The reason I was seriously considering the WD Caviar Black is because it is supposed to continuously look for and repair errors.

Thanks,
edsager
0
 
SnibborgOwnerCommented:
I must admit I have always had issues with WD drives in the past.  I've never had a failure with a Maxtor.  however, with the consolidation in the HDD industry it's had to tell who makes drives under what names.

My recommendation is that you swap that drive out ASAP.  Use Acronis or Ghost to take a backup of the drive, I use both quite successfully, having used Ghost last week to replace a hard drive on a IBM Thinkpad.

You might want to check the temperature of the suspect drive.  If, as you suspect, it is a bearings problem (the most common moving part issue) it will cause a large increase in heat given off.  I had a WD Caviar that had bearing problems and believe me you'll know if it's overheating.  Just be careful not to touch it!

The other issue, which is by far the most common, is that it may well be an electronics problem caused by a fault on the controller board.  This is more likely in the light that the problem is intermittent.  Mechanical faults do not tend to manifest in this way.

Let us know how you get on.

Snibborg
0
Free Tool: Site Down Detector

Helpful to verify reports of your own downtime, or to double check a downed website you are trying to access.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

 
davidlevans13Commented:
I would get the WD Caviar Black 640GB, or a Seagate 750 7200.12 and just do a straight up clone from old drive to new one.
Then you can just extend the partition or make a new one with the extra space. I have done this several times using a Parted Magic disk and the terminal function in it....   by this method...  http://www.justlinux.com/forum/showthread.php?threadid=134457

Or you could use Casper...  haven't used it myself but have heard good things.  http://www.fssdev.com/products/casper/

Or any other cloning software you feel comfortable with. Everybody seems to have a favorite.
0
 
DavidPresidentCommented:
Let me explain a few things first:
 - S.M.A.R.T. tests are predictive failure tests only .. It is designed to tell you if the disk is in a degrading condition and is going to fail "soon".  It is not infallible.  A dying drive can pass a S.M.A.R.T. test, and a perfectly good disk can report a false error.  
 - Bad blocks do not necessarily mean a disk should be replaced.  A large drive will have THOUSANDS of spares.  
 - It is foolish to reject or embrace a HDD manufacturer based on anecdotal information from some nameless blogger about good or bad experiences they have had with a particular vendor.  First, they all ship MILLIONS of disks per month, which includes both cheap consumer and high-end enterprise disks that go for over $1000/drive. Go to anyplace that discusses HDDs, choose any manufacturer, and you will find people that say that this particular manufacturer should be avoided, and you will also find people who say that they are the best ones out there.
 - Chkdsk is NOT a media test. It is a file system test.  A drive that fails S.M.A.R.T. and/or media tests can pass chkdsk, and vise-versa.  It should never be used as a diagnostic tool beyond the context of repairing file system corruption, and discovering a small percentage of corrupt data files.  

Now you do get bad blocks.  The best way to prevent this is to never turn on the computer.  The 2nd next best way is to get premium HDDs that have 100X better ECC so they can auto-correct.  Third best way is to go with RAID1.  With RAID1, an unrecoverable read error on block #X, will be repaired automatically because the RAID controller/software will read that block from the good HDD and then repair the bad block on the other one.

So if you want to make these problems go away, invest in RAID1, and buy enterprise class disk drives.
0
 
SnibborgOwnerCommented:
Or if you can't afford Enterprise disks invest in Spinrite: http://www.grc.com/sr/spinrite.htm.  It's a better test than using S.M.A.R.T.
0
 
edsagerAuthor Commented:
Snibborg:  Thank you for your suggestions.  The HDD with which I am having problems is the Seagate one.  I have had problems with Seagate HDDs and not WD.  I guess we all have our own experiences, so my desire is to stay away from Seagate (my superstition).  

The temperature of the HDD seems to never get beyond about 40 degrees C, with the normal operating temperature being about 32 degrees C.  

I am not sure if it is a ball bearings problem, but it makes a sound when in use which sounds like ball bearings scratching against each other. Another description of the sound (which I have heard since purchasing and installing the HDD) might be a muted clicking sound.  Or if you go to http://www.datacent.com/hard_drive_sounds.php, the sounds I hear from my HDD are most similar to the following sounds:
“Hitachi 60GB laptop drive with stuck spindle can't spin up, makes humming/buzzing noise”
“Toshiba laptop hard drive with stuck spindle trying to spin up(heard if taken close to your ear)”
“Fujitsu 40gb desktop drive with bad media making scratching noise.”

If I could upload an audio file, I would record the sounds of my HDD…if that would help.

Spinrite sounds like a good product, but it has not been updated in years.  I wonder if it would be compatible with Windows 7 64-bit....

dlethe:  I thank you for your suggestions and information.  I like your idea of RAID 1 and my motherboard supports SATA RAID 0, RAID 1, RAID 5, and RAID 10.  If I do the RAID 1, do I need to have the same HDD as the 2nd drive, or do I only need to have the same size of HDD?

As far as Enterprise level HDDs, they sound great, but appear to be a little out of my price range.  This is why I was considering the WD Caviar Black HDDs, which also come with a 5-year warranty, rather than the normal 3-year warranty given on other non-enterprise level HDDs.  Would this be a good 4th option (use RAID 1 with Caviar Black)?

So, if the chkdsk only checks the file system, what might be causing all the errors in the file system and what can I do to not have those errors?

Does anyone know if the Caviar Black drives might reduce these errors (“WD's Data Lifeguard technology automatically finds, isolates, and repairs problems that may develop over the life of a hard drive.”)?

Thank you,
edsager
0
 
edsagerAuthor Commented:
davidlevans13: Thank you for the information.  

It sounds like all I really need to do to clone the disk is have some sort of disk imaging software (e.g., Acronis TI, GBM, Windows 7 backup, etc.), I use the software to make an image of the HDD, then...this is where I am stuck.  How do I get the cloned image to be on the new HDD and for that new HDD to be the boot disk?

Does the cloning come from the HDD software?

Thank you,
edsager
0
 
edsagerAuthor Commented:
dlethe:  One more question:  if the HDD I have only suffers from file system problems, can I use it as the 2nd disk in the RAD 1 setup?  Or should I get 2 new HDDs?

Thank you,
edsager
0
 
edsagerAuthor Commented:
Oyks!  It seems as though I need to prepare the disks for RAID prior to installing the OS.

That sounds like I would need to make an image of the HDD, set up the RAID array (which will format the HDDs), install the OS, and then restore the imaged HDD.

Does this seem correct?
0
 
SnibborgOwnerCommented:
I believe that spinrite works with all current systems, but I would drop them an email and ask them.  Just because its an old system, do not dismiss it.  Disc technology has not fundamentally changed since the 1970's.

My main desktop computer has the following setup - 5 x 500Gb hard drives.  2 are set up as mirrors for the operating system and downloaded utilities, etc.  The RAID 5, which takes up the rest of the system holds my main data.

I am, however, paranoid.  If you read my bio you'll see why.  All the disaster recovery will have that effect.  As a result I have an external Buffalo RAID5 drive that acts as a backup to the internal RAID 5.  I use KarenWare's Replicator (http://www.karenware.com/powertools.html) to keep a mirror of my data on it.  This happens overnight each night.

In the end I decided that the this was my best alternative.  Having said that, I have had 2 drives fail on me - 1 Maxtor (Mirrored) and 1 Seagate (RAID 5).  I have never lost a byte of data with this configuration, but I did come close.

When I set this up, disks were much more expensive than now.  I have a new hitachi 2TB drive on my desk at the moment that cost just over 100 GB Pounds.

When choosing the drives you are going to work with, make sure they are all the same - in my case 500Gb, with 3mb cache.  For kicks, when I first got them, I tried mixing and matching them to see what happened.  Despite them all being the same specification, when I put the 2 Maxtors with one of the Seagates, the motherboard would not create the array.  I use an Asus Striker Extreme motherboard.

Sorry if I have gone on a bit, but I thought you'd like to see how I approached the problem.

Good luck, whichever your choices and remember we are always here.

Snibborg
0
 
davidlevans13Commented:
I have never worked with a RAID array so I don't know the steps necessary for transferring a system image to that array.
But I am sure there are lots of folks around here s=who could direct you to a tutorial about that. I find it beneficial to Google what you want to do and then read several articles to get a feel for what is needed to get your desired effect.
0
 
edsagerAuthor Commented:
Thank you all for your comments.

I guess I am still wondering if I need a new hard drive based upon what is happening with the hard drive now.  It is still under warranty, so if I get a new drive, I want to eb able to point to some objective evidence that the drive needs service or replacement.

Anyone help with that?  Snibborg suggested Spinrite, and the Seagate and WD software give conflicting results.  Any software with a free trial or for free which you would recommend to get an objective assessment of the condition of the hard drive?

Thank you,
edsager
0
 
DavidPresidentCommented:
You need to spend your money on enterprise disks and implement RAID1.  This will not only protect against a drive failure, but more importantly, it will auto-correct bad blocks. The enterprise disks also have 100X better error correction circuitry, which is statistically significant once you have a TB or more of data.

Spinrite is a software package that MAY be able to correct an unreadable block. However, by the time this happens, it is too late.  Your O/S isn't going to instantly wait for you to try to correct a bad block, RAID1 does this for you and will prevent data loss. (Any RAID level with parity will as well).  

A few bad blocks will PASS diagnostics.  SMART is an indicator of impending failure.  Bad blocks are not.

Seatools is free, and will run the embedded diagnostics built into the firmware, and you could also run a media check, but that will tell you nothing you don't already know.  Bad blocks are a fact of life, and not an indication of doom.  If you want to prevent data loss as a result of bad blocks, implement RAID1.   If you want to minimize the frequency of unreadable/uncorrectable data, get an enterprise class disk.
0
 
ocanada_techguyCommented:
Applause @ dlethe  Excellent.

a bit more trivia to add:
Technically checkdsk has a command line switch for checking for bad blocks, but just read the description of SpinRite at the GRC website and you'll know right away why the experts are telling you chkdsk is for fixing the file allocation and NOT for badtracking, cause it SUCKS at recovering anything, so it's really only ok when formatting, and even then not very thorough.
SpinRite is great and I can highly recommend it.  That said, I once had a bug with SpinRite trying to handle a particularly large drive with excessive bad sectors.  You could also look into HDDRegenerator.
If you're going to redo the drives preparing them for RAID is fine, but might I suggest that you actually thoroughly bad sector scan them just before that.  So yes, get a full good backup, 'cause the  plan calls for reformatting  the drives.
The various brands of drives have diagnostic tools put out by the drive manufacturer.  ALWAYS use the diagnostic tool SPECIFIC to the MANUFACTURER (and model, such as IBM/Hitachi Think/DeskStar unique)  So, your WD Tools on Seagate drive and seatools on WD drive not working, is NOT surprising and NOT indicative of anything too telling.  These tend to be none too user-friendly, command lineish, and can be somewhat difficult to deal with, maybe hard to tell the difference between a non-destructive and desctructive (kiss your data goodbye) scan.  DIagnostics can check your firmware, provide an avenure to update it, test the logic board, etc.  
SpinRite is pretty easy, manufacturer independent, although clearly it CANNOT test your firmware or logic board as that would be manuf specific tests.  The nifty thing about SpinRite is it can be used as a PREVENTATIVE measure.  If you scan your drive with it once a year or so, it will set aside questionable sectors before they turn into bad sectors that put data in jeapoardy.
Most all drives auto badtrack.  The SMART parameters will show how much, and whether thresholds are being approached/exceeded.  The difference between enterprise and consumer drives is the extent to which they will do this succesfully, and the quality of the construction which equates to the MTBF (mean time between failure) rating of it's estimated life.  Yes the length of warranty can sometimes be an indication of the expected quality.  
Think of it like cars, different manufacturers have their lemons, even lemons have warranties but you still don't want a lemon; models, years. sometimes a bad batch, what have you.
Sometimes some models might not handle bad blocks the best.  Sometimes a firmware update may help. Sometimes the drivers written for Windows can be buggy.
It's entirely possible you may have your BIOS set to have your SATA controller in IDE compatibility mode, in which case the drives "appear" as if they are plain ol IDE, and Windows is using a generic IDE controller driver.  Does it happen that many people do this unknowingly as it's kind of the default behaviour, yes.   Could the controller probably operate faster in RAID mode, probably.  Make sure you have the correct drivers for your motherboard and built-in RAID or dedicated RAID controller card(s).
To be clear, SpinRite doesn't run simultaneously with your OS as some sort of add-on, it is a BOOTABLE disk and it actually runs under FreeDOS, bad tracking.  So you can see, when the bad blocks happen, the OS needs to deal with it, badly.
You were smart enough to have disk redundancy, the two 500G backup up to 1TB nightly, and so it seems like a fairly reasonable approach to instead have 2xmirror or better yet 4 drive raid5 or raid10.
The differennce between ON-THE-FLY deal with badsectors which are bound to happen over time of RAID versus "auto badsectoring" hope it was able to move the data versus occaisional scan with a tool is what it boils down to.
0
 
edsagerAuthor Commented:
Thank you all for your responses.  

This is my plan of action (please comment if I have this wrong, or even if I have it right):

So what I see as an optimal choice would be to get Enterprise level HDDs (on the way now), bad sector scan drives to be used in RAID, set up RAID array, then use whatever software I used to image the hard drive to restore the image (I generally use Windows 7 backup, but also have Acronis TI, GBM, and a few others) to the now blank HDD, which means that the image will be restored to both drives on the RAID array?

Thank you,
edsager
0
 
SnibborgOwnerCommented:
Thats how it will work, yes.  Once the drives are mirrored, the restore will autpmatically go to both disks.  

Snibborg
0
 
DavidPresidentCommented:
Actually the SpinRite FAQ is technically correct, but some things are misleading.  The biggest one is that it dances around the scandisk by acknowledging it WILL do a media scan, but then says that they have a better one.  Fair game, but you will also notice that they do not do file system repairs, which means you still need to run the MSFT code.

So, the microsoft scandisk/chkdsk will scan & remap bad blocks, and fix file system as best as possible.   Spinrite will attempt to recover bad blocks, but then you have to run the MSFT software to perform filesystem repairs.  Many times it is not unusual to have unreadable blocks where there is no data.  This is actually the norm.  

As for the HDD manufacturer's tests .. there are a lot of dirty little secrets about those, and some of your statements are incorrect, but not for the reasons you are aware of.  Just more reasons for me to write a whitepaper.  But anyway, the manufacturer's self-test programs are part of an ANSI specification.  Everything is standardized from the bytes you send to the drive to kick them off, to resulting data structures.   Only reason you want to run the SEAGATE tests on seagate disks is that their software will ignore a non-seagate drive.

There are a few vendor-specific things, but the freebie consumer software generally doesn't run them.

0
 
edsagerAuthor Commented:
What about using the tools found on UBCD?  I am not averse to purchasing some software to to scan for bad disks before setting up the RAID array, but is using the Western Digital Lifeguard will not cut it, what will?

By the way, I have chosen to go with the WD RE3 500 GB.

Again, thank you all for your responses,
edsager
0
 
edsagerAuthor Commented:
I mean I am not averse to spending money to scan for bad SECTORS, not bad disks, as indicated above.

I am guessing chkdsk /r would not be sufficient.

Thank you,
edsager
0
 
DavidPresidentCommented:
the windows scandisk DOES detect bad sectors. It is one of the 2 checkboxes that defaults off.  It does not attempt to read the HDD and repair them, like spinrite does, but one of the things I am saying is that under normal situations odds are higher that the sector in question is not in use in any way.
0
 
edsagerAuthor Commented:
Well, it did not go as planned.  

I made a sector-by-sector image, set up the RAID array and tried to restore the drive image to the RAID array.  

Well, I did not know that the RAID array needed to have 100MB on each disk for miscellaneous purposes, so the sector-by-sector image was the exact size of the HDD, and thus, could not be restored.  

Should I not have done a sector-by-sector image?

Thank you,
edsager
0
 
DavidPresidentCommented:
If you were trying to take the usable area from one RAID environment, and copy to another, and the NEW controller uses less metadata then the former, then below would have worked..  You needed to do a sector image starting at the usable area of the HD.   Example, say the metadata takes 100 (512-byte) blocks.  Total capacity written on the disk is 1000 blocks (again, to make math easy).  So the RAID controller tells you the total on the disk is 900blocks (1000 - 100).

The RAID array presents PHYSICAL DISK block #100 as your LOGICAL block #0, on your 900Block disk drive.  

Had you imaged block #100 - 899 (with this example), to physical block #0 of a non-RAIDed physical disk, then your O/S would boot OK, and say you have a 1000 block disk, with  100 blocks unused.

Now this also assumes that there is no metadata at the tail end of the disk, and, of course, you have to figure out the logical/physical mapping.  Too difficult to walk somebody through the process, so  buy someting like Runtime.org's reconstructor to do this.  

=====
A more elegant technique would be to take an image backup using the LOGICAL drive i.e, C:\, with something like acronis or ghost, that supports both controllers,  then build the new array, format it, and restore to the new array.  No sector-by-sector imaging as I described above, it is too hard for most people to  get right.

0
 
edsagerAuthor Commented:
dlethe:

Thank you.  

I did not go from one RAID environment to another; rather, I went from 1 HDD and setting up a brand new RAID array, then tried to restore the sector-by-sector backup of the C: drive to the RAID array.  Currently, I have no RAID setup on my computer.

Let me see if I have it straight:  make an image backup of the C: drive with Acronis rather than what I used (EASEUS), then create the new RAID array and then restore the image to the new array.  

But does not Acronis make a full image the entire disk, whether parts are used or not?  Will I not run into the same problem with the backup image being slightly too large for restoring to the array?

Thank you,
edsager

P.S.  Acronis scares me as I have not had good experiences with it in the past.
0
 
DavidPresidentCommented:
You need to treat this as if there was no RAID at all (mostly).  Use a bare metal backup product such as acronis, that can see the "C:\" drive when it runs.  Create the backup.   then create the target disk and format it (with the RAID controller), so your backup software sees a "D" drive, or a different "C" drive.  then restore.

By going down the path of physical sectors, you destroyed metadata, which is vendor/product/controller/firmware unique.   Also by using  physical sectors, you technically have incorrect  partitioning information as it relates to size of the physical disk and each partition.   The backup software will properly handle restoring to different make/models of disks and capacities.

So as long as your backup software can see the logical disks when they run, on both controllers, then you will be fine.
0
 
ocanada_techguyCommented:
You had it straight what I was saying yes.  No I was not suggesting a raw block-by-block exact image copy like a RAID mirroring would do, but rather more of a BACKUP/RESTORE/COPY, like gHost does (ghost saves time and keeps copies/images small by only copying data not the entire drive empty or no)  Whether you use the disk backup feature with Acronis True Image, or Acronis Disk Backup or any of their many products that do so makes little difference, doing the correct KIND of operation with it does (I suggested the product with the free trial)

For a teeny bit more insight into RAID versus RAID, you may understand more clearly if take a look at the transcript of Security Now! with Leo Laporte (tech tv, The Lab, etc) and Steve Gibson (GRC, SpinRite, instrumental in embarassing MS into making firewall in SP2, hotfixes automatic, etc)
Straight from an experts' expert.
http://www.grc.com/sn/sn-112.txt
Look for RAID

I highly recommend SpinRite 6 for recovering data on bad sectors, but a) it's not a free tool b) if your RAID hardware / auto-correcting drive already "handles/ed" the bad sectors under the blankets, either it did so well/sucessfully or poorly/it corrupted the file while eliminating the bad block; either way may be after the horse is out of the barn as far as SpinRite recovery goes, in which case i) recover said files from backup/restore or else Windows file recovery/repair and/or ii) check your RAID logs/diagnostics, is a hard drive failing/about to fail more than RAID can reasonably deal with iii) do you need a better RAID level/redundancy?

Have I mentioned enough: "have good backups and verify them once in awhile"?

Not sure if you can read other EE posts (possibly not) but here's a different problem that has some explanations that somewhat relate to your problem.  http://www.experts-exchange.com/OS/Microsoft_Operating_Systems/Windows/XP/Q_26022925.html
0
 
edsagerAuthor Commented:
OK, now I am a little lost.  

EASEUS backup software saw the C & D drives, even after creating the RAID array.  However, it would not restore the backup because the backup image of the pre-RAID C drive was 100 MB too large due to the RAID disks each having 100 MB used for RAID purposes.

So, even if backup software properly handles restoring, have I caused my disk some harm by using the sector-by-sector backup and therefore, destroying metadata?

As far as wrong partitioning for the disk and partition, both disks are the same size (or they report to have the same size) and each has not been partitioned, except for possibly by the OS.  After configuring the RAID array, each disk also lists the same size and number of sectors.

I guess the main question I have is:  what do I do when the image of the pre-RAID drive is 465GB and the capacity of the same HDD post-RAID setup is 465GB less 100MB?

Thank you,
edsager
0
 
DavidPresidentCommented:
You either get a bigger HD then repair the partitioning manually (difficult unless you know what you are doing); or do file-system level backup that incorporates sufficient intelligence to resize partitions and not back up empty space ... and make sure that you have 100GB free space in the filesystem before starting out.  

Sorry, I thought you meant earlier that the older array was 100MB larger so you would have 100MB worth of empty space.
0
 
ocanada_techguyCommented:
$^&* that's an problem eh
I don't know whether you did a sector-by-sector copy as you think you did, or it did a partition image backup, BUT you could have encountered this problem either way.
There's a big difference between 100Gigabytes and 100Megabytes, >I< read quote "MB" unquote.
I suppose it's possible the RAID set aside 100MB for a larger bootstrap or something, but frankly it is MORE likely that in prepping the drives it found BAD sectors which account for your now having 100MB less space to work with.
Let's say you were backing up a 500GB hd that was 50% full.  In the good-ole XP days, gHost would need 250G to back it up, and gHost could recover it onto drives smaller than 500GB, as long as it was big enough to hold the amount of full data.  In the good-ole days, a program like PartitionMagic could shrink that 500G partition down to 250+ as long as there'd be enough to hold all the data, and PM could move the partitions around by doing sequential copy-paste with tempspace.  If your backup was 250G you could restore it on 250G.
HOWEVER, in the Vista/Windows7 days, with VSS (volume shadow copy) and additional metadata added in, the OS now chooses to put data in the middle and near the end of the partition on purpose by design.  PartitionMagic no longer works and Symantec who bought it say it likely won't ever be updated.  Vista/Windows7 does have built-in ability to shrink/expand partitions right in Disk Management now, BUT, it cannot move, cannot push the start of a partition over to make space for a preceeding partition to be expanded.  The built-in disk image backup included in Vista/Windows7 refuses to restore to smaller, will only restore to the same or larger.  The shrink operation will only reclaim empty space AT THE END of the partition, and since things are put near the end on purpose, that's often shrink not very much or zero.  You can read about it here: http://www.howtogeek.com/howto/windows-vista/working-around-windows-vistas-shrink-volume-inadequacy-problems/

Now, the problem is your backup insists on restoring to a drive at least as big as when it was made.

The solution is convoluted, be prepared to be underjoyed.
One route would be to UN-raid the drives, restore the backup onto one, REDO the backup so that it will fit on the intended target, then re-raid and restore.  HOWEVER, as I postulated, your 100MBs lost may be lost forever (or half of it on one half of it on the other since some RAID treat an error on one as requiring sparing/slipping on both), SO... you might not even be able to restore the backup you took back onto the drive it came from anymore.  If you could, then your next interim problem is to SHRINK the partition by at least 100MB.  Use Disk Manager on that Windows7 to do it.  If it should happen that Windows7 says it can shrink 0 (cannot shrink any), then, save yourself a ton of grief and do NOT go through all the rig-a-ma-roll suggested in the previous paragraph article reference but instead use the EASIER solution he recommends which is obtain the free trial of PerfectDisk11, defragment the partition, (use the defragment on boot option if necessary to defragment otherwise unmovable files) and THEN you should be able to SHRINK the partition, and back it up for eventual recovery on the RAID.
(People sometimes ask, hey, why did you leave a little bit of space on the drive unallocated, why didn't you use 100% of the drive for the partition?  Ta-daa, now you know the answer)
Another route may be almost the same, except you'll leave the RAID made like you have it, and instead you need to find a larger disk you can erase/use temporarily.  Recover to that, do the shrinking on that, then backup/restore or drive-to-drive copy the shrunken partition to the RAID, relinquish the spare drive, scrubbing it of confidential data if necessary.
Ugh.
Good luck.

0
 
edsagerAuthor Commented:
ocanada:

You gave me the solution to this last problem, I believe.  I did get rid of the RAID array because I need to use my computer.  I restored the image to the HDD after deleting the 100 MB partition on each of the 2 HDDs I was using.

But what I am getting from you is to use software which can non-destructively reduce the size of the partition, leaving at least 100 MB unallocated, then do the image backup (thus, preparing the HDDs for that), then create the RAID array and then restore the image onto the RAID array.

Yes?  If so, this is definitely doable, but will take some patience and time.

I have Acronis Disk Director Suite, which has done quite well for me when changing partition sizes non-destructively.

The backup software I used (EASEUS) has an option for sector-by-sector backup, which is what I chose.  I kinds hope it was really such a backup.

Thank you,
edsager
0
 
ocanada_techguyCommented:
Yup, you're right on the money
Ah ok, disk director suite eh, I'll have to find out if it deals with the Vista/7 natively unable to shrink issue
0
 
edsagerAuthor Commented:
Thank you.  I feel much more comfortable with what my options are.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Cloud Class® Course: Certified Penetration Testing

This CPTE Certified Penetration Testing Engineer course covers everything you need to know about becoming a Certified Penetration Testing Engineer. Career Path: Professional roles include Ethical Hackers, Security Consultants, System Administrators, and Chief Security Officers.

  • 14
  • 7
  • 4
  • +3
Tackle projects and never again get stuck behind a technical roadblock.
Join Now