6.7 Configuring RAID 5: Disk striping with parity
RAID 5, disk
striping with parity, offers fault tolerance with less overhead and
better read performance than disk mirroring. To configure RAID 5, you
use three or more volumes, each on a separate drive, as a striped set, similar to RAID 0. Unlike RAID 0, however, RAID 5 adds parity
error checking to ensure that the failure of a single drive won’t bring
down the entire drive set. In the event of a single drive failure, the
set continues to function with disk operations directed at the remaining
disks in the set. The parity information can also be used to recover
the data using a process called regeneration.
RAID 5 works like this: Each time the operating system writes to a RAID-5
volume, the data is written across all the disks in the set. Parity
information for the data, used for error checking and correction, is
written to disk as well, but it’s always written on a separate disk from
the one used to write the data. For example, if you are using a
three-volume RAID-5 set and save a file, the individual data bytes of
the file are written to each of the disks in the set. Parity information
is written as well, but not to the same disk as one of the individual
data bytes. Thus, a disk in the set could have a chunk of the data or
the corresponding parity information, but not both. This, in turn, means
that the loss of one disk from the set doesn’t cause the entire set to
fail.
Like any type of RAID, RAID 5 has its drawbacks as well. First, if
multiple drives in the set fail, the entire set will fail and you won’t
be able to regenerate the set from the parity information. Why? If
multiple drives fail, there won’t be enough parity information to use to
recover the set. Second, having to generate and write parity
information every time data is written to disk slows down the write
process (and, in the case of software RAID, reduces processing power).
To compensate for the performance hit, hardware RAID controllers have
their own processors that handle the necessary processing—and this is
why hardware RAID is preferred over software RAID.
OK, so RAID 5 gives you fault tolerance at some cost to performance.
It does, however, have less overhead than RAID 1. By using RAID 1, you
have 50 percent overhead, which effectively cuts the amount of storage
space in half. By using RAID 5, the overhead depends on the number of
disks in the RAID set. With three disks, the overhead is about
one-third. If you have three 750-GB drives using RAID 5, you use 2250
GBs of space to store about 1500 GBs of information. If you have
additional disks, the overhead is reduced incrementally, but not
significantly.
To create a RAID-5 set, start Disk Management. In Graphical View,
press and hold or right-click an area marked Unallocated on a dynamic
disk and then choose New RAID-5 Volume. This starts the New RAID-5 Volume Wizard. Tap or click Next. Create the volume . The key difference is that you must select free space on three or more separate dynamic drives.
When you tap or click Finish, you’ll return to the main Disk Management window and Disk Management will create the RAID-5
set. During the creation of the mirror, you’ll see a status of
Resynching. This tells you that Disk Management is creating the RAID-5
set. When this process finishes, you’ll have three or more identical volumes,
all of which will show the same drive letter in Disk Management. Users,
however, will see the RAID-5 set as a single volume. The volume status
should be listed as Healthy. This is the normal status for volumes. If
the status changes, you might need to repair or regenerate the RAID-5
set .
6.8 Breaking or removing a mirrored set
Windows Server 2012 provides two ways to stop mirroring. You can
break a mirrored set, creating two separate but identical volumes. Or
you can remove a mirror, which deletes all the data on the removed
mirror.
To break a mirrored set, follow these steps:
-
In Disk Management, press and hold or right-click one of the volumes in the mirrored set and then choose Break Mirrored Volume. -
Confirm that you want to break the mirrored set by tapping or
clicking Yes. If the volume is currently in use, you’ll see another
warning dialog box. Confirm that it’s OK to continue by tapping or
clicking Yes.
Windows Server 2012 will then break the mirrored set, creating two independent volumes.
To remove a mirror, follow these steps:
-
In Disk Management, press and hold or right-click one of the volumes
in the mirrored set and then choose Remove Mirror. This opens the Remove
Mirror dialog box. -
In the Remove Mirror dialog box, select the disk from which to remove
the mirror. If the mirror contains a boot or system volume, you should
remove the mirror from the secondary drive rather than the primary. For
example, if Drive 0 and Drive 1 are mirrored, remove Drive 1 rather than
Drive 0. -
Confirm the action when prompted. All data on the removed mirror is deleted.
6.9 Resolving problems with mirrored sets
Occasionally, data on mirrored volumes can get out of sync.
Typically, this happens if one of the drives in the set goes offline or
experiences temporary I/O problems and, as a result, data can be written
only to the drive that’s online. To reestablish mirroring, you must get
both drives online and then resynchronize the mirror, but you must
rebuild the set using a disk with the same partition style—either MBR or
GPT. The corrective action you take depends on the drive status.
Note
When mirroring boot volumes, Windows requires you to use the same partition style. With data volumes, you can mirror between MBR and GPT.
The Missing or Offline status usually happens if drives have been
disconnected or powered off. If the drives are part of an external storage
device, check the storage device to ensure that it is connected
properly and has power. Reconnecting the storage device or turning on
the power should make it so that the drives can be accessed. You then
must start Disk Management and rescan the missing drive by selecting
Rescan Disks on the Action menu. When Disk Management finishes, press
and hold or right-click the drive and choose Reactivate Volume. The
drive status should change to Regenerating and then to Healthy. If the
volume doesn’t return to the Healthy status, press and hold or
right-click the volume and then choose Resynchronize Mirror.
A status of Failed, Online (Errors), or Unreadable indicates I/O
problems with the drive. As before, try rescanning the drive, and then
try to reactivate the drive. The drive status should change to
Regenerating and then to Healthy. If the volume doesn’t return to the
Healthy status, press and hold or right-click the volume and then choose
Resynchronize Mirror.
If these actions don’t work, you must remove the failed mirror,
replace the bad drive, and then rebuild the mirror. To do this, follow
these steps:
-
Press and hold or right-click the failed volume, and then select Remove Mirror. -
You now must mirror the volume on an Unallocated area of free space
on a different disk. If you don’t have free space, you must create space
by shrinking a volume, deleting other volumes, or replacing the failed
drive. -
When you are ready to continue, press and hold or right-click the
remaining volume in the original mirror and then select Add Mirror. This
opens the Add Mirror dialog box. -
Use the Disks list to select a location for the mirror, and then tap
or click Add Mirror. Windows Server 2012 begins the mirror creation
process, and you’ll see a status of Resynching on both volumes.
6.10 Repairing a mirrored system volume
When you mirror a system volume, an entry that allows you to boot to
the secondary mirror is added to the system’s boot configuration data.
So, if a system fails to boot to the primary system volume, restart the
system, and select the Boot Mirror—Secondary Plex option for the
operating system you want to start. The system should start up normally.
After you successfully boot the system to the secondary drive, you can
schedule the maintenance necessary to rebuild the mirror if desired.
Rebuilding mirrored system volumes on MBR disks
To rebuild the mirror, you must complete the following steps:
-
Shut down the system, and replace the failed drive. Then restart the system using the secondary drive. -
In Disk Management, press and hold or right-click the remaining
volume in the mirrored set and choose Break Mirrored Volume. Tap or
click Yes at the prompts to confirm the action. -
Next, press and hold or right-click the volume again, and choose Add
Mirror. Use the Add Mirror dialog box to select the second disk to use
for the mirror, and then tap or click Add Mirror.
If you want the primary mirror to be on the drive you added or replaced, perform these additional steps:
-
Use Disk Management to break the mirrored set again. -
Make sure that the primary drive in the original mirror set has the
drive letter that was previously assigned to the complete mirror. If it
doesn’t, assign the appropriate drive letter. -
Press and hold or right-click the original system volume, select Add Mirror, and then re-create the mirror.
Rebuilding mirrored system volumes on GPT disks
For GPT disks, rebuilding mirrored
system volumes is a bit different. To rebuild the mirror, shut down the
system and replace the failed drive, and then restart the system using
the secondary drive. In Disk Management, press and hold or right-click
the remaining volume in the mirrored set and choose Break Mirrored
Volume. Tap or click Yes at the prompts to confirm the action.
6.11 Resolving problems with RAID-5 sets
Most problems with RAID-5 sets have to do with the intermittent or
permanent failure of a drive. If one of the drives in the set goes
offline or experiences temporary I/O problems, parity data cannot be
properly written to the set and, as a result, the set’s status will show
as Failed Redundancy and the failed volume’s status changes to Missing,
Offline, or Online (Errors).
You must get all drives in the RAID-5 set online. If the status of
the problem volume is Missing or Offline, make sure that the drive has
power and is connected properly. You then must start Disk Management and
rescan the missing drive by choosing Rescan Disks from the Action menu.
When Disk Management finishes, press and hold or right-click the drive
and choose Reactivate. The drive status should change to Regenerating
and then to Healthy. If the volume doesn’t return to the Healthy status,
press and hold or right-click the volume and then tap or click
Regenerate Parity.
A status of Failed, Online (Errors), or Unreadable indicates I/O
problems with the drive. As before, try rescanning the drive, and then
try to reactivate the drive. The drive status should change to
Regenerating and then to Healthy. If the volume doesn’t return to the
Healthy status, press and hold or right-click the volume and then tap or
click Regenerate Parity.
If one of the drives still won’t come back online, you must repair
the failed region of the RAID-5 set. Press and hold or right-click the
failed volume, and then select Remove Volume. You now must press and
hold or right-click an unallocated space on a separate dynamic disk with
the same partition style—either MBR or GPT—and choose Repair Volume.
This space must be at least as large as the region to repair, and it
can’t be on a drive that’s already being used by the RAID-5 set. If you
don’t have enough space, the Repair Volume option is unavailable and you
must free space by shrinking a volume, deleting other volumes, or
replacing the failed drive.
|