|
General Data Recovery FAQ
|
What NOT to do when confronted with a data recovery
situation
My disk appears to be physically damaged, what can I do?
I have accidentally formatted a partition, what can I do?
I have accidentally deleted one or more files, what can I do?
I have lost or deleted one or more partitions, what can I do?
DiskPatch, iRecover or iUndelete: which tool to pick for the job
I will buy a license if you guarantee success
My hardware RAID 0 array is no longer detected
"Invalid system disk" (Windows 9x versions only)
Diagnosing and resolving NT boot issues
General data recovery procedures
Terminology: an explanation of terms used in data recovery and disk
usage
|
What NOT to do when confronted with a data
recovery situation |
These 4 golden rules
apply (until you know exactly what you're dealing with):
- DO NOT run "chkdsk" or "scandisk"
or any other partition repair utility
- DO NOT use the Windows recovery
console commands "fixmbr" or "fixboot"
- DO NOT attempt to recreate a
deleted partition with "Fdisk" or "Disk Management"
- DO NOT change the contents of the
disk or partition that needs to be recovered
If you need help diagnosing the problem please look here to find
out how to get started.
|
| My
disk appears to be physically damaged, what can I do? |
|
If the hard disk makes unusual noises (clicks, grinding,
scratching etc.):
STOP using it immediately and do not attempt any software
based recovery! Recovering data from a physically defective
disk is very expensive so consider the need for this type of
recovery carefully.
The hard disk spins but is not detected in the BIOS:
This can be caused by a number of things ranging from
defective IDE cables to physical damage:
- Make sure the drive is properly configured in the BIOS
(the AUTO setting should do well in most circumstances)
- Make sure the drive is properly jumpered
- You can try if the disk is detected using different
hardware; a different PC, different cables etc.
- Use diagnostic software like HD Workbench or a drive
manufacturer diagnostic tool; If HD Workbench sees the
drive see if you can clone it.
Sometimes swapping the drive circuit board with one from an
identical disk (type AND revision) can resolve this.
Drive doesn't spin up:
Consider if the value of the data outweighs the cost of
data recovery by an outfit equipped to deal with physical
defects (expertise, tools, cleanroom environment etc.).
Drive contains unreadable sectors:
Clone the drive immediately using DiskPatch.
If required repair logical structures on the clone or extract
data from the clone using a tool like iRecover.
|
| I
have accidentally formatted a partition, what can I do? |
|
Important: avoid writing data to the accidentally
formatted drive/partition!
Successful recovery of deleted files depends on certain
factors:
- The file system type that is in use (FAT or NTFS)
- The level of fragmentation
- Is all data still available (or already overwritten)
Background: when formatting a drive a large portion
of the administrative and meta information is re-initialized,
but not all of it. Whether recovered data will be intact
highly depends on the fragmentation of the file system: since
there's no way of telling exactly what clusters "belonged" to
specific files, recovery software will assume that the file
system was not fragmented and files were stored in consecutive
clusters. The effect of fragmentation on the quality of the
recovered files is higher for FAT(32) than it is for NTFS.
What to do: if the partition was reformatted with a
different file system, or the same file system using different
parameters (for example different cluster sizes) than this
will further complicate the recovery. When the program used
for the recovery allows it (as iRecover
does), specify exactly what file system you are "looking for".
So if for example a FAT32 drive was accidentally formatted
NTFS, tell the software to "treat" the partition as if it were
a FAT32 drive.
The only way to verify if files can be recovered intact is
by trying. In general, if the FAT(32) drive was recently
defragmented chances for a good recovery are high. Large files
have a higher chance of being fragmented and so have files
that are frequently modified. Again, for drives that were
originally formatted NTFS, this is less of an issue.
For this specific scenario we advise against the use of
tools that attempt "in place" repairs; that is, tools trying
to repair the file system. In general those repair attempts
are almost never successful and they can't be undone, which in
the end makes data recovery even harder.
Suggested tool for the job: iRecover.
|
| I
have accidentally deleted one or more files, what can I do? |
|
Important: if you have deleted or lost files,
immediately STOP writing to the partition that contained the
lost files!
Successful recovery of deleted files depends on certain
factors:
- The file system type that is in use (FAT or NTFS)
- The level of fragmentation
- Is all data still available (or already overwritten)
Background: when a file is deleted its actual
contents initially remain intact. However, in the partition's
book-keeping of clusters in use by the file, these clusters
are marked as free for use. The deleted file's name is
modified so the deleted file no longer appears in the
directory list. As it is no longer known exactly what clusters
"belonged" to the deleted file, the successful recovery of
files that were not stored in consecutive clusters is
uncertain. If fragmentation is high the chances of a good
recovery become less.
What to do: whatever tool you use to undelete files,
do NOT copy the tool to the partition from which the files
were deleted, and do not recover the deleted files by copying
them to another folder on the same partition!
Suggested tool for the job:
If you have additional problems with the partition, use iRecover.
If it's just a simple undelete job and the partition is fine,
use iUndelete.
|
| I
have lost or deleted one or more partitions, what can I do? |
|
Important: if you have deleted or lost partitions do
NOT try to recreate them using Fdisk, Disk Management, or any
on-board or third party volume management utility!
Normally you can repair this type of damage with DiskPatch
by rebuilding the partition table. Alternatively, you can
extract data rather than fixing the damaged disk using
iRecover to copy data from the deleted partitions to another
disk.
Please note the functional difference between the two
approaches:
- DiskPatch repairs structures on-disk, thus
changing the contents of the disk. Changes made by the
automatic repair can be undone.
- iRecover extracts files from
the disk, not changing the contents of the disk.
Which
approach to take depends on the nature of your problem (and
perhaps even personal preference). When posting a request for
support, we will suggest which
tool is best suited for the job.
|
|
DiskPatch, iRecover or iUndelete: which tool to pick for the job |
For the uninitiated it may be tricky to pick the right tool for the
problem you're currently facing. Let's see if we can make some
things a bit clearer.
Our three main tools for Data Recovery are DiskPatch, iRecover and iUndelete. There is some
minor overlap in functionality but each tool has
its own specific uses:
DiskPatch: performs actual
repairs on the (physical) disk
iRecover: lifts data from disks
that may suffer from all sorts of problems
iUndelete: undelete files, pure
and simple
Obviously this is the short version of the description but it will
give you an idea of which tool to pick for your recovery job. We'll
dive a bit deeper now, and see
what differentiates the tools.
DiskPatch:
This tool is aimed at working with a hard disk in general and puts
the actual disk center stage. This means that a lot of the things
you can do with DiskPatch are based on fixing disk related problems.
Use DiskPatch to clone
(copy) a disk that has read problems, or to wipe
a disk that needs to cleared. You can also check
the disk's surface and get an idea of your disk's health. One of
the more elaborate functions though is one step above the hardware
and involves the volumes that are on the disk. As you may or may not
know, all volumes on a hard disk have a number of structures that
define them. A partition table for instance defines the size and
types of the volumes, and a boot sector tells the operating system
where to find important bookkeeping information in the volume. You
can read more about these structures in this faq
item. If any of these structures is damaged you can not access the
volume. DiskPatch can identify problems in these structures and repair
them. It's important to understand that DiskPatch actually
corrects things on the disk: this is the main difference
between DiskPatch and our other tools; iRecover and iUndelete do not
actually repair things, they copy files to a safe location.
One thing that might be difficult is to recognize a certain problem and
determine if DiskPatch is the way to go. An example would be a
damaged boot sector; this would prevent you from accessing the data
normally. You could use iRecover to copy the data to safety but
repairing the boot sector would be much quicker and much more
convenient. So in this case you would need DiskPatch to repair
that boot sector. To help you identify such problems we have
created the DiskPatch walkthroughs and getting
started guide. Look at the getting
started page and see if your particular problem is listed.
Remember: you can at any time ask us for help, free of charge. Post
your problem description in our support
forum and we will suggest the right course of action.
iRecover:
This is the tool you need if you have problems accessing your
data. If a volume has been damaged or corrupted by an accident
and you no longer have normal access to the files, use iRecover to
analyze the volume and copy the files to safety. You can even
analyze whole disks and copy files to safety from any volume that
was on that disk. Causes for volume corruption could be virus
damage, operating system crashes, partition management actions (like
merging or resizing) that went wrong, or even read errors. iRecover
can be used to read data from disks that have read problems (bad
sectors) but in general it is best to clone such a disk first
(cloning can be done with DiskPatch).
One of the strong points of iRecover is working with RAID disks. If
a RAID set has sustained damage you can use iRecover to analyze the
individual disks and copy files from the RAID set to safety, as if it
was a single volume: iRecover handles all the actions that are
required to "reconstruct" the RAID set and access the
files. All you would have to do is select the files and have
iRecover copy them to a safe place. Please note that iRecover does
not actually repair things; the RAID set is not repaired. Rather,
the RAID parameters are determined by iRecover and are used to
access the disks as if they are still a functional RAID set.
The original (problem) disks are never changed. This is a good
thing: if things didn't work out after the first try you can simply
try again (adjusting some parameters if needed) and see what the
second attempt brings you. If the disks would be edited by iRecover
that would not be possible.
iRecover is somewhat "heavy duty", meaning that it is a
professional recovery solution that can be used for just about every
possible scenario. There's even support for Linux based volumes.
iRecover is the tool to have in a PC repair shop, or an IT
department. This doesn't mean that home users should be intimidated:
iRecover's scope and wizard-like interface will make sure that every
user of any level can recover his or her files.
iRecover also has a digital image
recovery mode: use this to recover pictures from cameras and
memory cards.
iUndelete:
True to its name, this tool undeletes files. It's the simplest way
to get files back from beyond the Recycle bin. iUndelete supports
most known file systems but doesn't handle damaged partitions or bad
disks. It's undeleting, pure and simple. Make sure you understand
that iUndelete can only be used if the disks and partitions are
okay; if anything is wrong with the partitions you'll have to use
one of our more advanced tools. Just look at iUndelete as an
extension to the Recycle bin.
|
| I
will buy a license if you guarantee success |
|
We can never and will never guarantee a positive outcome. All
we can tell you is that we will make every effort to help you
recover your data, if things do not initially work out when
using our software. You are free to contact us to see if we
can make an estimate on the recoverability of your data. Find
information on how to contact us here.
You can verify if the software actually runs and detects
your disks by running the demo versions. The demo versions
will give a fair indication of the possibility of data
recovery:
- You can test to see if the physical disks are detected;
if the demo version does not find the disks, the
registered version won't either
- They will detect or not detect lost partitions; if the
demo version does not find any partitions, the registered
version won't either
- Using the demos (DiskPatch
especially) you can generate log files: you are welcome to
submit the log files to us, we will always give a fair
assessment of the chances for a successful recovery.
|
| My
hardware RAID 0 array is no longer detected |
|
Users have reported successful recoveries using DiskPatch. We
can not guarantee that DiskPatch will work with each and every
hardware RAID 0 setup. Please note that the following scenario
applies to hardware RAID 0 arrays.
Scenario:
A RAID 0 array is no longer valid or detected as an array.
Using the RAID setup configuration to re-create the array
results in a warning message indicating that all data on the
disks will be lost.
Solution:
Even after continuing and ignoring the warning message,
users frequently reported that DiskPatch was able to recover
all partitions on the re-created array using the "normal"
procedures for recovering lost partitions.
Explanation:
It appears only the MBR on the array was blanked and this
resulted in data loss. DiskPatch can almost always recover all
partitions after they were lost due to a wiped MBR. Whether
DiskPatch does this on a disk or a hardware RAID array makes
no difference.
|
| "Invalid
system disk" (Windows 9x versions only) |
|
This description is relevant for Windows 9x versions ONLY.
First, make sure there is no diskette in the floppy drive.
If there is, remove the diskette and reboot your computer. If
the problem persists and is not floppy related, do the
following:
Important! Do not perform any of the following actions if the
partition can not be accessed!
- Boot the computer from the Operating System boot
diskette.
Run Scandisk on the C: drive to detect and repair any disk
errors.
- Make sure your hard drive is not compressed and that no
anti-virus programs are running.
- Insert a bootable diskette that contains the same system
files as the operating system on your hard drive. For example,
if you have Windows 95b on your hard drive, insert a bootable
floppy that was created using Windows 95b.
- Run the program "SYS.com" as follows: type "SYS A: C:"
<enter>. The file SYS.com can usually be found in the
c:\Windows\Command folder.
- The necessary system files will be transferred from the
floppy to the appropriate location on your hard drive. When
the process is complete, remove the bootable floppy and
restart your computer. You should now be able to boot your
operating system.
Note that damage to the partition's boot sector may be
beyond what can be repaired by using the above procedures. For
instance, damage to the BPB (which contains important meta
information about the file system) can not be repaired this
way; this type of damage can often be repaired by DiskPatch.
|
| Diagnosing
and resolving NT boot issues |
|
The following will help you diagnose boot problems for
Windows NT, and recover the operating system if necessary. The
following files must be located on the Windows NT System
partition:
NTLDR -- The "NT Loader" takes control of the
entire boot process until control is given to NTOSKRNL.EXE.
(This occurs when the NT system menu option is selected.)
BOOT.INI -- This file contains the location of the NT boot
partition and displays the menu options on bootup.
BOOTSECT.DOS -- This file takes control from NTLDR if you
are running an operating system other than Windows NT.
NTDETECT.COM -- This file detects the installed hardware
and adds the hardware list to the registry.
If any of the above files are missing or corrupt, an error
message will be reported that may prevent Windows NT from
loading properly. The following paragraphs outline some common
error messages and their solutions.
"Couldn't find NTLDR. Please insert another
disk."
This error message will occur during bootup if NTLDR is
missing, or if there is a diskette in the floppy drive. To
solve the problem, run an NT Repair.
"Windows NT could not start because the
following file is missing or corrupt:
winntrootsystem32ntoskrnl.exe"
This error will almost always indicate that the boot.ini is
pointing to the wrong partition. It can also be reported if
NTOSKRNL.EXE has become corrupted. To solve the problem first
verify that the boot.ini points to the boot partition. If the
boot.ini is correct, run an NT Repair; if the NTOSKRNL.EXE is
corrupt, it will be repaired.
Note: If the BOOT.INI is missing, NTLDR will attempt to
boot NT from the default directory of the active partition. If
this fails, the above error message will also be reported.
"I/O Error Accessing Boot Sector File..{ARC
Path..}"
This error message will occur when attempting to boot to
another operating system other than Windows NT. This occurs
when BOOTSECT.DOS is missing. To solve the problem run an NT
Repair.
"NTDETECT v1.0 Checking Hardware...NTDETECT
failed"
This error message will occur if NTDETECT.COM is missing or
corrupt. To solve the problem run an NT Repair.
|
| General
data recovery procedures |
|
This article describes a general data recovery procedure.
By following these steps, you increase your chances of a
successful recovery and work in the safest manner possible!
Note: this article does not address "boot-issues" (software
/ O.S. related startup problems).
Step 1: Cloning
Cloning the disk (if possible) should always be the first
step in a recovery process. Since you can then perform the
recovery on the clone, the clone provides you with a safety
net: if the recovery damages the clone, the original is not
lost and you could create another clone. Also, cloning the
problem disk might give you some additional information about
the problem disk, like the status of the disk surface, the
status of the controller or the status of the disk itself
(screechy noises, clicking noises etc.).
When a disk is going bad (bad sectors appear and/or data
becomes unreadable) cloning the disk is vital because each
read on the problem disk may generate more damage. In fact, a
dying disk should be treated as if the next read action will
be the last read action.
To reduce "noise" caused by data that is already present on
the destination disk it is advised you wipe the destination
disk first. By wiping we mean that all sectors on the
destination should be zeroed, only deleting the partitions or
reformatting partitions on the destination disk is not
sufficient! By wiping the destination disk you also verify the
condition of this disk.
Both disk cloning and wiping can be performed using DIY
DataRecovery DiskPatch.
Step 2: Diagnosis
The first thing that needs to be determined is the physical
condition of the problem disk. As mentioned, observations made
during the cloning process can provide useful information.
If your tool of choice to clone the disk doesn't detect the
disk there are probably electronical or mechanical defects. If
the disk is detected but the disk produces abnormal noises
(clicking, scratching), the disk is physically damaged. If
there is no sound at all, the disk is not spinning up.
If the disk isn't detected, produces noise or doesn't spin
up, you should send the disk to a data recovery company that
is equipped and qualified to deal with this. For example, DiskLabs.
If the disk can be read and the cloning tool creates a log
file, this log file may include information that is useful for
assessing the problem. When requesting
support always include the relevant log file (DiskPatch
creates a detailed log that contains information on what could
and could not be read, always include this log file when
requesting support).
If the cause for the data loss is known, for example
because you have accidentally deleted a partition that you
shouldn't have deleted, the diagnosis is pretty
straightforward. If data just disappeared "all of a sudden"
additional examination and analysis is required.
Avoid these common mistakes:
- If you are using "on-board" partitioning tools to
assess the situation, be very careful; for example Fdisk
deletes key structures when re-creating a partition. You
will re-create the partition, but lose its contents!
- Running on-board or 3rd party "diagnostics and fixers": by this we mean Windows Chkdsk, Scandisk or 3rd
party Chkdsk utilites like Norton/Symantec Disk Doctor. If
one or more volumes can not be accessed, do not use any of
the aforementioned utilities! If those utilities were
already used and as a result files disappeared, the
contents of those files may have been written to .CHK
files. To determine if .CHK files contain usable data you
either need to open those files (with a binary file
editor) and manually inspect their content, or a
specialized tool like CHK-Mate can be used. Note that if
files were fragmented before Chkdsk or Scandisk was run,
the contents of a file may be scattered among multiple .CHK
files.
- The Windows Recovery Console's Fixboot command; this
command frequently makes the damage worse! Do not use this
command if a volume can not be accessed!
Your first step (after you have cloned the disk) is to
verify if volumes or partitions are present. Each disk must
contain at least one partition or volume before it can be used
for data storage. A partition/volume is defined in the
partition table. In Windows 2000/XP/Vista/7/8 a volume can also be
defined in the LDM database; managing the LDM database is
handled transparently by the Windows Disk Manager.
In general a volume or partition is assigned a drive letter
(by the operating system) through which you can address the
volume. If you are no longer able to boot the operating system
you need to boot the PC from a diskette or bootable CD.
FAT and FAT32 partitions/volumes, if present and intact,
can be accessed using a DOS boot diskette. DOS can not access
NTFS volumes without the help of a driver (for example
NTFS4DOS, currently supported by Avira).
If one or more volumes or partitions can not be found, the
problem is probably in the partition tables. Even if the
damage extends beyond that, the next step is determining the
partition locations (their start position, their size).
DiskPatch can be used to get additional information about the
partition tables, and DiskPatch can rebuild the partition
tables.
If ALL volumes can be found but one or more volumes can not
be accessed, the damage is probably specific to those volumes
(the boot sector, the FAT, the MFT, etc.). DiskPatch is often
able to indicate if the problems come from a damaged boot
sector, and several types of boot sector damage can be
repaired by DiskPatch.
If all volumes are found and can be accessed while only a
few files or folders on a particular volume have disappeared
or can not be accessed, there is probably only minor file
system damage (FAT, MFT or directories are damaged to some
degree). In this case running a DiskPatch surface scan can
reveal if this damage is caused by unreadable sectors.
Step 3: Repair or Extraction
- Repair:
Before attempting in-place repairs you should verify that
the repair utility has an undo feature! As stated earlier, the
safest way to conduct repairs is to work on a clone. At the
very least, the software used to perform the repairs should
create backups of areas that are repaired or allow the user to
backup areas of the disk he or she is modifying.
In our experience it is safe to make in-place repairs if
the data loss is caused by a corrupt master boot record,
corrupt partition tables or a corrupt boot sector. It is this
type of repair that DiskPatch can be used for. DiskPatch fixes
the partition tables and boot sectors, after the user has
selected which partitions to repair. DiskPatch allows the user
to create a clone, and can undo all modifications made during
the automatic repair.
- Extraction:
If in-place repairs are impossible or undesired, software
like iRecover can be used to
extract data from the damaged disk, and copy it to an intact
disk.
Software like iRecover scans a disk to determine volume
locations and their file system specific parameters. Since
files are addressed using cluster addresses relative to the
start of a data area (like the O.S. does), it is vital that
the correct start and end values for a partition or volume are
known, and that the software is able to determine the cluster
size even if the boot sector is damaged.
If damage extends to the file system's main bookkeeping
structures (FAT, MFT), recovery of fragmented files may be
impossible (the files may be recovered, but they're probably
corrupt/incomplete).
Sometimes all vital file system information can be lost. In
this case it may still be possible to recover certain files
based on their headers (recovery software checks the file's
contents to see what type of file it is and then copies the
connected clusters to a file). With this type of recovery file
sizes and filenames are often generic. After the files are
recovered they must actually be opened one by one to determine
their content.
After files have been recovered, and before re-formatting
the problem disk (if possible or desired), the recovered data
needs to be verified! It may appear files were recovered (the
folder structure is present, files with recognizable names are
present), but this is no guarantee the files are actually
intact.
|
| Terminology:
an explanation of terms used in data recovery and disk usage |
| Terminology: |
Description: |
| Sector |
Data
is stored on a disk by arranging it in bytes, simply put. It is however
not possible to read an individual byte from the disk; you have to read
512 at a time. 1 sector contains 512 bytes, so a sector is the smallest
addressable area on a disk. |
| Cluster |
An operating system
doesn't access the disk per sector, it accesses the disk per cluster. A
cluster is a number of sectors that are read or written as a group. So,
a
partition is divided into clusters, which in turn are made up of
sectors, which in turn are made up of bytes. |
| Partition / Volume |
A region of a disk that is used for file storage. An empty disk by
itself can not be used to store data: partitions (or volumes) must be
created on the disk so the operating system knows where to put the
files. A disk can be divided into many partitions. Usually 1 disk
contains 1 partition which takes up all the disk's space, but since the
introduction of Windows 2000 there is actually no real limit to how many
partitions can be on a disk. In fact, it's often considered good
practice to create 1 partition for the operating system and 1 or more
partitions for data storage. Partitions need to be formatted before they
can be used (as explained under "partition types: formatted
as..."). |
| Disk
facts |
A hard
disk (the physical device) is made up of platters. To be able to find a
specific location on the disk it is divided into cylinders, each
cylinder is divided into tracks (or heads) and each track is divided
into sectors. In earlier days these numbers were vital. Nowadays we use
LBA addressing and the cylinder/head/sector numbers are no longer used
to identify a disk location, but they still play a part when partitions
are created. It is important that the PC recognizes the disk correctly
(size, head and sector values must correspond to what the manufacturer
states). |
| LBA
address |
LBA
stands for Logical Block Addressing. This is simply a way to identify
each sector on a disk: the first sector on a disk (the MBR) is numbered
0 and all following sectors are numbered sequentially. So a LBA address
is nothing more than a sector number. Keep in mind that the count starts
at 0; LBA sector 10 is the 11th sector on the disk. |
| Basic / Dynamic
disks |
A description of disk types that is
used in Windows (2000 and up): a Basic disk contains simple volumes
(using the "old fashioned" partition table to describe partitions), a
Dynamic disk allows for more types of volumes such as spanned/mirrored/raid/striped.
Read more in the Microsoft knowledge base (link). |
| Disk
geometry |
Geometry consists of 2 values: heads (per cylinder) and
sectors (per track). These values can differ from disk to disk, but are
generally 255 (heads) and 63 (sectors). Although modern operating systems
don't need to do this anymore, these values are still used when
creating (basic) disk partitions. As such it is important that the PC has
the correct values for the disk; if the values are wrong the operating
system will not be able to find certain important structures on the disk
and fail to boot, or even fail to access the partition completely.
|
| S.M.A.R.T. |
Self Monitoring,
Analysis and Reporting Technology.
This is a monitoring system for computer hard disks to detect and report
on various indicators of reliability, in the hope of anticipating
failures. All modern day hard disks have SMART technology on board.
SMART keeps an eye on various aspects of the hard disk (spin up counts,
bad sector count, number of hours used etc.) and attempts to predict
problems by analyzing the data that is gathered and comparing that against
certain threshold values.
For instance, say the disk has developed a number of bad sectors. Usually
that doesn't have to be a problem since these sectors are taken out of
commission and are replaced by sectors from a spare pool. But this can not
go on indefinitely; at some point either the spare sectors run out or the
amount of bad sectors is growing so rapidly that the disk will cease to
function within the near future. SMART would keep an eye on these
emerging bad sectors and once the amount of bad sectors has crossed a
threshold value, SMART takes note.
Understand that SMART itself will not notify the user; a SMART capable
utility or operating system is needed to receive these warnings. For
instance, the SMART tool that is included in the DiskPatch distribution
can check the status of your disk by reading and interpreting the SMART status.
|
| MBR |
The Master Boot Record, the first
sector on the disk. Contains the partition tables (a list that describes
the locations of partitions on the disk). |
| EPBR |
Like
the MBR, this sector contains a partition table. But unlike the MBR, the
EPBR can be anywhere on the disk. An EPBR describes a logical partition
and can be found on disk just before the position where the logical
partition starts. |
| Partition
Tables |
These
describe the various partitions on the disk: where they start, how big
they are, what type they are. The MBR contains a partition table.
A partition table has space for 4 entries, meaning that 4 partitions of
any type is the maximum that can be created. This can be circumvented by
creating at least 1 extended partition, which in turn can contain many
more logical partitions. A common setup is 1 or 2 primary partitions and
1 extended that contains the rest of the partitions. Primary partitions
are the only ones that can be used to start an operating system from.
If Dynamic disks are used the partition table explanation as described
here does not apply anymore. |
Partition
types:
primary / extended / logical |
Primary
partitions are the only types that allow you to boot an operating
system. In general, you would install the operating system in a primary
partition. A primary partition can be set "active", which allows the
computer to locate the operating system that needs to be started. Only 1
primary partition can be active, but more primary partitions can exist.
If more than 1 primary partition is set active, the computer will not
boot. Also, when no primary partition is set active, the computer will
not boot. Only the partition table in the MBR can contain primary
partitions.
Extended partitions are nothing more than "containers" for
logical partitions. In general, all space that is not used by primary
partitions must first be added to an extended partition before logical
partitions can be created. Only 1 extended partition can exist per disk.
The partition table entry that describes the extended partition is in
the MBR.
Logical partitions are the partitions that are created in the
extended partition area. More than 1 logical partition can exist in the
extended partition area.
Note that when using Dynamic disks these rules don't apply anymore.
More information can be found in this knowledgebase article (link). |
Partition
types:
formatted as FAT / NTFS / etc. |
A
partition needs to be formatted before it can be used. Different
operating systems offer/require different format types:
DOS and older Windows versions (9x/ME) use a FAT type format. Newer
Windows versions (NT/2000/2003/XP/Vista/7/8) can also use NTFS type formats
(NTFS is preferred, it is more reliable than FAT). Other operating
systems use different formats: Linux uses EXT2/EXT3, OS/2 uses HPFS,
etc.
NTFS stands for New Technology File System. |
Partition
states:
active / hidden |
A
primary partition can have 2 flags enabled or disabled to signify a
certain state. These flags are kept in the partition table in the MBR.
The states are: active/not active/hidden/not hidden. A primary partition
that holds a bootable operating system needs the "active" flag to be
set. Only 1 partition can be set active (which makes sense; you can only
boot 1 operating system at a time). The "hidden" flag is not as
important as it was in the days of Windows 9x: earlier Windows versions
couldn't deal with more than 1 primary partition, so if you had more
than 1 the others had to be "hidden". Windows 2000/2003/XP/Vista/7/8 have no
problems with 2 or more primary partitions.
These flags can be set with DiskPatch (from the MBR operations
menu). |
| Boot
sector |
The
very first sector of a partition. This sector contains a lot of
information that the operating system needs to be able to use the
partition. If this sector is damaged or deleted, the partition can not
be accessed. |
| FAT |
The
File Allocation Table. A table that is used to keep a record of which
parts of the disk are in use. Only used when a partition is formatted as
a FAT type partition. |
| LDM
database |
A region of the disk that is
claimed by Windows 2000/2003/xp/vista/7/8, containing a description of all
volumes on all disks in the system. The LDM database is located at the
end of the disk and is invisible when working with Windows. The LDM is
synchronized to all disks on the system when either volumes are created,
or volumes are added/changed (when using Windows disk manager). The LDM is only created and used when Dynamic disks are used. |
| Type 42h
partition |
When the disks are Dynamic and the
LDM is used to describe the volumes on the disk, Windows creates a
partition type in the MBR that envelopes the entire disk. This way
windows knows that the "old fashioned" partition table is not used
anymore; the LDM is used to identify volumes. |
| Boot
code |
Both the MBR and a
boot sector can contain a small "program" that allows the operating
system to find the stuff that is needed to boot. This
"program" (boot
code) is only required if one partition is an active primary partition,
and if an operating system is booted from that partition. The boot code
in the MBR differs from the boot code in a boot sector; the MBR's boot
code is needed to find the partition that has to be booted (this boot
code is used by the computer's BIOS), the boot sector's boot code is
needed to find the necessary files on that partition (this boot code is
used by the operating system). Since the boot code in the boot sector is
used by the operating system, it is different for each operating system. |
| "Fdisk damage" |
When a FAT
partition is deleted, people sometimes recreate the partition (using
fdisk or a similar tool) in the hope that it will then show up again.
This is not the case. This procedure actually damages the FAT area and
requires additional repairs to make the partition accessible again.
DiskPatch can detect and repair this damage. |
| Root
cluster |
The cluster in the
partition that contains the start of the root directory. Only applicable
to FAT32 partitions. The location of the root cluster is needed to rebuild
a damaged boot sector. |
| MFT |
The Master File
Table. A structure on a NTFS formatted partition that contains
information on where files and other important NTFS structures can be
found. If this structure is damaged the partition will be inaccessible.
In general, the MFT can not be repaired. If MFT damage exists,
file-oriented recovery should be used (copy files from the damaged
partition using a file recovery tool like iRecover). |
|
|