Help - Dynamic Disc Woes - Win2k :(

etones · #1 (**permalink**) 02-01-03, 09:48

Hi everyone, im having some severe problems, hoping some of you guys/ladies
can help me out. Im using Win2k Server, SP3.

To cut a long story short, I have 3 hard drives connected to Promise EIDE
controller. On Thursday all the drives vanished, hmmmm... very strange
(worked perfectly for over a year), I opened the system, played with the
cables, nothing.

So, I bought a new EIDE controller, tried it, nothing. Put the old
controller back in, and the drives have somehow appeared into Disk
Management, however if I try and re-activate them i get the error:

'INTERNAL Error - Disk Group has no valid configuration copies'

I looked over http://support.microsoft.com/?kbid=294244 and it suggests a
re-booting the system, again, no luck. After a search of news, I see a
program called RepoMan is recommended: http://www.diydatarecovery****/

However, im unsure is it will help me as i have a strange disc config,

disc #1 and disk #2 are connected as a spanned volume (200gb total).
disc #3 is split into 4 partitions (40gb total).

In total, there is 240gb worth of data that i really cant loose, I must get
it back. Please help.

Thanks all.
Taz

tripl333 · #2 (**permalink**) 05-16-03, 02:32

same problem, and is is a big pain...

i have a 120gig that needs recovering, but i too have NOOOOOOO idea ...

i tried putting it in another box and it still gives the same error when trying to import the new drive....

if you find a solution, hit me back....in the meantime, ima look for some software that might do SOMETHING.......grrrrr

tripl333

etones · #3 (**permalink**) 05-16-03, 03:43

Hiya, luckyly i did manage to find a solution.

MS have a until to restore the new MBR style partition it places at the end of the drive, called dmpss.exe, its $100 however i managed to blag it for free

If you need it I can mail it to you, i however coule not get it to work

However, i did find an app called r-studio which re-constructed the drive and let me copy across all my data, if that isnt an option, let me know and i'll send you the file, regards

Taz

ainsleyiv · #4 (**permalink**) 06-11-03, 16:27

did you get your data back online, i have a 80g "data" drive that i tried to import after i formatted and reinstalled XP on the 40g o/s app drive. please let me know what worked, thanks.

HaroldM · #5 (**permalink**) 07-20-03, 20:41

Ok I am having a similar situation let me explain my setup first off.

I have 2 30gig drives (one is my main os drive and the second is a slave off that drive) which are plugged into the primary IDE port on the mobo. I also have 2 120gig drives. Each is set as a single drive plugged into a IDE Controller card.

Everything was fine until I had to move the IDE controller card to a different PCI slot. Now I have the problem where the slaved 30gig and the second 120gig are not showing in Win2K and when I try to import the drives in Disk Management I get the error that says "INTERNAL Error: Disk has no valid configuration copies."

I've searched all over the net for solutions and haven't found much. I came across this post and wondered if anyone had a solid solution and if so could they share it with me. I am not the stupid computer user but some of the stuff I've seen related to this problem is a little foreign to me. If someone could either reply here or message me on AIM (HSDHollywood) or email me (hmartin@mchsi.com) it would be greatly appreciated. I really need to recover this data as it is customer information that I use for my web design business I do on the side. Thanks in advance for any help you all may provide.

etones · #6 (**permalink**) 07-21-03, 12:10

Hi,

1) Do you have a spare computer u can hook the drives upto
2) Do you have some spare HD space you can copy the data to.

If so, i can help you. If not, i can give you the email microsoft sent me (which I couldnt use, but worth a try).

Taz

HaroldM · #7 (**permalink**) 07-21-03, 12:55

Yes I have one of the 120s are still being picked up in Windows so I got space to put data on. I don't have another Win2K Machine unfortunately or any other machine for that matter. I do have a friend at work who is running XP Pro and said he'd be willing to try putting my drives in importing them to see if that works. Why do you ask these two question? Thanks for replying btw.

mossebo · #8 (**permalink**) 07-27-03, 16:36

Hi,
I have the same problem. I get the "INTERNAL Error: Disk has no valid configuration copies." error message when trying to reactivating my dynamic offline disks.

Have also searched for solutions, but haven't found much more than this...

Please post to this thread if you find a solution

on how to recover data from the disks or how to make them online again...

Best Regards
/Dan

etones · #9 (**permalink**) 08-14-03, 11:48

Hi all, damm stupid thing lost my original reply, so here goes again.

1) Sorry for not getting back to many of the guys who emailed me for help
2) I've attached the MS utility to this post along with the help documentation they sent me
3) I see no problem attaching the file/help as it was sent to me free of charge by microsoft.

So. here goes

Quote:

Hi Taz,

The following example is a sample output from the Dmpss
utility that displays the syntax for each command:

dmpss - Logical
Disk Manager Recovery Tool
Usage:

dmpss dumpconfig
<source disk number>? <target
file>
dmpss restoreconfig <source
file>? <target disk number>
dmpss
copyconfig <source disk number>? <target
disk number>? <target diskid>
dmpss
repairconfig <target disk number>
dmpss getheader
<source disk number>
dmpss initheader
<target disk number> [/restore]
dmpss
changeheader <target disk number>?
<attribute1>=<value1> <attribute2>=<value2> ...

diskid = guid
hostid = guid
dgid = guid
dgname = string
bootsig = unsigned decimal
disksetid = guid
newdisksetid = guid
puboffset = unsigned decimal
publen = unsigned decimal
toc = unsigned decimal
alttoc = unsigned decimal

dmpss zapconfig <target disk number>
(This is a hidden command.)

The definitions for each command
are:

The dmpss dumpconfig command: This command can read the 1 megabyte (MB) (2,048
sectors) Logical Disk Manager database from the end of
the source disk and save
it to a file. For example, dmpss dumpconfig 2
c:\ldm2.dsk.

The dmpss restoreconfig command: This command can restore a previously saved Logical Disk
Manager database from a source Logical Disk Manager database file and restore
it to the end of the target disk. This command can overwrite the existing
Logical Disk Manager database. For example, dmpss restoreconfig c:\ldm2.dsk 2.

The dmpss copyconfig command: This command can copy the 1 MB (2,048
sectors) Logical
Disk Manager database from a healthy dynamic disk to a target dynamic disk that
displays a "missing" message or that you are
attempting to repair. This command
requires the correct "diskid:<guid>" that had been
originally assigned to the target disk. For example,
dmpss copyconfig 1 2 45260b21-e0e7-4bd6-b817-2e75b94de9b6.

This information can be extracted from the DMDiag tool
report under the following two sections:

The "LDM Disk Config Hard disk#" section as noted below - search for the word Detached.

---------- LDM Disk Config Hard diskx ----------
...
..
.

#Record 3: type=0x0034 flags=0x0000 gen_flags=0x0004 size=120
#Blocks: 7 8
Disk: Disk2 rid=0.1027 updated=0.1061
assoc: diskid=45260b21-e0e7-4bd6-b817-2e75b94de9b6 detached

lastdevice=SCSI\DISK&VEN_SEAGATE&PROD_ST19171W&REV _0024\4&1079D045&0&050
flags:

The "DMAdmin Verbose Query" section. A device is not associated with this missing disk:

---------- DMAdmin Verbose Query ----------
...
..
.

Disk: Disk2
info: diskid=45260b21-e0e7-4bd6-b817-2e75b94de9b6
assoc: nodevice (last diskset:
00000000-0000-0000-0000-000000000000)

The Dmpss repairconfig command: This command is used to repair a LDM data that shows
"Invalid Magic Number" errors:

---------- LDM Disk Config Harddisk1 ----------

#Config copy 01

ERROR: Error (File block 47): Invalid magic number in the
configuration copy
#Header nblocks=5924 blksize=128 hdrsize=512
#flags=0x0100 (CLEAN)

This command only addresses the situation when there are corrupted
blocks in the configuration copy. It does not fix disk header, table of
contents, or config copy header corruptions. There is nothing we can do if the
corrupted blocks contained some real volume records.
These records may be
deleted when you use this command.

The dmpss getheader command: This command can read the Logical Disk Manager header
from the source disk and display the header
information. There are three copies
of the header. Two copies are located in the Logical
Disk Manager database
itself. The first copy is located in the last sector
of the physical disk, and
the second copy is located at the last sector of the physical disk minus 191
sectors. The third (tertiary) header is in sector 6 of
the physical disk and is
used when you extend hardware raid arrays so that you
can move the Logical Disk
Manager database. This command can only attempt to
read the first two headers
that are located in the Logical Disk Manager database.
For example, dmpss getheader 2.

Header
======
diskid: 45260b21-e0e7-4bd6-b817-2e75b94de9b6
group: name=Ftclass2Dg0
id=fa248f78-b5de-4102-856f-a2153ad10e96
flags: private autoimport
import: bootsig=836838673 hostid=1b77da20-c717-11d0-a5be-00a0c91db73c
diskset: id=00000000-0000-0000-0000-000000000000
version: 2.11
iosize: 512
public: slice=0 offset=63 len=17767827
private: slice=0 offset=17781064 len=2048
update: seqno=0.5 time=-273336016
headers: 2047 1856
configs: count=1 len=1481
logs: count=1 len=224
tocs: 1 2046

If the header is corrupt, or it does not exist, for example, on a
basic disk, the dmpss getheader command can return the following error message:

dmpss getheader 2
Error: Invalid disk header: A
format error has been found in the private
region of the disk.

The dmpss initheader [/restore] command: This command can re-initialize a generic header on the
target disk. The command can only initialize the two headers in the Logical
Disk Manager database, not the tertiary backup header.
The DMDiag tool cannot
extract information from the Logical Disk Manager
database without a proper
header. If the header that is created is not a
complete accurate header, you
must use the changeheader command to modify the fields
that are incorrect. For example, dmpss initheader 2.

NOTE: The /restore switch is an optional switch that is used to copy the tertiary
Logical Disk Manager header (located in sector 6 of
the target disk) onto the
primary and secondary Logical Disk Manager headers.
The /restore switch must not be used unless Windows 2000 Service Pack 3 or the
hotfix from the following article (Q294244) had been previously applied before
a problem occurred.

For
additional information, click the article
number below to view the article in
the Microsoft Knowledge Base:

KBLink:294244.KB.EN-US: Extending Hardware Raid Array May Cause Dynamic
Unreadable/Offline Error

Header
======
diskid: 00000000-0000-0000-0000-000000000000
group: name=Bogus id=00000000-0000-0000-0000-000000000000
flags: private autoimport invalid-primary invalid-alt
import: bootsig=0 hostid=1b77da20-c717-11d0-a5be-00a0c91db73c
diskset: id=00000000-0000-0000-0000-000000000000
version: 2.11
iosize: 512
public: slice=0 offset=63 len=17767827
private: slice=0 offset=17781064 len=2048
update: seqno=0.31 time=-849813984
headers: 2047 1856
configs: count=1 len=1481
logs: count=1 len=224
tocs: 1 2046

etones · #10 (**permalink**) 08-14-03, 11:48

Quote:

If the target disk is basic, or it contains an unknown partition
table entry, when you attempt to initialize a new
header, you can receive the
following error message:

dmpss initheader 0
Error:
Basic or unknown partitions were detected on
the target disk.

The dmpss changeheader command: This command is used to modify a newly initialized
header after you have run the initheader command, or
you want to modify the existing header attributes to
make repairs. To make effective repairs, you must
first run the dmdiag.exe -v command to extract vital information from the Logical Disk
Manager database.

The following list has header attribute names and
their descriptions.

NOTE: The header attributes that have an asterisk (*) are required to
be modified after you run an initheader command before
the dynamic disk can be recognized.

- * diskid: Is the globally unique identifier (GUID) of that physical
disk that is extracted under the "DMAdmin Verbose Query" section of the
DMDiag tool report. Every physical disk contains a unique GUID.

- hostid: Is not applicable for Windows 2000. Hostid is a value that is
used only on UNIX-based computers. On Windows-based computers, this
value always has the same fixed value.

- * dgid: Is the disk group-ID GUID that is extracted under
the "LDM Disk Config Hard disk X"
section of the DMDiag tool report. This
header attribute is also stored in the
registry under the following key in a
value called ID:REG_SZ:<Guid>:

HKEY_lOCAL_MACHINE\System\CurrentControlSet\Servic es\Dmio\BootInfo\Prima
ryDiskGroup

- * dgname: Is the name of the disk group. This header
attribute is usually the computer name
that is attached with DG0 (IE:
Ftclass2Dg0) that is extracted from the
Group record that is under the "LDM
Disk Config Hard disk X" section of the
DMDiag tool report. This header
attribute is also stored in the registry
under the following key in a value
called NAME:REG_SZ:<dg_name>:

HKEY_LOCAL_MACHINE\System\CurrentControlSet\Servic es\Dmio\BootInfo\Prima
ryDiskGroup

- * bootsig: This value is the disk signature of the boot disk that
contains the Windows operating system files (%Systemroot%) converted to
decimal. As displayed on the DMDiag tool report, disk-0 disk signature

= 31e12511 hex that makes bootsig=836838673.

- * disksetid: By default, this header attribute is a GUID
that contains all zeros. When you create
a disk group, all disks have the same
disksetid of zero. However, whenever a
disk is removed from a disk group, the
remaining disks receive a new disksetid.
The current disksetid (if non-zero) is
also stored in the following registry in
a value called
DiskSetId:REG_SZ:<Guid>:

HKEY_LOCAL_MACHINE\System\CurrentControlSet\Servic es\Dmio\BootInfo\Prima
ryDiskGroup

An an example, you can have a disk group with 5 disks on
computer A. Each disk has an identical copy of the
Logical Disk Manager
configuration. If you move 2 disks to another
computer, computer B, and then
you import them, you can change the configuration (for example, you create a
new volume). You have 3 disks with the previous configuration copy and 2 disks
with a new configuration copy. Then, you shut down the computers, move the 2
disks from computer B back to computer A, and then you restart the computer
with all 5 disks together again. Logical Disk Manager
has to decide which
configuration copy to load for that disk group: The
one from the 3 disks that
remained on computer A or the one from the 2 disks
that moved around. The
solution is to stamp a GUID called "disksetid" into
the header of each dynamic
disk. When Logical Disk Manager is confronted with
this problem, it reads the
configuration copy from disks that have the same
disksetid as the one that had
been stored in the registry. The other disks are not brought online at all
because Logical Disk Manager makes the assumption that
they were reconfigured
on another computer. Therefore, their previous configurations do not match
anymore with the data on disk. In this example,
Logical Disk Manager is going
to read the configuration from the 3 disks that
remained on computer A, and
then Logical Disk Manager is going to mark the other 2 disks as foreign. When
the disks are imported, they are displayed as offline.

- newdisksetid: It is only used to specify a new disksetid when it is
needed.

- puboffset: It is usually 32 or 63 decimal, based on the drive
geometry, and it is the starting sector where the first partition can
start on the disk. However, if there is an OEM Enhanced Industry
Standard Architecture (EISA) partition at the beginning of the disk,
puboffset is set at the end of the OEM partition.

- publen: It is the geometric size of the disk, minus the puboffset
value, and represents the usable disk space for data volumes. Because
Logical Disk Manager attempts to hide the last cylinder of the disk,
after you make repairs, you may have to reduce the publen value,
especially when you copy a Logical Disk Manager database from different
size disks by using the copyconfig command. If the last fraction of a
cylinder is less than 1 MB, the end of the public region is going to be
less than the geometric size of the disk to accommodate the 1 MB
private region.

- toc: It is a pointer to the table of contents of the private region,
and it must not be altered.

- alttoc: It is a pointer to the alternate table of contents of the
private region, and it must not be altered. For example: dmpss
changeheader 2 bootsig=836838673 dgname=ftclass2Dg0
dgid=fa248f78-b5de-4102-856f-a2153ad10e96
diskid=45260b21-e0e7-4bd6-b817-2e75b94de9b6

- zapconfig. It is a hidden switch that can enable you to completely erase
the 1 MB (2,048 sectors) Logical Disk
Manager database at the end of any
physical disk. This switch can erase the
last 2,048 sectors irregardless of the
disk type (for example, basic or
dynamic). This switch is useful to erase
unnecessary Logical Disk Manager
databases from basic disks that had been
dynamic at one time. (The DMDiag tool
had to report such unnecessary
information.)

For example: dmpss zapconfig 2

If you use this switch, you may receive the following
message:

Warning: Zapconfig is a hidden option. Running
it may cause
irreversible damage to your data.
Are you sure you want to
erase the Logical Disk
Manager configuration of disk 2 (Y/N)? If you
enter "Y", you may receive the following
message:

The Logical Disk Manager configuration of disk 2 has been successfully
erased.

PICTURE OF DYNAMIC DISK LAYOUT
==============================

sector# (last 2048 sectors)
0 6 63(puboffset) <LDM Database>

|---------------------------------------------------------|
|M H | | H H|
|B D | USER SPACE FOR VOLUMES | D D|
|R R | (Publength) | R R|
| 3 | | 2 1|
|---------------------------------------------------------|
^ ^ ^ ^
Last Sector is
|>- Tertiary header points to database-->| | |<-
Main Header
|
|<- 191 Sectors back
is Backup Header

HaroldM · #11 (**permalink**) 09-22-03, 05:11

I meant to reply but have been busy. Thanks etones. You are a lifesaver. Worked great. Got my two drives back. Much thanks.

ed3 · #12 (**permalink**) 12-13-03, 18:52

Quote:

Originally posted by etones

I have a similar problem as you had before. I had a 80G dynamic hard disk as my secondary drive. I have a 20G hard disk (basic disk) installed with XP. The volumn of the dynamic hard disk disappears while I copying some files from an external hard disk. I reboot the system and found the disk 1 (dyanmic disk) becomes offline. When I try to reactivate it from disk manager. It shows an error in the event log "INTERNAL Error - There are no valid log copies in the disk group (C1000082)."

I saw your post and download the dmpss.exe and try it. I was able to get the header information of the primary disk ("dmpsss getheader 1"). But I was getting an error when I try "dmpss getheader 0"
I get the following error

Error: Invalid disk header:
A format error was found in the private region of the disk

I don't know how the disk ID is changed. I tried disk ID greater than 1, it will give me path not found error.

I have tried to restore the header 0. But I got an XP windows exception and prompt me to send the debug information. I wonder you have any idea of how to bring the dynamic disk online again. Do you have another utility that can help if the dmpsss doesn't work? Thank you in advance.

HFloyd · #13 (**permalink**) 01-16-04, 14:38

HaroldM,

Could you outline the steps you used to restore your drives? I have an 80GB drive that went offline yesterday and haven't been able to fix. I used the command

dmpss initheader 2 /restore

But nothing changed.

Thanks so much,

Heather

HaroldM · #14 (**permalink**) 01-16-04, 17:26

Honestly I don't remember it was so long ago now. I am really sorry. I just read the documentation that was posted here and I did the copyconfig one or something. Wish I really remembered.

SycHo_SNaiL · #15 (**permalink**) 03-05-04, 12:43

So I guess this is a bit late to start this up again but anyway maybe I can get some help here as there isn't alot available from Microsoft's web site and the best reply on google is this. I have a PC running Windows 2000 Server with 5x HDD's. Disk 0 = 80Gig IDE, 1 = 80Gig IDE, 2 = 80Gig IDE, 3 = 20Gig IDE, 4 = 40Gig SCSI. I boot off Disk 4, Disk 0 + 1 are configured as a software mirror, the rest are Simple Volumes. Disk 4 failed in the week. I moved Disk 1 to a Windows XP PC as I have many times before with other Dynamic Disks and got the Internal error - Disk group has no valid configuration copies. I have since installed a new 9Gig SCSI as disk 4 and loaded Windows 2000 Server. Connected all 4 IDE disks and still get the same error. I got the DMPSS file from this post and got all the headers. What do I have to do to get the IDE disks up and running again though? I was thinking of copying the configuration from disk 0 to disk 4 as I can always reload disk 4 if I need. Would realy like some help on this one.

Thanks §

Internet Services

Web Development

Digital Marketing

Consumer Tech