
6 Software RAID with Red Hat Enterprise Linux 4
www.dell.com | support.dell.com
Restoring the RAID Configuration After a Drive Failure
When a hard drive in a software RAID 1 array fails, you can restore the RAID array onto a new drive
by following a three-step process:
• Replacing the failed drive
• Partitioning the replacement drive
• Adding RAID partitions back into the md devices
If you suspect a drive failure, you can check the status of each RAID device by using the
following command:
cat /proc/mdstat
For example, a system missing a partition from the md0 device would show the following:
md0 : active raid1 sda1[0]
104320 blocks [2/1] [U_]
This output indicates that
md0
is active as a RAID 1 device and that partition
sda1
is currently
active in that RAID device. The output [
2/1
] denotes that two partitions should be available to
the device (the first value), but only one is currently available (the second value). The output [
U_
]
shows that the first partition is available (denoted by the letter "U") and the second partition is
offline (denoted by the underscore).
Replacing a Failed Disk Drive
Once a hard disk drive fails, replace it immediately to preserve the data redundancy that
RAID1provides.
See your system documentation for instructions on replacing your failed hard drive with a
new hard drive.
Partitioning the Replacement Drive
Once the failed disk drive has been replaced, restore the partitions that were saved earlier in
the
/raidinfo
directory.
For replacing drive
sda,
restore the original partition scheme for
sda
to the new hard drive
by typing:
sfdisk /dev/sda < /raidinfo/partitions.sda
For replacing drive
sdb,
restore the original partition scheme for
sdb
to the new hard drive
by typing:
sfdisk /dev/sdb < /raidinfo/partitions.sdb
Commentaires sur ces manuels