This morning, just come into my office, I read my mail that the server sends to the account of administration and with great surprise, I learned that one of the hard drives configured in RAID 1 (mirroring) has abandoned me. The peculiarity of the error is in the "degradation" of all configured RAID 1 array. When an event happens like this, where they degrade the arrays of an entire hard disk, it is highly likely that the disc is damaged and needs to be replaced. But here is what error I was notified through e-mail and as I proceeded to remedy.

On the email I received two messages containing this object:
"DegradedArray event on / dev/md1: server-studio.miodominio";
"DegradedArray event on / dev/md2: server-studio.miodominio";

Given that there are only two on my server RAID array (md1 and md2), dealing respectively with the / boot partition and the /, and the fact that both arrays were degraded to the fuoriscita by mirroring partitions residing on the second hard drive, it is highly likely that the latter is a terminally ill patient. Running the command "cat / proc / mdstat", in fact, were not present partitions hdc (the second hard drive).

When this happens it is essential to proceed first with a nice full backup of your data. And 'what I did, moving my few gigs on some DVDs and a third hard drive. Then I attempted to repair the array this way ...

With the command "fdisk-l" I checked what were the two hard drive partitions to call into the mirror and found this:

Disk / dev / hda: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors / track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/ Dev/hda1 * 1 13 104391 fd Linux raid auto-sensing
/ Dev/hda2 14 30401 244091610 fd Linux raid auto-sensing
Disk / dev / hdc: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors / track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/ Dev/hdc1 * 1 13 104391 fd Linux raid auto-sensing
/ Dev/hdc2 14 30401 244091610 fd Linux raid auto-sensing

As you can see, paying attention to the parts shown in bold, hda1 and hdc1 must be part of the same array (the sizes of the partitions are identical); hda2 and HDC2 must be in the same array.

I continued, then, using the mdadm command like this:

# Mdadm-add / dev/md1 / dev/hdc1

Then I ran the command:

# Mdadm-add / dev/md2 / dev/hdc2

Currently RAID is being rebuilt:

# Cat / proc / mdstat
Personalities: [raid1]
md2: active raid1 HDC2 [2] hda2 [0]
244091520 blocks [2/1] [U_]
[=====> ...............] Recovery = 29.2% (71281408/244091520) finish = 97.6min speed = 29493K/sec
md1: active raid1 hdc1 [1] hda1 [0]
104320 blocks [2/2] [UU]

It can be seen that the raid array md1 has been rebuilt, while the md2 array is rebuilding.
The next step will be to see if the secondary hard drive (hdc) is suffering from some hardware problem. To do this, the software will use smartctl