[ale] Where Did The Disk Go???

[Phil Turmel]

Hi Gene,

On 11/27/2013 11:18 AM, Gene Poole wrote:
> I'm running CentOS 5.9 x86_64 on a machine I built myself that has 6 SATA 
> II hard drives (4 - 1 TB drives; 2 - 1.5 TB drives) all in several RAID-1 
> arrays. These arrays were created when I did the original installation 
> with CentOS 5.1 and each created partition (both standard and LVM) were 
> built raid 1.
> 
> Due to some things happening around the house that required most of my 
> attention, I saw some alerts concerning drive /dev/sdb but I didn't have 
> time to address it and since I am running raid 1...  One of the tings that 
> happened was my home took a lightening strike and we were down for 4 days.
> 
> When we got our electricity back and I brought the machine back up I 
> noticed that there were only 5 drives listed  and I wasn't getting any 
> more alerts.  What was /dev/sdb was no longer listed with any command 
> (fdisk; cat /proc/mdstat; mdadm --list).

Before you do anything else, document what you have.  I suggest you
start with "lsdrv" [1] for the bird's-eye view, then:

"smartctl -x" for each drive,
"fdisk -l" or "gdisk -l" for each drive,
"mdadm -D" for each array,
"mdadm -E" for each array member

> I do have a spare drive and my questions are:
> 
> Anyone know how I can find the serial number of the bad drive?

After the above, turn off the system and look at the drive stickers.
Find the one that doesn't match anything documented.

> Once replaced, what is the best way to get the partitions recreated?

If your drives are still MBR, I'd just "dd" the pre-partion 1 area onto
the new drive and issue "blockdev --rereadpt" (the new drive must be the
same size as the original).  If your drives have GPT partitions, use
gdisk's backup and restore features.

> Thanks,
> Gene Poole

HTH,

Phil

[1] http://github.com/pturmel/lsdrv