Help! Raid 5 rebuild question - Dell tech no help

G

Guest

Guest
Archived from groups: microsoft.public.win2000.general (More info?)

I have a situation with a Dell server with a 3 disk Raid 5 configuration that
I broke down for regular cleaning. When reassembled I could attempt a log-on
but before the admin profile would load it would freeze. After attempting to
reboot, one of the drives appeared to be offline. I called Dell's extended
warrantee department and the tech had me pull the drives, swap the order of
drives, reconfigure the controller and attempt a reboot - a couple of times.
After several errors of no operating system I finally got off the phone with
him and was practicall helpless. I tried a few different orders for the
disks and finally got the Windows screen. Once it gets past that I get the
following stop error: 0x0000007B (0x85E9D330, 0xC0000032, 0x0000000)
INNACCESSABLE BOOT DEVICE.

Unfortunately this is the only server and without it the business is
practically paralyzed The Dell tech said that rebuilding is not something we
want to do, but wouldn't the parity bit help us recover from this?

I need some serious help here. Thanks in advance.
 
G

Guest

Guest
Archived from groups: microsoft.public.win2000.general (More info?)

Having been in your situtation more than once, I feel your despair.

First off, the drive order should make no difference to anything. The RAID
signatures are on the drives and the controller reads those to determine the
striping order. This is required for any reliable hot-swapping. (Unless it is very old
the PERC controllers are LSI Logic-based and as such are very solid.)

So the fact that the drive order >does< seem to make a difference is suggesting to
me that you probably have - or had - a marginal cable or connector - which is about 4 out
of 5 system problems in general. Every time you swap drives around you are
mechanically altering it which causes drives to drop in and out of the array.
Nothing is more detremental to a RAID5 array.

RAID5 is a good compromise in theory, but turns out to be very fussy in practice.
I no longer even specify it for any application - it offers only mediocre performance
and is prone to the very sorts or problems you are having now.

At this point you are into a disaster recovery scenario. I would first try all three
combinations of two drives, removing a third drive entirely, to see if there is one
pair you can boot from in degraded-array mode. If so, you want to rebuild that
array onto a fresh third drive (not the drive you have).

If that doesn't pan out, you can try an in-place upgrade/install of Windows to try to get a
bootable O/S. That will let you at least see what you have left on the rest of the partitions.
Before you do this you may want to boot to the repair console and do a chkdsk /p to make
sure the MFT is in good enough shape to use. (Prepare yourself for a possible shock there.)

After that, you just have to pull out the backups.

Steve Duff, MCSE, MVP
Ergodic Systems, Inc.

"dtremain" <dtremain@discussions.microsoft.com> wrote in message news:521A611A-171B-43B7-91CD-68242890A392@microsoft.com...
>I have a situation with a Dell server with a 3 disk Raid 5 configuration that
> I broke down for regular cleaning. When reassembled I could attempt a log-on
> but before the admin profile would load it would freeze. After attempting to
> reboot, one of the drives appeared to be offline. I called Dell's extended
> warrantee department and the tech had me pull the drives, swap the order of
> drives, reconfigure the controller and attempt a reboot - a couple of times.
> After several errors of no operating system I finally got off the phone with
> him and was practicall helpless. I tried a few different orders for the
> disks and finally got the Windows screen. Once it gets past that I get the
> following stop error: 0x0000007B (0x85E9D330, 0xC0000032, 0x0000000)
> INNACCESSABLE BOOT DEVICE.
>
> Unfortunately this is the only server and without it the business is
> practically paralyzed The Dell tech said that rebuilding is not something we
> want to do, but wouldn't the parity bit help us recover from this?
>
> I need some serious help here. Thanks in advance.
 
G

Guest

Guest
Archived from groups: microsoft.public.win2000.general (More info?)

Are you getting the Stop 0x0000007b when the Windows splash screen is
displayed? If so, that often indicates a problem with a filter driver. If
Steve's suggestions don't work, boot to the Recovery Console, browse to the
Winnt\System32\Config folder, and rename System to System.old and Software
to Software.old. Next, copy System and Software from the Winnt\Repair
folder. These hives are created when Windows is first booted successfully
after install or upgrade. Restart the server and try to boot to Windows. If
successful, look at Knowledge Base article 826901 "Stop 0x0000007B" error
message after you remove Dell OpenManage
(http://support.microsoft.com/?id=826901). That may not be the exact filter
driver that is causing the problem, but steps similar to the article (you
would open System.old, not System, for instance) can be used to look for
filter drivers.

VERY IMPORTANT NOTE: You need to be careful removing filter drivers because
you can create more problems by making an incorrect Registry change. Either
make a copy of the entire hive or export the specific key you're working on
prior to deleting anything. If you have questions about the process, please
post them here.

Kevin McNiel, MCSE/MCSA
Platform Server Setup Group

This posting is provided "AS IS" with no warranties, and confers no rights.
Please reply to the Group only, This address cannot receive incoming
messages.


"Steve Duff [MVP]" <ergodic@ergodic-systems.com> wrote in message
news:OrO8e5AlFHA.2396@TK2MSFTNGP12.phx.gbl...
> Having been in your situtation more than once, I feel your despair.
>
> First off, the drive order should make no difference to anything. The RAID
> signatures are on the drives and the controller reads those to determine
> the
> striping order. This is required for any reliable hot-swapping. (Unless it
> is very old
> the PERC controllers are LSI Logic-based and as such are very solid.)
>
> So the fact that the drive order >does< seem to make a difference is
> suggesting to
> me that you probably have - or had - a marginal cable or connector - which
> is about 4 out
> of 5 system problems in general. Every time you swap drives around you are
> mechanically altering it which causes drives to drop in and out of the
> array.
> Nothing is more detremental to a RAID5 array.
>
> RAID5 is a good compromise in theory, but turns out to be very fussy in
> practice.
> I no longer even specify it for any application - it offers only mediocre
> performance
> and is prone to the very sorts or problems you are having now.
>
> At this point you are into a disaster recovery scenario. I would first try
> all three
> combinations of two drives, removing a third drive entirely, to see if
> there is one
> pair you can boot from in degraded-array mode. If so, you want to rebuild
> that
> array onto a fresh third drive (not the drive you have).
>
> If that doesn't pan out, you can try an in-place upgrade/install of
> Windows to try to get a
> bootable O/S. That will let you at least see what you have left on the
> rest of the partitions.
> Before you do this you may want to boot to the repair console and do a
> chkdsk /p to make
> sure the MFT is in good enough shape to use. (Prepare yourself for a
> possible shock there.)
>
> After that, you just have to pull out the backups.
>
> Steve Duff, MCSE, MVP
> Ergodic Systems, Inc.
>
> "dtremain" <dtremain@discussions.microsoft.com> wrote in message
> news:521A611A-171B-43B7-91CD-68242890A392@microsoft.com...
>>I have a situation with a Dell server with a 3 disk Raid 5 configuration
>>that
>> I broke down for regular cleaning. When reassembled I could attempt a
>> log-on
>> but before the admin profile would load it would freeze. After
>> attempting to
>> reboot, one of the drives appeared to be offline. I called Dell's
>> extended
>> warrantee department and the tech had me pull the drives, swap the order
>> of
>> drives, reconfigure the controller and attempt a reboot - a couple of
>> times.
>> After several errors of no operating system I finally got off the phone
>> with
>> him and was practicall helpless. I tried a few different orders for the
>> disks and finally got the Windows screen. Once it gets past that I get
>> the
>> following stop error: 0x0000007B (0x85E9D330, 0xC0000032, 0x0000000)
>> INNACCESSABLE BOOT DEVICE.
>>
>> Unfortunately this is the only server and without it the business is
>> practically paralyzed The Dell tech said that rebuilding is not
>> something we
>> want to do, but wouldn't the parity bit help us recover from this?
>>
>> I need some serious help here. Thanks in advance.
>
>
 
G

Guest

Guest
Archived from groups: microsoft.public.win2000.general (More info?)

Thanks for the help Steve and Kevin. I was finally able to get the DC
started after obtaining the right Perc drivers from Dell. The chkdsk /p
followed by chkdsk /a resolved the problem. Several OS files were corrupted
along with some of the OS prgs (ie TS) as was the Veritas BU prg, but the
good thing is that the company data is still there. It was thought that
maybe a boot sector virus hit the DC but I'm not 100% sure of that. Monday
brings me to a parallel install of Windows.

Thanks again for the help.

"Kevin McNiel [MSFT]" wrote:

> Are you getting the Stop 0x0000007b when the Windows splash screen is
> displayed? If so, that often indicates a problem with a filter driver. If
> Steve's suggestions don't work, boot to the Recovery Console, browse to the
> Winnt\System32\Config folder, and rename System to System.old and Software
> to Software.old. Next, copy System and Software from the Winnt\Repair
> folder. These hives are created when Windows is first booted successfully
> after install or upgrade. Restart the server and try to boot to Windows. If
> successful, look at Knowledge Base article 826901 "Stop 0x0000007B" error
> message after you remove Dell OpenManage
> (http://support.microsoft.com/?id=826901). That may not be the exact filter
> driver that is causing the problem, but steps similar to the article (you
> would open System.old, not System, for instance) can be used to look for
> filter drivers.
>
> VERY IMPORTANT NOTE: You need to be careful removing filter drivers because
> you can create more problems by making an incorrect Registry change. Either
> make a copy of the entire hive or export the specific key you're working on
> prior to deleting anything. If you have questions about the process, please
> post them here.
>
> Kevin McNiel, MCSE/MCSA
> Platform Server Setup Group
>
> This posting is provided "AS IS" with no warranties, and confers no rights.
> Please reply to the Group only, This address cannot receive incoming
> messages.
>
>
> "Steve Duff [MVP]" <ergodic@ergodic-systems.com> wrote in message
> news:OrO8e5AlFHA.2396@TK2MSFTNGP12.phx.gbl...
> > Having been in your situtation more than once, I feel your despair.
> >
> > First off, the drive order should make no difference to anything. The RAID
> > signatures are on the drives and the controller reads those to determine
> > the
> > striping order. This is required for any reliable hot-swapping. (Unless it
> > is very old
> > the PERC controllers are LSI Logic-based and as such are very solid.)
> >
> > So the fact that the drive order >does< seem to make a difference is
> > suggesting to
> > me that you probably have - or had - a marginal cable or connector - which
> > is about 4 out
> > of 5 system problems in general. Every time you swap drives around you are
> > mechanically altering it which causes drives to drop in and out of the
> > array.
> > Nothing is more detremental to a RAID5 array.
> >
> > RAID5 is a good compromise in theory, but turns out to be very fussy in
> > practice.
> > I no longer even specify it for any application - it offers only mediocre
> > performance
> > and is prone to the very sorts or problems you are having now.
> >
> > At this point you are into a disaster recovery scenario. I would first try
> > all three
> > combinations of two drives, removing a third drive entirely, to see if
> > there is one
> > pair you can boot from in degraded-array mode. If so, you want to rebuild
> > that
> > array onto a fresh third drive (not the drive you have).
> >
> > If that doesn't pan out, you can try an in-place upgrade/install of
> > Windows to try to get a
> > bootable O/S. That will let you at least see what you have left on the
> > rest of the partitions.
> > Before you do this you may want to boot to the repair console and do a
> > chkdsk /p to make
> > sure the MFT is in good enough shape to use. (Prepare yourself for a
> > possible shock there.)
> >
> > After that, you just have to pull out the backups.
> >
> > Steve Duff, MCSE, MVP
> > Ergodic Systems, Inc.
> >
> > "dtremain" <dtremain@discussions.microsoft.com> wrote in message
> > news:521A611A-171B-43B7-91CD-68242890A392@microsoft.com...
> >>I have a situation with a Dell server with a 3 disk Raid 5 configuration
> >>that
> >> I broke down for regular cleaning. When reassembled I could attempt a
> >> log-on
> >> but before the admin profile would load it would freeze. After
> >> attempting to
> >> reboot, one of the drives appeared to be offline. I called Dell's
> >> extended
> >> warrantee department and the tech had me pull the drives, swap the order
> >> of
> >> drives, reconfigure the controller and attempt a reboot - a couple of
> >> times.
> >> After several errors of no operating system I finally got off the phone
> >> with
> >> him and was practicall helpless. I tried a few different orders for the
> >> disks and finally got the Windows screen. Once it gets past that I get
> >> the
> >> following stop error: 0x0000007B (0x85E9D330, 0xC0000032, 0x0000000)
> >> INNACCESSABLE BOOT DEVICE.
> >>
> >> Unfortunately this is the only server and without it the business is
> >> practically paralyzed The Dell tech said that rebuilding is not
> >> something we
> >> want to do, but wouldn't the parity bit help us recover from this?
> >>
> >> I need some serious help here. Thanks in advance.
> >
> >
>
>
>