i have a RAID 5 array consisting of drives 1-4, and drive 0 which is a hot fix. now drive 4 has broken (red light in the array, log copied below) and drive 0 has 'stepped in', the system is saying 'missing hot fix'.
i assume i need to replace the dead drive 4 with a new HDD, make it part of the array again and then revert drive 0 back to being the hot fix. my hardware support guy's reply is 'just replace dead drive 4 and make it a hot fix'. is this correct? surely that will change the arrangement of the array making drive 4 as the hot fix, rather than drive 0 which it always has been. shoud i do this?
what is the correct procedure for replacing the broken drive 4 and having a hot fix drive, and through what interface do i do the fix (SCSI BIOS, Storage console or storage console plus?).
LOG COPIED FROM STORAGE CONSOLE:
0 Mon Nov 05 10:01:05 2007 1 13 Array Drive 1: Logical Drive 4 <SCSI-A, ID 4, LUN 0> failed
0 Mon Nov 05 10:01:05 2007 1 14 Array Drive 1: FAIL state entered
0 Mon Nov 05 10:01:25 2007 1 22 Array Drive 1: Hot Fix 0 activated
0 Mon Nov 05 10:01:25 2007 1 16 Array Drive 1: failed drive replaced by Logical Drive 0
0 Mon Nov 05 10:01:25 2007 1 41 SCSI-A, ID 4: Auto Hot Plug for Hot Fix started for slot 4
0 Mon Nov 05 10:01:25 2007 1 26 Array Drive 1: drive rebuild started
0 Mon Nov 05 12:59:55 2007 1 21 Array Drive 1: drive rebuild finished successfully
Now, as far as removing the defective drive from the configuration, it all depends on the controller and the software. Most will simply recognize the new hot spare as soon as it's replaced and you won't have to configure anything. Some do not, however and will require you to manually add a hot spare back into the fold. Even if that were the case, you don't have to worry about the order. You can configure the new hot spare first and then remove the old one if it's still hanging around. My guess is that it wouldn't, though.
The goal for array software is bare minimum administrative effort. The software should be smart enough to know that the array after the failure should look the same as it did before the failure.