Hey Brink,
So, I've got Windows 7 on this raid 1 array that is buggy. I've had a couple of crashes but managed to recover so far. I want to replace the disks, as you might guess. I have a standby disk installed in the same machine (all of the disks are 500GB). Is this the way to go? Create a backup, replace disks, and then recover? Or should I not be able just to replace a disk with a new disk somehow, since it is a raid 1 array? (I actually cannot see that happening, since the array wouldn't be recognized with a new disk in place.)
Sorry to impose on you like this. Any suggestion would be greatly appreciated.
---------------------------------------------------
Update on progress so far...
I ran into a fairly common problem, it seems, with running the system image backup. The routine goes through the typical preparatory stages successfully, even to the point of determining the needed space for a backup, in my case less than 10% of the available. Fine. Then the routine starts in earnest, with the typical sliding progress bar for my amusement. Around 5% of the task being completed, an error pops up declaiming my lack of available disk space and rebuking me for not having available space amounting to about 1/10 of that which, in fact, I have available.
After quite a lot of searching for fixes (I won't bother to report the error code, since this seems relatively arbitrary, that is, the solution apparently fixes many differently coded errors), I ran across a discussion of the USN journal, and how that pops its ugly little head up during the process. It seems the Maximum Size reported in the journal for the SR partition (my z: drive) was a ridiculously high value (indeed larger than my c: drive, which fills the remainder of the 500GB disk).
Various solutions are mooted for the fix of this problem, the most elegant of all being the deletion of the journal via:
fsutil usn deletejournal /d z:
I have now found that after a while, without the disable switch (/n), the journal comes back with a more reasonable Maximum Size.
As a caveat, I cannot claim that this is a direct fix of the problem, since I do not know how the USN journal might come into play with the backup procedure. I do know that the process of deleting the journal worked for me, and has not had any negative side effects (so far). In fact, I do not know what a/the USN is, how used, why stored. But just such questions keep me out of the bars and off the streets.
I will update again once my new drives arrive. I hope this post will be useful to those a little shy of this procedure.
--------------------------------
Final Update
Could not do it. After hours of trying to get the recovery program to recognize my new raid 1 array, I had to give up.
I was able to format and partition the array using the recovery program's DOS window, and the recovery program found the drive, started the process, and then a pop-up error would inform me that 'no drive could be found that was suitable to take the recovery'. I wish I could be more specific, but I couldn't figure a way to get a screen grab of the error message. I do remember that it gave a checklist of three options for fixing the problem, none of which applied to the current situation. Could it be that it is necessary to actually do an install of Win 7 onto the array before you can recover to it?
The whole mess was pretty frustrating, and has called into question why anyone would bother to make a system image, given that it is so difficult to use in an actual recovery situation.
For my machine, however, all was not lost since I actually had a functioning system disk. As you will remember, I was replacing a raid 1 array that had crashed a couple of times, though each time had been able to rebuild itself. The procedure I used to replace this array was to replace one disk at a time, and let the IRST procedure rebuild the array (rinse and repeat). As time consuming as this was, I would suggest it as the first thing to try for anyone replacing a raid 1 array.
I would like to hear some comments on this. Especially from anyone who has recovered to a blank array successfully.