Question : Problem: Intel Matrix RAID10 degraded drive is not recognized

I messed up my 4 drive RAID10 configuration, and now my system won't boot.

I have a Gigabyte EP45-DS3R motherboard with an intel ICH10R SATA RAID chipset.
I have 4 Seagate 500GB drives connected.
I configured 2 volumes, 120GB RAID10 boot volume, and 800GB data volume.
Vista x64 is installed on the boot volume, and the data volume is used for data files.

I was messing around with the BIOS settings, restore optimized defaults, and this changed the RAID option back to AHCI.
I rebooted to let the optimized defaults take effect, then changed the RAID option back to RAID.

After I rebooted this time the RAID manager reported that the two volumes were degraded, and it reported that disk0 was a non-raid disk, while disk1 to 3 were still reported as part of the two RAID10 volumes.

At this point I would expect the system to function correctly, with no data loss due to the RAID1 config, but without the benefit of RAID0.
Booting the system resulted in a fulure of a missing bootmanager file.

I would also have expected that by pressing Ctrl-I to enter the RAID BIOS config I would be able to add the non-raid disk as the missing disk.
Entering the BIOS gave me no options to add the non-RAID disk as the missing disk.

The Intel KB article states that on entering the Ctrl-I utility I will be prompted to add the drive, I am never prompted, and there are no options to do anything with drive0.
See:
http://support.intel.com/support/chipsets/imsm/sb/CS-021017.htm
 
I booted the system from WinPE, and both volumes are present, bt strangely the boot volume shows up as drive D and the data volume as drive C. If the ntloader interprets the drives the same way, it may explain why the loader is reported as missing.

I booted from a Vista x64 DVD and seletec the repair option, no windows installations were detected.

This leaves me baffled:
1. Why did one boot into AHCI mode loose the RAID drive config and made disk0 a non-raid drive?
2. Why is the Intel RAID BIOS not showing me the dialog to join a disk as explained in the Intel KB?
3. Why would vista fail to boot if 3 out of 4 drives in a RAID10 is present and the volumes are just degraded?
4. The Intel KB on a failed RAID10 also mentions using the Windows RAID Storage Manager utility to reconfigure the drives, but if I can't boot, I can't run this utility.

And most importantly:
a. How do I get my system to boot again?
b. Why did this happen?
c. How do I prevent this from happening again?

Thanks
Pieter

Answer : Problem: Intel Matrix RAID10 degraded drive is not recognized

I found a solution, well, I should not really call it a solution, more like a workaround for what appears to be a problem with the Intel RAID BIOS and with the RAID Storage Manager.

This thread helped me out:
http://www.tomshardware.com/forum/244932-32-ich9r-raid-drive-volume

Basically you have to get a Windows OS to boot from a drive that is not part of the RAID arry, then you connect only the drive that is supposed to be RAID but isn't, then inside Intel Storage Manager you right click that drive and mark it as spare, then you reboot with all drives connected, then the rebuild will automatically start.

Here is a summary what I ended up doing:
1. Vista Ultimate x64 SP1, 8GB RAM, 4x500GB HDD, 120GB RAID10 for OS, 800GB RAID10 for data, Gigabyte EP45-DS3R, ICH10R RAID chipset, Q9650 3GHz.
2. Update BIOS from F9 to F10, load optimized defaults, reboot.
3. The loading of the optimized defaults changed the drive config from RAID back to AHCI, I think this somehow confused the RAID chipset.
4. Changed BIOS config back to RAID, rebooted.
5. Now the Intel RAID BIOS reported that drive-0 is non-RAID, that drive-1 to drive-3 is RAID, and that the two RAID10 volumes are degraded.
6. I entered the RAID config by pressing Ctrl-I, but there were no options to let me move the drive back to RAID.
7. I found an Intel KB that says I am supposed to see a dialog that lets me re-assign the drive, but I had no options to do anything with the non-raid drive, see: http://support.intel.com/support/c [...] 021017.htm
8. So I figured I can still boot since the RAID1 portion of the RAID10 will still work. Two things went wrong.
9. Booting Vista gave me a boot volume can't be mounted bluescreen error. I think this was because the non-raid disk was listed before the raid boot volume in the BIOS.
10. I changed the BIOS to put the single non-raid drive after the two raid volumes.
11. I booted again and this time Vista told me it can't find winload.exe.
12. I also tried booting with the non-raid drive unplugged, same problem.
13. I booted from the Vista DVD and selected repair, but the repair could not find any Vista installs.
14. I ordered a replacement drive and paid for next day delivery, turns out I don't need it, but I will set it up as a spare.
15. I wanted to get the Intel Storage Manager software running, but I need a working OS to do this.
16. I plugged in an external eSATA drive, unplugged all internal drives, and installed Vista and the Intel version 8.6 storage manager on the eSATA drive.
17. I plugged all the internal drives back in, and booted from the eSATA drive.
18. On logging in the Intel storage manager tray icon told me that the drive is degraded, and that I must right click on the new drive and to select rebuild.
19. I right clicked on the non-raid drive and selected it for the rebuild, but got the stupid error that lists the various possible causes of the problem, but not the actual problem.
20. Intel, this is the dumbest error message ever, I want to know exactly what failed, not some list of possible causes.
21. One of the possible causes mentioned was that there is data on the drive. I opened disk manager and saw that the non-raid disk had an assigned drive letter, and had a 120GB partition on the 500GB disk.
22. I deleted the partition, rebooted, tried again, and the same winload.exe error.
23. Now that I had the (stupid) error dialog, google pointed me to this thread.
24. This is what really helped me; I removed all but the non-raid drive, and selected that disk for use as a spare.
25. I shutdown, reconnected all drives, rebooted, still booting on the esata drive, this time the rebuild automatically started.
26. After a few hours the rebuild completed, I ran chkdsk on the raid volumes, the first volume had some errors, repaired them, the second volume was fine.
27. Now I had to get Vista booting from the original boot volume again.
28. I changed the BIOS settings to place the raid volumes first, and I rebooted, but I still got the winload.exe error.
29. I booted from the Vista DVD, and selected repair, this time the repair tool found two install, the one on the boot volume and the one on the eSATA drive, and performed some repair actions.
30. On booting again I had two Vista Ultimate (repaired) OS boot options, the first entry booted from the eSATA drive, the second entry booted from the RAID boot volume.
31. I powered down the eSATA drive and rebooted. I ran bcdedit.exe and deleted the first entry.
32. After rebooting the correct OS automatically loaded.
33. I think if I had ran the boot repair with the eSATA drive disconnected I could probably have skipped the bcdedit step.

This still leaves me with some questions:
1. Why did the drive turn into a non-RAID drive, this is of greatest concern to me for I am bound to update the BIOS again, and switching from RAID to AHCI is bound to happen, intel bug?
2. Why did the Intel RAID BIOS not allow me to fix the problem from the BIOS, but forced me to do it from Windows, Intel bug?
3. Why exactly did the right-click on the drive fail, and required that the drive be individually marked as a spare, Intel bug?

If somebody can provide insight they can have the points.