Parity:

Parity is additional information stored along with the data that enables the
controller to rebuild ("reconstruct") data after a drive failure. Only RAID 3 and 5 have
parity. RAID 1 uses mirroring, not parity, which allows the controller
to reconstruct data after a drive failure.

Parity Check/Repair:

Parity check scans the segments in a logical unit and checks the parity for
each segment. RAID 1 does not use parity, but parity check can still
be performed on a RAID 1 unit. In this case, parity check compares the data
on the mirrored drives.
Parity repair corrects any parity inconsistencies found during parity check.
However, parity repair corrects only parity errors. If the errors
were caused by corrupted data, the data is still corrupted.
You can set a time to run an automatic parity check/repair by using Options
* Auto Parity Settings in the Maintenance/Tuning application.
You can also initiate a manual check/ repair using the Recovery application.

RAID Level:
The RAID level determines how the controller reads and writes data and
parity on the drives The storage management software supports 4 RAID levels:

* RAID 0 - In RAID 0, data is striped across the drives in segments. There
is no parity data, so RAID 0 uses the full capacity of the drives. However, 
there is no redundancy, so if a single drive fails, all data on the
logical unit is lost.

* RAID 1 - In RAID 1, data is mirrored. Each data drive has a corresponding
mirrored drive with identical data. If one drive in the mirrored pairs fails,
the data from the other drive is used. In fact, a RAID 1 logical unit
can suffer multiple drive failures without losing data, as long as none
of the failed drives are mirrored pairs. Because RAID 1 mirrors data,
a RAID 1 logical unit has only half the capacity of the assigned drives for
example, if you create a four-drive RAID 1 logical unit with 18 MB drives,
the resulting data capacity is 2 x 18 = 36 MB)

* RAID 3 - In this version of the storage management software, RAID 3 is
identical to RAID 5.

* RAID 5 - In RAID 5, data and parity is striped across the drives in the
logical unit in segments. Because of this parity, if a single drive fails,
data can be recovered from the remaining good drives.
Two drive failures cause all data to be lost. A RAID 5 logical unit has the
capacity of all the drives in the logical unit less one, for
example, a five-drive RAID 5 logical unit with 18 MB drives has a
data capacity of 4 x 18 = 72mb.


Running Parity Check:

The RAID Modules use parity information in RAID 3 and 5 logical units to
enable data redundancy. When a single drive fails in a RAID 3 or 5
logical unit, the controller can reconstruct the data on the missing
drive by using the parity information stored on the drives. RAID 1
logical units do not use parity, but you can still run Parity
Check/Repair on RAID 1 logical units. In this case, the check
compares the data on the mirrored drives. RAID 0 logical units do not
have parity and therefore cannot be checked. A non-optimal RAID 1, 3,
or 5 logical unit (such as a degraded units) also cannot be
checked. When you run Parity Check/Repair, you can specify 
whether you want the controller to automatically fix any parity errors found.
If the error is found on a RAID 3 or 5 logical unit, the controller changes
the parity information so that the parity is correct. If an error is found
on a RAID 1 logical unit, the controller changes the data stored on
the mirrored drive so that it is the same as that on the other drive.
Note that repair only guarantees that the parity information, not the data,
is correct. The data in the logical units may be corrupted even though
the parity information is correct. If you get a parity error, you may need
to locate the file containing the error and restore the file from back-up.
To ensure that a redundant logical unit will be able to recover after a
failure, you should set your RAID Modules to perform 
Automatic Parity Check/Repair on a regular basis. You can set it to
run either daily or once a week. At times, you may need to run
Parity Check/Repair manually as part of the recovery process.