I have a new system with a boot drive and 2 drives that I want to use for data. On the 2 drives I setup a mirrored volume and a spanned volume. I wanted to test the robustness of LDM in Windows (better now than when there's actual data that I want safe). I pulled one of the disks and rebooted (simulating a cable failure). The mirrored volume came back up with "failed redundancy", the spanned volume came up as failed, and a "missing" disk shown, all as expected. Shut down, reattach the second drive. I get my Disk2 called "foreign" and the "missing" drive is still shown. The missing drive IS the foreign drive! "Reactivate volume" fails with "the plex is missing" message. Same error for "reactivate disk" on the missing disk. "Import foreign disk" shows "foreign disk group (1 of 1 disks) -- wrong it should be 1 of 2. It sees the volume info but proceeding gives the warning dialog "Some of the volumes you are importing will lose data because you have not moved all your disks to this system". Ahh -- this isn't the data safety that I was hoping for!! I downloaded ldmdump (http://technet.microsoft.com/en-us/sysinternals/bb897413.aspx). And it looks as if windows has changed the disk group ID of the disk that remained: Disk 1 "PRIVATE HEAD" relevant data: Disk Id : ab2e4dc7-e61d-11de-aad5-00241dd53e77 Host Id : 1b77da20-c717-11d0-a5be-00a0c91db73c Disk Group Id : 71af879b-e81a-11de-b75f-00241dd53e77 Disk Group Name : KURTSHTPC-Dg0 Disk 1 "VBLK" disk data: 0x000011: [00000C] <Disk> Name : Disk1 Object Id : 0x0002 Disk Id : ab2e4dc7-e61d-11de-aad5-00241dd53e77 0x000012: [00000D] <Disk> Name : Disk2 Object Id : 0x0003 Disk Id : ab2e4dca-e61d-11de-aad5-00241dd53e77 Disk 2 (the one that was pulled, but now is "foreign"): Disk Id : ab2e4dca-e61d-11de-aad5-00241dd53e77 Host Id : 1b77da20-c717-11d0-a5be-00a0c91db73c Disk Group Id : ab2e4dc6-e61d-11de-aad5-00241dd53e77 Disk Group Name : KURTSHTPC-Dg0 So the online Disk1 knows it's Disk2 has the same ID as the "foreign" disk. The Disk Group Ids are different though. Did windows change the first disk's disk group id when it thought disk2 was gone forever? Is there a way to write disk 2's LDM data to match Disk1's group ID?? I'm guessing that would just create other problems... Not impressed by the robustness... Thanks for any help! -Kurt

Replying to myself for your googling benefit: I went ahead with the "import foreign disk" despite the scary message. The spanned volume came right back online, no problem. The mirrored volume came back online as 2separate simple volumes and the second was given a new drive letter. I deleted the new simple volume intending to mirror the first in the new empty space. I clicked create spanned volume by mistake. I think if I had clicked create mirrored volume instead that I would have been back to normal.

EXACTLY the same behaviour here. I'd booted with one disk disconnected, having had c: mirrored, and a large spanned d: The state after replacing the drive got me somewhat worried (although I have backups), but after importing the foreign disk I also ended up in same state - spanned fine, mirror needing recreating. There's clearly a bug here -- the id should be the same, and a simple reactivate should suffice. Am using WIndows 7 x64 with all windows-update fixes........

Similar problem here, running Win 7 x64 Ultimate. Although I had a reboot from a BSOD instead of a disconnected disk that broke the mirror array. Without reading the above posts I don't know if I would have tried to import the foreign disk because of the dire warning messages. But I did the same steps as above and my mirror is resynching at the moment. The reason I'm using this software RAID is that my current motherboard's onboard RAID controller ended up being flaky. I haven't been thrilled with this method either -- too many resynchs that take hours and slow down my system.

I ran into about the same scenarios as described... I've got a very simple setup, running Win7 x64 Pro on two 1.5TB drives set to Dynamic, with a System Reserved volume and the C: volume, and both are mirrored across the two drives. I removed one drive and booted, then reconnected and booted again, and the reconnected drive shows up as Foreign, plus the system still shows an extra drive entry as the missing drive. Importing the Foreign drive gets me to where the reconnected drive shows the original two volumes now as simple volumes, and no indication of any mirroring. The volumes on the unaffected drives also appear to have lost any knowledge of prior mirroring. I noticed that the reconnected drive is initially still listed as Dynamic... but when I deleted the simple volumes to prepare for recreating the mirrors, the drive suddenly changed back to Basic. I carried on, recreating the mirrors on both volumes, and right now the C: volume is still resynching... which actually should say "synching" because "resynching" implies it's recovering from a prior mirror, but instead I'm actually creating a new mirror again. Still, this looks exactly as it did when I first created the mirrors, so I'm assuming things will work normally from here. Has anyone else had any better experiences than this? I'm in agreement with the previous posters, this appears basically functional but with some bugs on the recovery activities.

Similar behavior, but on WHS 2011. This has happen to me three times, this is a very uncomforting bug. Last time this happened was after a reboot to finish installing some microsoft updates, was not playing with the hardware! This is a WHS 2011 machine with three drives, the boot drive and a two 2 TB mirror. I have a similar Windows 7 Pro machine that so far has not exhibited this behavior yet, different motherboard. I have decided to break the mirror on the WHS 2011 machine and do a lazy (once a day) mirror sync, as the data on these drives does not change often. Ron

Has microsoft fixed this problem yet? Anybody knows? Thanks in advance, BJBJ

Updating this old thread to keep it all together, as I too seem to have been plagued by this or similar issue. Win7 64 Bit. 8 GBs RAMDual eSata card Dual eSata external enclosure with 2 matching 1TB drives. Software RAID 1 in W7, worked fine for about a year. 1 drive on the Raid died. I removed it, put it into a external drive caddy, yes, it's dead. The other one still operates fine without removal from the external case. The message in Storage management is that the drives RAID had failed (or words to that effect). I replace the dead drive with a new 1 TB drive. I attempt to restore the RAID. Message is "the plex is missing". Nothing more. No options to do anything but cancel. I found this site on the web. So WTF does it want you to do about a 'missing plex'? I assume I put two drives that I have emptied into the enclosure and recreate the mirror again. But I thought that adding a new drive into the RAID allowed recovery automatically. What is the solution next time? I didn't try the Import the foreign volume routine though. Didn't know it was an option. Seems similar to Sanhoser's issue, and maybe the OP. Hope this helps MSFT get to the root of this, or give us better restoration efforts. I think that next time I'm going to buy an external case with all RAID internal to the case and just use a single esata port to run the data in and out. Get Windows out of the way of running my RAID.

Same "plex is missing" error here. I'm not very happy with this Win 7 RAID solution, the only reason I paid an extra $40 for Win 7 "Pro". 1.) Is there a way to get notified when a RAID volume has gone offline? I only found the problem by checking manually. Not something I want to do every night. 2.) I couldn't find a way to simply re-syncronize the volume that had gone offline. Had to erase/repartition the volume and re-sync. With only 200GB of data, took several hours. Thinking of investing in a FreeNAS solution...

Another bug is the inability of Windows to inform you that your drive has failed. I've searched for hours on this trying to find a way to do it, but it's pointless and MS publicly knows about the issue. On the other hand, when I used Intel's Matrix Storage Manager, it continuously rebuilt the drive every time I came out of sleep because there is no configurable timeout for a drive powering up, and they (like MS) see no reason to fix such things. I guess when you get as big as those companies, quality is no longer a concern. I literally know of no other reasonable way to raid my drives than these two programs. This whole "no plex" issue just hit me. I actually didn't know my drive had failed until Norton Ghost kept annoyingly asking me if I wanted to run a backup on my D drive over and over, even though I selected yes. I saw that Norton showed the disk as "Unavailable" but it also shows my backup drive where the backups are stored as unavailable all the time, another quality issue, so I just thought it was a fluke at first. Had to dig into the disk manager to find the failed drive. Speaking of quality, the two drives I am mirroring are 60 GB SSDs Corsair F60s. The sync takes about 5 minutes for 60 GB, so at least they're fast. I have had to return under warranty 4 of these in the last 2 years, so I'm not surprised one has already failed again. I purchased a Samsung 830 series 120 GB for the windows boot drive, but alas recently windows comes out of sleep and can't find that drive now either. It's tough running a small business when you can trust NOTHING to secure your information.