Problem: RAID 1 system disks will no longer boot to Windows via PCI RAID card. Individual drives WILL boot to Windows when connected directly to Motherboard SATA port. How do I get my RAID restored?
System: XP Pro 32bit SP3, Gigabyte GA-K8NS MB, AMD Athlon 64 3700, 3GB DDR400 Ram, GeForce 6600GT AGP Video, ASUS DVD Burner (SATA port 1 of 2), 1x120GB WD SATA150 "Failsafe" XP installed (SATA port 2 of 2), Generic Silicon Image 3512 SATA150 2-port PCI RAID Controller connected to 2x250GB WD SATA300 disks in RAID 1 mirror config as Primary C Windows Drive.
History: Generally very stable system. Running RAID 1 for system drive redundancy (after getting burned in the past on a system drive failure). At some point in the past few weeks, however, Windows performed an overnight automatic update with an automatic restart, but the system did not completely shut down and I awoke to "windows is shutting down" on the screen for who knows how many hours. I manually pressed the power reset button on the case to reboot. At bootup, I was greeted with a system warning: "RAID in critical state; system will rebuild after bootup" or something very similar. As windows booted to the desktop I was greeted with a "Windows updated" message. However, on all subsequent proper restarts, I was told the same "RAID in critical state; system will rebuild after bootup" message... weeks later, even.
In an attempt to diagnose and solve this problem, I booted up on RAID drive "A" only connected to the RAID card , which worked. I then tried to boot on RAID drive "B" only but I got a BSOD before Windows would load: STOP 0x0000007B (0XF78A2528, etc). I figured this was the corrupted drive of the pair. I booted up on RAID drive A only again okay, but when both drives were connected up at the same time, I again got the BSOD message at boot. Disconnecting drive B did not allow Drive A to boot as it had before... it's like the drive corruption had jumped from the bad drive to the good drive after they were both booted together. No Safe Mode boot, either.
Connecting RAID Disk B to one of my 2 motherboard SATA ports (unplugged DVD-Burner) and Rebooting onto the "Failsafe" Windows drive WD SATA 120, I was able to do an Error Check and Repair on Disk B with system tools, then did the same for RAID Disk A.
Good news, either drive WILL BOOT all the way to Windows when connected to the motherboard SATA port.
Bad news, both Drive A and Drive B WILL NOT BOOT boot when connected to the RAID Card (one at a time). I get the same BSOD as before.
The BIOS sees the drive connected to the RAID card, but neither RAID Disk will boot to windows. If I leave the system connected as such but boot on my failsafe SATA 120 drive, windows sees the RAID card but not the disk. Even Partition Magic does not see the disk on the RAID card.
SO, to summarize, I seem to have successfully repaired both RAID Disks on an individual basis (when run from the Mainboard SATA ports) yet they will not boot or even be recognized by Windows when connected to the RAID card. How do I re-establish RAID sync without destroying the existing Windows install and data on those disks?
My thought at this point is to delete the RAID set (not format the Disks) in the BIOS RAID Utility and try to create a new RAID 1 set with the existing data on Drive A. Just not sure if I will have any better luck getting the disk to boot through the RAID card than I am now.
I have successfully deleted the RAID set via the BIOS RAID Util while keeping my data and OS install perfectly intact. Actually, I had to do this two times; once for Disk A and again for Disk B, as I connected only one drive at a time to the RAID card to perform this operation.
Once the set was deleted, I am now able to boot (singly) from either drive, either from the ports on the RAID card, or the MB SATA ports. It should only be a matter of creating a new RAID set and rebuilding at this point.
I am relieved that I do not have to recover or reinstall everything. Like last time. Sucked.
RAID 1 successfully rebuilt.... only took about 20 HOURS to do an offline rebuild
Up and running again like nothing happened.
It wasn't explicitly mentioned in the original post, but this same problem occurred to me back in September, too. Also caused by a Windows update which did not properly shut down. Hard power reset seems to corrupt the RAID. Drive A would boot at first, Drive B would BSOD at boot. Somehow, in an attempt to diagnose and repair, I corrupted BOTH drives, and could not get either one to boot to Windows. This seems to occur once I reconnect both drives to the RAID card at the same time. I was able to recover my data, but I couldn't figure out how to restore the OS. I don't believe I did the XP Error-Check utility at that time. I ended up formatting and reinstalling XP and all my software. Not my ideal solution.
When I came upon this same problem this time, I ran the Error Check on both drives. I know that Drive B had many errors to fix, Drive A check ran overnight so I never did see the report on error checking... I don't believe it was as bad. I think the Error check may have helped once the problem escalated.
Let's see if I've learned anything:
Don't improperly shut down my PC. Set Windows Update such that it does not automatically shut down my PC.
If I do end up with a RAID problem, and I determine which drive is the corrupted drive, do not reconnect both drives back to the RAID card at the same time because that will corrupt both drives (WTF??). Immediately go into RAID BIOS to do an offline rebuild (and expect to leave the system alone for a whole day). If I do an online rebuild, I am not confident that I will ever get the drives back in sync (I had been operating in this mode for several weeks before trying to diagnose this problem, the sync never having caught up... although the data on the drive was up-to-date, unlike the time previous where it had dropped off and stayed behind by several weeks.)
Be sure to run Error-Check.
If the Drives still will not boot, delete the RAID set from each drive, then create and offline-rebuild a new RAID set in the BIOS RAID Util.
I am not convinced there is a Hard Drive problem per se. I believe the Disk is operating nominally, until an improper shutdown occurs. Then it either becomes slightly unsynced from Master Drive A (and the RAID card complains and slowly tries to re-establish sync as the computer is used) or the MBR becomes corrupted outright. I think it is the cheap RAID card that is to blame, actually. Or Windows for not accomplishing the proper shutdown.
Since the offline rebuild replaced disk data bit-for-bit on Drive B, I'm pretty confident there is no corruption right now.
Nevertheless, I have considered grabbing another WD2500YS 250 GB SATA II, to use as an occasional backup RAID mirror, periodically copying over and keeping off-site. But at 20+ hours for an offline rebuild, this is not a very convenient backup solution. Also seems a bit overkill. I do try to divide and backup my data across many drives, in addition to offsite DVD-R's.
The value of raid-1 for protecting data is that you can recover from a hard drive failure quickly. (not so quick this time)
It is for servers that can't afford any down time.
Recovery from a hard drive failure is just moments.
Fortunately hard drives do not fail often.
Mean time to failure is claimed to be on the order of 1,000,000 hours.(100 years)
Raid-1 does not protect you from other types of losses such as viruses,
software errors,raid controller failure, operator error, or fire...etc.
For that, you need EXTERNAL backup.
If you have external backup, and can afford some recovery time, then you don't need raid-1.
It depends on the controller actually. Good ones, knowing that the data sits on both drives, will read data from both drives. Lower end (read as cheap) controllers will only read from one. Writes of course are single drive only. A good high end controller should be able to read just as fast as a AID0 array. How useful it is depends on how much time you spend reading data from your drives. If it takes 20hrs to repair from this failure, your probably better off using a single normal drive with a USB drive for backup of important data.
RAID 1 on anything but a server is pointless.
What you need is a good backup/disk imaging program like Acronis, and backup your entire drive to another drive. Acronis can be set to do an exact mirror to another drive initially, then it does incremental backups as needed.
If you are worried about your data, this is the proper way to do it, not run RAID 1.
Anything important is burned onto a DVD and kept in a binder and/or a copy is placed on my wifes computer. Because I do have two computers here, I'm not that worried about if one computer goes down for a week while a new drive comes. My computer doesn't bring me money, so I can afford downtime.