Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We have a RAID 5 setup (Megaraid SAS ROMB) with 4 930 Gb physical disks running as a 2.7 TB VD. Recently (a few months back) we had a drive fail and had to leave it in degraded mode for ages (various factors, wish we didn't but it's not in the scope of the question). We replaced the disk and rebuilt. The last few weeks we have noted persistent unrecovered medium errors on one of the physical disks - basically the same error that predated the first failure, and as frequently as right before it happened (roughly 5 incidences per minute logging 4-8 errors per incident). Thankfully budget is in a much nicer place right now, so we have a replacement disk ready.

The only problem, the only thing new to me here, is that we are performing the steps on an array which is still, as far as MegaRAID Storage Manager is concerned, Optimal. There's lots of information available around the place on replacing a disk in a degraded array (not to mention I would just be repeating the same steps I've done before) but I can't find any extra info on how to do it proactively, before the drive fails.

I believe that the process is essentially the same, with an added first step, but I'd really appreciate correction if I've missed the mark here:

  1. In Storage Manager, mark the physical disk in question as offline (if I have the right end of the stick, this will cause the VD to go into degraded mode)

  2. Turn off the computer

  3. Replace the drive

  4. Turn on the computer

  5. Begin the rebuild

Now, how badly have I mangled those steps? (this might be worth noting, and is the reason for my double checking here, the computer is question is also the DC)

share|improve this question
If you know how to put out a fire if a house is on fire, how do you put out a fire if the house is not on fire? Easy ... you start a fire first. – David Schwartz Mar 15 '13 at 7:18
up vote 1 down vote accepted

I've managed/mangled a number of RAID 5 arrays over the years and I always take big breath when working this scenario. You should be able to take those exact steps (without step 1) and the controller should figure out what you did and rebuild the array. Basically the controller is going to see that the drive at slot is missing and that the new one needs to be folded into the array.

I think anyone would agree that before taking these steps it's always a good idea to back up the critical data first - I mean I'm sure you already are :)

What should be the worse case is that the controller doesn't do anything with the new drive and it just runs the array degraded until you replace the missing drive. If this were an older controller I would be more afraid but given this a a late model controller it should definately be able to do this intelligently.

share|improve this answer
Definitely have backups. So I basically treat it as though the controller had already detected that drive as failed, it seems? – Chris O'Kelly Mar 15 '13 at 5:03

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.