3ware HTML Bookshelf



3ware® CLI Guide > CLI Syntax Reference > Controller Object Commands > /cx add type=<RaidType> disk=<p:-p> [stripe=Stripe] [noscan] [group=<3|4|5|6|7|8|9|10|11|12/13|14|15|16>] [nocache] [autoverify] [noqpolicy] [ignoreECC] [name=string] [storsave=<protect|balance|perform>] [rapidrecovery=all|rebuild|disable] [v0=n|vol=a:b:c:d]

/cx add type=<RaidType> disk=<p:-p> [stripe=Stripe] [noscan] [group=<3|4|5|6|7|8|9|10|11|12/13|14|15|16>] [nocache] [autoverify] [noqpolicy] [ignoreECC] [name=string] [storsave=<protect|balance|perform>] [rapidrecovery=all|rebuild|disable] [v0=n|vol=a:b:c:d]
This command allows you to create a new unit on the specified controller. You specify type, disks, and optional stripe size. By default the host operating system will be informed of the new block device, write cache will be enabled, a storsave policy of protect will be set, and the drive queuing policy is enabled. In case of RAID 50, you can also specify the layout of the unit by specifying the number of disks per disk group with the group attribute.
 
Note: By default, write cache is enabled. However, if the controller does not have a BBU installed, a message will warn you that you could lose data in the event of a power failure.
Enabling write cache will improve write performance greatly, but you are at risk of losing data if a power failure occurs when data is still in the cache. You may want to obtain a UPS to safeguard against power loss.
/cx is the controller name, for example /c0, /c1, and so forth.
type=RaidType specifies the type of RAID unit to be created. Possible unit types include raid0, raid1, raid5, raid6 (9650SE and higher only), raid10, raid50, single, and spare.
Example: type=raid5
When a new unit is created, it is automatically assigned a unique serial number. In addition, users can assign the unit a name.
 
Note: The unit’s serial number cannot be changed.
The following table shows supported types and controller models.
 
9650SE and 9690SA
Yes (SATA only)

1
Models 9500S, 9550SX, and 9590SE

disk=p:-p consists of a list of ports (disks) to be used in the construction of the specified unit type. One or more ports can be specified. Multiple ports can be specified using a colon (:) or a dash (-) as port index separators. A dash indicates a range and can be mixed with colons. For example disk=0:1:2-5:9:12 indicates port 0, 1, 2 through 5 (inclusive), 9 and 12.
If you have a 9690SA controller, the syntax is the same even though you are technically addressing vports.
stripe=Stripe consists of the stripe size to be used. The following table illustrates the supported and applicable stripes on unit types and controller models. Stripe size units are in K (kilobytes). If no stripe size is specified, 64K is used by default, if applicable. If you need to change the stripe size after the unit is created, you can do so my migrating the unit.
9650SE and 9690SA

1
Models 9500S, 9550SX, and 9590SE

group=3|4|5|6|7|8|9|10|11|12/13|14|15|16 indicates the number of disks per group for a RAID 50 type. (This attribute can only be used when type=raid50.) Group=13-16 is only applicable to 9690SA.
Recall that a RAID 50 is a multi-tier array. At the bottom-most layer, N number of disks per group are used to form the RAID 5 layer. These RAID 5 arrays are then integrated into a RAID 0. This attribute allows you to specify the number of disks in the RAID 5 level. Valid values are 3 through 16. However, no more than 4 RAID 5 subunits are allowed in a RAID 50 unit.
Note that a sufficient number of disks are required for a given pattern or disk group. For example, given 6 disks, specifying 3 will create two RAID 5 arrays. With 12 disks, specifying 3 will create four RAID 5 arrays under the RAID 0 level. With only 6 disks a grouping of 6 is not allowed, as you would basically be creating a RAID 5.
The default RAID 50 grouping varies, based on number of disks. For 6 and 9 disks, default grouping is 3. For 8 disks, the default grouping is 4. For 10 disks, the default grouping is 5, and for 12 disks, the disks can be grouped into groups of 3, 4, or 6 drives (the group of 4 drives is set by default as it provides the best of net capacity and performance). For 15 disks, the disks can be grouped into 5 drives (3 drive groups would make 5 subunits, you can have a maximum of 4 subunits). For 16 disks, the disks can be grouped into groups of 4 or 8 drives.
Note that the indicated group number that is supported depends on the number of ports on the controller. group=16 is the maximum and it is available on the 9690SA.
noscan attribute instructs CLI not to notify the operating system of the creation of the new unit. By default CLI will inform the operating system. One application of this feature is to prevent the operating system from creating block special devices such as /dev/sdb and /dev/sdc as some implementations might create naming fragmentation and a moving target.
nocache attribute instructs CLI to disable the write cache on the newly created unit. Enabling write cache increases write performance at the cost of potential data loss in case of sudden power loss (unless a BBU or UPS is installed). By default the cache is enabled. To avoid the possibility of data loss in the event of a sudden power loss, it is recommended not to set nocache unless there is a BBU (battery backup unit) or UPS (uninterruptible power supply) installed.
autoverify attribute enables the autoverify attribute on the unit that is to be created. For more details on this feature, see /cx/ux set autoverify=on|off. This feature is not supported on model 7000/8000. For 9650SE and 9690SA controllers that support basic verify, autoverify will be set to ON by default for a new unit. For other 9000-series controllers that do not support basic verify, autoverify is set to OFF by default for a new unit.
noqpolicy attribute instructs CLI to disable the qpolicy (drive queuing for SATA drives only) on the newly created unit. The default is for the qpolicy to be on (in other words, noqpolicy is not specified). For a spare unit, drive queuing is not meaningful, so the noqpolicy cannot be set. During unit creation, specifying noqpolicy for a spare returns an error. (If the spare unit becomes a true unit, it will adopt the qpolicy of the “new” unit.) For more about drive queuing, see /cx/ux show qpolicy and /cx/ux set qpolicy=on|off.
ignoreECC attribute enables the ignoreECC/OverwriteECC attribute on the unit that is to be created. For more details on this feature, see /cx/ux set ignoreECC=on|off. The following table illustrates the supported Model-Unit Types. This table only applies to setting this feature at unit creation time. IgnoreECC only applies to redundant units. For the 7/8000 series, this setting is only applicable during rebuild; it is not applicable during creation.
 
9650SE
and
9690SA

1
Models 9500S, 9550SX, and 9590SE

name=string attribute allows you to name the new unit. (This feature is for 9000 series and above controllers.) The string can be up to 21 characters and cannot contain spaces. In order to use reserved characters (‘<‘, ‘>’, ‘!’, ‘&’, etc.) put double quotes (" ") around the name string. The name can be changed after the unit has been created. For more information, see /cx/ux set name=string and /cx/ux show name.
storsave=protect|balance|perform attribute allows user to set the storsave policy of the new unit. This feature is only for 9000 series SX/SE/SA controllers. For more information, see /cx/ux set storsave=protect|balance|perform [quiet].
rapidrecovery=all|rebuild|disable attribute specifies the Rapid RAID Recovery setting for the unit being created. Rapid Raid Recovery can speed up the rebuild process, and it can speed up initialize and verify tasks that may occur in response to an unclean system shutdown. Setting this option to all applies this policy to both these situations. Setting it to rebuild applies it only to rebuild tasks. If the policy is set to disable, then none of the tasks will be sped up.
 
Notes: Once the rapidrecovery policy has been disabled for a unit, it cannot be changed again. Disabling this policy is required if you want to move a unit to a controller that does not have the 9.5.1 firmware.
There is some system overhead from setting rapidrecovery to all. If you have a BBU, you can set rapid recovery to rebuild, as a BBU provides protection against data loss in the event of an unclean shutdown.
v0=n or vol=a:b:c:d may be used to divide the unit up into multiple volumes.
v0=n can be used if you only want two volumes, in which case v0=n is used to define the size of the first volume, and the second volume will use the remaining space. One way in which this can be useful is if you want to create a special volume to function as a boot volume, with a separate volume for data.
vol=a:b:c:d can be used to specify sizes for up to four volumes.
The value(s) should be positive integer(s) in units of gigabytes (GB), with a maximum of 32 TB. If you specify a size that exceeds the size of the unit, the volume will be left “uncarved.”
Both v0=n or vol=a:b:c:d work in conjunction with auto-carving, if that feature is enabled. When auto-carving is used, v0=n and vol=a:b:c:d are used to specify the size of the first few volumes, after which the auto-carve size is used for additional volumes. (For more about auto-carving, see /cx set autocarve=on|off and /cx set carvesize=[1024..32768].)
Example of RAID 5 unit created with first volume set to 10 GB:
//localhost> /c0 add type=raid5 disk=2-5 v0=10
 Creating new unit on Controller /c0 ... Done. The new unit is /c0/u0.
 Setting write cache=ON for the new unit ... Done.
 Setting default Command Queuing Policy for unit /c0/u0 to [on] ... Done.
After the unit creation, a subsequent show command for the unit shows the the volume size(s):
 //localhost> /c0/u0 show
 
 Unit    UnitType  Status  %RCmpl  %V/I/M  VPort  Stripe Size(GB)
 --------------------------------------------------------------
 u0       RAID-5    OK      -       -       -     64K    1117.56   
 u0-0     DISK      OK      -       -       p2    -      372.519   
 u0-1     DISK      OK      -       -       p3    -      372.519   
 u0-2     DISK      OK      -       -       p4    -      372.519   
 u0-3     DISK      OK      -       -       p5    -      372.519   
 u0/v0    Volume    -       -       -       -     -      10        
 u0/v1    Volume    -       -       -       -     -      1107.56
Example of RAID 0 unit created with volume sizes set to 2000, 500, 1024, and 700 GB: 
The example below combines auto-carving and vol=a:b:c:d. Notice that the last volume (u0/v5) is odd-sized (247.188 GB).
Volumes 0 through 3 are carved using the first four sizes as specified.  Volumes 4, 5, and 6 are the auto carved volumes (1024 GB each). Volume 6 is the remainder of the carve size.
 //localhost> /c2 add type=raid0 disk=0:1:2:4:5:6:
7 vol=2000:500:1024:700
Creating new unit on controller /c2 ...  Done. The new unit is /c2/u0.
Setting default Command Queuing Policy for unit /c2/u0 to [on] ... Done.
Setting write cache=ON for the new unit ... Done.
After the unit creation, a subsequent show command for the unit shows the volume sizes:
//localhost> /c2/u0 show
 
Unit     UnitType  Status         %RCmpl  %V/I/M  Port  Stripe  Size(GB)
------------------------------------------------------------------------
u0       RAID-0    OK             -       -       -     64K     6519.19   
u0-0     DISK      OK             -       -       p0    -       931.312   
u0-1     DISK      OK             -       -       p1    -       931.312   
u0-2     DISK      OK             -       -       p2    -       931.312   
u0-3     DISK      OK             -       -       p4    -       931.312   
u0-4     DISK      OK             -       -       p5    -       931.312   
u0-5     DISK      OK             -       -       p6    -       931.312   
u0-6     DISK      OK             -       -       p7    -       931.312   
u0/v0    Volume    -              -       -       -     -       2000      
u0/v1    Volume    -              -       -       -     -       500       
u0/v2    Volume    -              -       -       -     -       1024      
u0/v3    Volume    -              -       -       -     -       700       
u0/v4    Volume    -              -       -       -     -       1024     
u0/v5    Volume    -              -       -       -     -       1024      
u0/v6    Volume    -              -       -       -     -       247.188  
 

Copyright AMCC 2004-2008
To contact Support, log in at https://www.3ware.com

www.3ware.com
1-858-535-6517
1-800-840-6055