howto replace faulty disk in intel matrix raid 1

gerbra

Honorable
Dec 19, 2013
2
0
10,510
I have a hardware raid 1 (mirrored) configuraton with 2 disks. Intel Matrix Storage raid controller.

The OS is linux/debian.

SMART is reporting that one disk has exceeded threshold values for surface defects. Hard Disk Sentinel recommends immediate replacement.

I have read the Intel Matrrix Storage manual. I am now seeking further confirmation before I proceed, as I don't want to make any mistakes when I replace the drive.

(1) Must the replacement disk be the same model, or at least have the same geometry (1938021 cyl, 16 heads, 63 sectors)? Or is it sufficient if it has equal or greater capacity?

UPDATED: I've just got an answer to this from http://www.tomshardware.com/forum/259807-32-raid: It's OK if the replacement disk is equal or greater capacity.

(2) Is the following procedure correct? Have I omitted anything? Are there any traps I should be careful not to fall into?

(a) Backup all partitions. Make sure I have a backup of the directory tree and files and not just partition images. Be confident that the backup is restorable.

(b) Remove the faulty disk from the RAID matrix. (Ctrl-I while booting to get into the Matrix Storage Manager ROM). I know the serial number of the faulty disk. Be careful to remove the correct disk. Do not convert to non-RAID, just remove the faulty disk.

(c) How will RAID react when it has only one disk? It can't be a RAID 1 configuration.

(d) Replace the faulty disk by a new disk. Do I need to partition the disk first, to match the partitions on the good disk? Or will Matrix Manager do that?

(e) Start the computer. Ctrl_I to get into Matrix Manager ROM. Select recovery options. Rebuild the new disk from the good disk. Wait for it to finish then continue with the startup.

Do I need to add the new disk to the RAID array before selecting the recovery options?

Have I forgotten anything important?

Thank you. Gerry
 
Solution
1) Yes as long as its the same size

2a) Yes that is always wise but not 100% needed

2b) Yes just remove it, turn it off, add the new one, turn back on and re-enter Raid Config

2c) Does matter. A raid 1 will still funtion. It will say Degraded because everything is on that one hard drive. Your pretty much running on just one drive right now.

2d) NO! Do NOTHING with the drive just add it

2e) Yes it should have the option to add the new disk to the current raid and it will rebuild. You should be able to use your system while it is doing this as well. If you have the time to leave it off its best to let it rebuild while in the Config. It will go faster and with out issue.

I Don't have a PC with Intel Raid that i can mess with and...
1) Yes as long as its the same size

2a) Yes that is always wise but not 100% needed

2b) Yes just remove it, turn it off, add the new one, turn back on and re-enter Raid Config

2c) Does matter. A raid 1 will still funtion. It will say Degraded because everything is on that one hard drive. Your pretty much running on just one drive right now.

2d) NO! Do NOTHING with the drive just add it

2e) Yes it should have the option to add the new disk to the current raid and it will rebuild. You should be able to use your system while it is doing this as well. If you have the time to leave it off its best to let it rebuild while in the Config. It will go faster and with out issue.

I Don't have a PC with Intel Raid that i can mess with and see what exactly need to be done. There should be an option in there and tell its to use that new hard drive and rebuild the raid.

Also the hard drive has to be EXACTLY the same size or BIGGER down to the BIT! If one drive is like 233.8 and the new drive is 233.7 it won't work.
 
Solution

gerbra

Honorable
Dec 19, 2013
2
0
10,510
Good morning drtweak. Thank you. I am about to order a replacement disk and install it. I will post a response when I finish. There seems to be a lack of step-by-step HOWTOs related to disk replacements in raid arrays. It might be useful if I document exactly what I do and post it.

ADDED: At least, there is not much step-by-step HOWTOs for hardware (Intel) raids, but I have found a couple of HOWTOs for a software raid, but the procedures are quite different.

Gerry
 
yea. If anything what i would do is remove bad drive, install new drive, boot into windows, and then use the Intel Raid management software. Search for that. This way you won't accidently mess anything up. Also you should be able to just right click on the Non Raid drive in the Management software and tell it to use it in the Raid and it will add it and boom all done.