User manual

RAID 10
(striping over the member RAID 1). Allows two hard disks to fail or being unplugged, but the hard disks must
in different array.
RAID 30
(striping over the member RAID 3). Allows two hard disks to fail or being unplugged, but the hard disks must
in different array.
RAID 50
(striping over the member RAID 5). Allows two hard disks to fail or being unplugged, but the hard disks must
in different array.
RAID 60
(striping over the member RAID 6). Allows four hard disks to fail or being unplugged, but each two hard disks
must in different array.
JBOD
Just a Bunch Of Disks. No data protection. Data is damaged or inaccessible if any hard disk fails or is
unplugged.
4.2 VG Migration and Expansion
To migrate the RAID level, please follow the steps below. If the VG is migrated to the original VG within
the same RAID level, then it is regarded as expansion.
1. Click “/ Volume confi g / Volume group”
2. Select the VG to be migrated and then click “ ” button in the RAID fi eld
3. Change the RAID level by clicking the “ ” drop-down list box. If the hard disk capacity
is not enough to support the new RAID level, there will be a dialog window popped-up, then click “
” to increase hard disks and then click “ ” to go back to the confi gure
page. When migrating to lower RAID level, such as the migration from RAID 6 to 0, the system will
evaluate if this operation is safe and display the “Sure to migrate to a lower protection array?” to warn you.
4. Double check the confi gures. If there is no problem, click“ “.
5. Finally a confi rmation page will show detailed RAID data. If there is no problem, click “ ”
to start the migration. The system will also pop up the warning message. “Warning: power lost during
migration may cause damage of data!” to warn user that if power lost is occurred during migration, the hard
disk will be in high risk.
6. In the “status 3” fi eld, you can view the migration status and the letter “M” will be displayed. In “/
Volume confi g/User data volume”, a “M” and complete percentage “R%” will be displayed in “Status 3”
Figure 4.2.1
Chapter 4 Advanced Operation
4.1 Rebuild
In a VG that is set as protected by RAID level (e.g.: RAID 3, RAID 5, or RAID 6), if one physical disk has
been failed, unplugged or removed, then, the VG status is changed. The system will search for available
space to rebuild the data to ensure data integrity. It will use the dedicated spare disk as rebuild disk fi rst,
then the global spare disk.
STARDOM SCSI Series Products support the Auto-Rebuild function. When the RAID level allows disk
failures and the VG is protected, such as RAID 3, RAID 5, RAID 6 and etc, STARDOM SCSI Series
Products start Auto-Rebuild according to the scenario below:
Take RAID 6 for example:
1. When there is no global spare disk or dedicated spare disk on the system, STARDOM SCSI Series
Products will be in degraded mode and wait until (A) one hard disk is assigned as spare hard disk, or
(B) the failed hard disk is replaced with new clean hard disk, then Auto-Rebuild starts. The new disk
will be a spare disk to the original VG automatically.
a. If the new added hard disk is not clean (with unknown VG information), it would be marked as RS
(reserved) and the system will not start the Auto-Rebuild.
b.
If this disk is belonging to existing VG, it would be FR (Free) disk and the system will start Auto-Rebuild.
2. When there is enough global spare disk(s) or dedicated spare disk(s), STARDOM SCSI Series Products
will start Auto-Rebuild immediately. In RAID 6 level, if there is another disk failure happening during the
rebuilding process, STARDOM SCSI Series Products will complete the above Auto-Rebuild scenario as
well. The Auto-Rebuild feature only works at runtime to prevent a confl ict with the “Roaming” function.
In degraded mode, the status of VG is “DG”.
When rebuilding, the status of PD/VG/UDV will display “R”, and “R%” in UDV will display the ratio in
percentage. After rebuilding is complete, “R” and “DG” will disappear. And the VG will become a complete
one.
Remind : The list will not appear if there is no VG or only the VG under RAID 0 or JBOD mode. It is because the users cannot confi gure
dedicated space under these RAID levels.
Sometimes rebuild is called recover; these two have the same meaning. The following table describes the
relationship between RAID levels and rebuilding.
RAID 0 (striping). No data protection. Data is damaged or inaccessible if any hard disk fails or is unplugged.
RAID 1
(mirroring). Allows one hard disk to fail or being unplugged. One new hard disk is required to insert to the
system to complete the rebuilding.
N-way mirror Extension of RAID 1 level. It has N copies of the disks and allows N-1 hard disks to fail or being unplugged.
RAID 3 (striping with parity). Allows one hard disk to fail or being unplugged.
RAID 5 (striping with interspersed parity). Allows one hard disk to fail or being unplugged.
RAID 6
(2-dimensional parity protection). RAID 6 allows two hard disks to fail or being unplugged. If two hard disks
need to be rebuilt at the same time, it will rebuild the fi rst one and then the other one.
RAID 0+1
(mirroring of the member RAID 0). Allows two hard disks to fail or being unplugged, but the hard disks must
in the same array.
58
Advanced Operation Advanced Operation 59