We have an old MD1000 (out of warranty) with 15 disks, one of which is a global hot spare (disk 0), the rest forms a RAID5 volume. Each disk's size is different, but the majority (lowest common value) is 700GB, and we made a single volume of: (15 - 2) x 700 = 9.1TB. Then...
1. One day, disk 11 failed, and the auto rebuild using disk 0 (hot spare) occurred.
2. When rebuilt was completed, user checked the volume but found some files are 'missing'.
3. A new disk 11 is put in & made a new global hot spare.
4. At this point, the host, a Dell 2950, complained the entire volume is unavailable. A check of Dell OMA reported ALL disks are marked as 'foreign'.
5. We accessed the PERC 5/E adapter which connects to MD1000, and requested a controller 'import/recover foreign config'. Command ran, and it came back with disk 0 & 11 as 'available' & got assigned as global hot spares, all other disks remained as 'foreign'.
6. We again issued a second & subsequent 'import/recover foreign config', and process failed with error: 'operation could not be performed'.
7. We repeated the process, while server reboots to text-only BIOS mode, using the PERC 5/E config screen, but the option to 'import config' is unavailable - probably because this is a PERC 5/E, and not PERC 6.
Next, we double-checked & upgraded the firmware, BIOS of Dell 2950, MD1000, PERC 5/E, OS (it runs Windows 2003 32-bit) all to the latest versions.
Now, we ran out of ideas of what else can be done to recover the volume, even if it is recovered to a non-optimal config - that is a volume in RAID5 but no parity drive, using all disks minus disks 0 & 11.
Have anyone come across this or similar scenario, or what would you suggest we can try?