Server did not boot with until I unplugged second mirror RAID drive

[Bob Simon]

I have not worked with RAID before so I hope someone can explain what
happened to cause the events I observed today in the office.

The computer is running Windows Server 2003. On startup, the system
showed that the mirror was valid. It successfully completed POST then
hung shortly after Windows displayed the splash screen. The system
did not respond to Ctrl-Alt-Del. Keyboard indicator lights would not
even toggle when I pressed CapsLock, ScrollLock, or NumLock.

I tried to reinstall Windows (repair) but after F6 and loading the
RAID drivers the system hung again when scanning the HD. I unplugged
the second mirrored drive and W2k3 started up fine.

Some critical file was apparently damaged on the second drive. What
typically causes this type of problem?

I would have expected both drives to be identical. What are some
probable reasons they were not?

 

[Michael Hawes]

I have not worked with RAID before so I hope someone can explain what
happened to cause the events I observed today in the office.

The computer is running Windows Server 2003. On startup, the system
showed that the mirror was valid. It successfully completed POST then
hung shortly after Windows displayed the splash screen. The system
did not respond to Ctrl-Alt-Del. Keyboard indicator lights would not
even toggle when I pressed CapsLock, ScrollLock, or NumLock.

I tried to reinstall Windows (repair) but after F6 and loading the
RAID drivers the system hung again when scanning the HD. I unplugged
the second mirrored drive and W2k3 started up fine.

Some critical file was apparently damaged on the second drive. What
typically causes this type of problem?

I would have expected both drives to be identical. What are some
probable reasons they were not?

What drives? SATA, SCSI? Motherboard or add-in RAID card? The drive you
have disconnected may have failed in a way that prevents the other drive
working. If you disconnect the working drive and reconnect the faulty drive,
see if detected and run manufacturers diagnostic on it. If passes, wipe it
and startup with both drives. Configure the wiped drive back in to array and
set to rebuild.If it failed, replace the drive and do the same.

Mike.

 

[Arno Wagner]

I have not worked with RAID before so I hope someone can explain what
happened to cause the events I observed today in the office.

The computer is running Windows Server 2003. On startup, the system
showed that the mirror was valid. It successfully completed POST then
hung shortly after Windows displayed the splash screen. The system
did not respond to Ctrl-Alt-Del. Keyboard indicator lights would not
even toggle when I pressed CapsLock, ScrollLock, or NumLock.

I tried to reinstall Windows (repair) but after F6 and loading the
RAID drivers the system hung again when scanning the HD. I unplugged
the second mirrored drive and W2k3 started up fine.

Some critical file was apparently damaged on the second drive. What
typically causes this type of problem?

I would have expected both drives to be identical. What are some
probable reasons they were not?

The RAID driver should not show you two disks, so Windows cannot
see anything different on the second drive. I would rather suspect
that there is something wrong with the second drive or its cabeling and
that the driver was written by incompetents and locks up instead
of reporting the problem.

Arno

 

[Bob Simon]

What drives? SATA, SCSI? Motherboard or add-in RAID card?

Seagate SATA drives. RAID done via software.

The drive you
have disconnected may have failed in a way that prevents the other drive
working. If you disconnect the working drive and reconnect the faulty drive,
see if detected and run manufacturers diagnostic on it. If passes, wipe it
and startup with both drives. Configure the wiped drive back in to array and
set to rebuild.If it failed, replace the drive and do the same.

The system is only six months old so the vendor replaced the bad drive
and mirrored it Thursday night. The server was put back in service
Friday morning and all is fine now. I just want to understand why if
one drive developed bad sectors, it would keep the system from
starting up.

 

[Bob Simon]

The RAID driver should not show you two disks, so Windows cannot
see anything different on the second drive. I would rather suspect
that there is something wrong with the second drive or its cabeling and
that the driver was written by incompetents and locks up instead
of reporting the problem.

Arno

I believe I proved that the second drive had some bad sectors since
the system started up with only drive 1 enabled. Your suggestion of a
defective RAID driver is consistent with what I saw. I had not
considered that possibility until your comment. Thanks Arno!

 

[Arno Wagner]

The RAID driver should not show you two disks, so Windows cannot
see anything different on the second drive. I would rather suspect
that there is something wrong with the second drive or its cabeling and
that the driver was written by incompetents and locks up instead
of reporting the problem.

Arno

[/QUOTE]

I believe I proved that the second drive had some bad sectors since
the system started up with only drive 1 enabled.

Bad sectors are one possibility. Might be the most likely one.

Your suggestion of a
defective RAID driver is consistent with what I saw. I had not
considered that possibility until your comment. Thanks Arno!

No problem.

Arno

 

[Arno Wagner]

On Fri, 10 Aug 2007 16:22:22 +0100, "Michael Hawes"

Seagate SATA drives. RAID done via software.

The system is only six months old so the vendor replaced the bad drive
and mirrored it Thursday night. The server was put back in service
Friday morning and all is fine now. I just want to understand why if
one drive developed bad sectors, it would keep the system from
starting up.

It should not. Linux, e.g., kicks the drive from the array after a
number of errors and timeouts. Any good RAID controller will do the
same, be it a hardware or software controller.

Arno