![]() |
The Array View |
![]() |
![]() |
![]() |
RAID 0 | RAID 1 | RAID 5 | RAID 10 | FAST W |
The Array View lists the attributes of the selected array. The array members are shown with a summary of their attributes.
This help is divided into sections :
Data Striping Data Mirroring Data Striping with Parity Mirrored Data Striping Resources which use the Fast Write Cache
( RAID 0, 1, 5 & 10 )
( Select the RAID type below )
RAID 0
3 to 16 member disks.
RAID 1
2 member disks.
RAID 5
3 to 16 member disks.
RAID 10
4 to 16 member disks, in steps of 2.
Fast Write
(RAID 10 Fast Write Array Shown)
Common Attributes: |
![]() |
Usage: The usage
of an SSA resource dictates how you can use the resource. All possible
usage classes are explained in the Resource List
help. Status: The array status depends on the
type of array.
An online array is currently active and can be accessed by
the selected adapter. An offline array has more than one member disk missing or it
is in an illegal network.
A temporary array name may be reported for an offline array if
the array configuration cannot be determined.
Online
Offline
Note:
All RAID types, except RAID 0, further define the online status as one of the following :
Non-Degraded | All members of the array are present and all parity is up to date. |
Exposed | One member has been lost, but no write operation has been received since the member was lost. Data can be read from the other member. The missing member can be brought back and the array can then be returned to its non-degraded state. |
Degraded | One member has been lost, and a write operation was received after the member was lost. The missing member is rejected from the array. If the disk returns, it will be listed under the 'Rejected' resource list. |
Rebuilding | Data is being reconstructed on the array. This occurs when an array is created and when a member is returned to, or exchanged into the array. |
RAID 10 can also have a status of :
Mixed | Different components in a RAID 10 array can be in different states. For example, one mirrored pair might be rebuilding while, in a different pair, one member disk drive is missing and the remaining members are in the Online - degraded state. A priority of array states is used when the state of an array is reported :
|
Number of Components: The number of physical disk components, or members, in the array.
Size: The capacity of the array, in giga bytes.
Block Size: The physical block size of the records in the array.
This must be 512. Support of 524 is for a specialised application.
Network ID: This identifies which SSA loop is being used.
Note: | Array members must all be on the same SSA loop. |
Additional 'Hot-spare' Attributes: (Common to RAID 1, 5 & 10) |
![]() |
All array types, except RAID 0, can make use of Hot-spare resources to further enhance the availability of the SSA subsystem.
Hot-spare Enabled:
This should be set to 'On' to enable the use of Hot-spares.
If an this is set to 'Off', the array cannot use any available
Hot-spare disks if required. The array will move to a
degraded state.
Default: On Supports: On, Off
Note:
Hot-spare Exact Size:
Default: | Off |
Supports: | On, Off |
When Hot-spare exact is set to 'On', only hot spares of exactly the same size as the failing disk will be used. When Hot-spare exact is 'Off', the Hot-spare used may have a greater capacity than the failing disk it is replacing.
Additional 'Hot-spare Pool' Attribute: (IBM Advanced SerialRAID/X Adapter Only) |
![]() |
With the IBM Advanced SerialRAID Adapter, all array types, except RAID 0, can make use of Hot-spare Pool resources. These pools further enhance the availability of the SSA subsystem.
Hot-spare Splits:
When Hot-spare Splits is set to 'On', and only one half of the RAID-10,
or RAID-1 array is available for use; Hot-spare
disks will still be used, if required, from the chosen or available pools.
When Hot-spare Splits is set to 'Off', hot-spare disks will not be used if
only one half of the RAID-10, or RAID-1 array is available. Hot-spare Preferred:
When Hot-spare preferred is set to 'On', only a hot spare from the
preferred Hot-spare pool will be used to replace a failing disk. When
preferred is 'Off', a suitable Hot-spare from a Hot-spare pool
other than the preferred pool may be used to replace a failing disk if
the preferred Hot-spare pool is empty.
If Hot-spare preferred is 'On', Hot-spare disks in the global
pool 0 will not be used.
Default: Off Supports: On, Off
Default: Off Supports: On, Off
Note:
Additional 'System (PC)' Attribute: (System 'PC' Resources Only) |
![]() |
Access: When running in Cluster Mode
arrays and disks can be made private to one adapter. That is, only one
adapter can access the resource. Access can be set only when the resource
is in the usage class 'System (PC)'.
Cluster Mode Only. The access can be set to three types:
Resource Number: The Resource Number is assigned when you 'Attach' the resource and make it available to the operating system. Resource numbering controls the order in which SSA resources are presented to the operating system , lowest first. Only 'System (PC)' resources have resource numbers.
![]() |
Additional 'RAID 0' Attributes: |
![]() |
Note: | RAID 0 arrays can consist of three to sixteen components. The larger the number of components, the more the data is spread across disks and performance is increased. The disadvantage of a large number of components is that as RAID-0 provides no redundancy and data is striped across all the member disks, the loss of a single disk results in the loss of more data. |
RAID 0 attributes depend on which adapter created the array.
Select your adapter from this list to view the relevant attributes.
Additional 'RAID 0' Attribute: when created with the RAID 0 Arrays created with this adapter have
the following attribute. This is in addition to the
common attributes outlined above. The Block Size can only be 512
with this adapter.
Strip Size:
This defines the number of contiguous blocks held on
a disk before switching to the next disk. Larger values may result in worse
performance for long transfers as the data may be striped across fewer disks.
Lower values can result in worse performance for writes. Additional 'RAID 0' Attribute: when created with the RAID 0 Arrays created with either of these adapters have
the following attribute. This is in addition to the
Common attributes outlined above.
Strip Size:
This defines
the number of contiguous blocks held on a disk before switching to the next
disk. Larger values may result in worse performance for long transfers as
the data may be striped across fewer disks. Lower values can result in worse
performance for writes. Additional 'RAID 1' Attributes: RAID 1 arrays can only consist of two components.
IBM SSA RAID Adapter (LL03 and LL11)
Note:
Default: 32 (Blocks) Supports: Min 8, Max 8192 in steps of 1
IBM Advanced SerialRAID/X Adapter
Default: 32 (Blocks) Supports: Min 8, Max 512 in steps of 8
Note:
The attributes of your RAID 1 arrays will depend on
the adapter that created the array. Select your adapter from this
list to view the RAID 1 attributes your adapter will present.
Additional 'RAID 1' Attributes: when created with the RAID 1 Arrays created with either of these adapters will have
the following attributes. These are in addition to the
Common and Hot-spare
attributes outlined above.
Initial Rebuild:
When Initial Rebuild is set to 'On', the array will initially be
rebuilding and will copy the data from the first component to the second.
If this bit set to 'Off', the array starts life with its contents undefined,
and will immediately become non-degraded. This attribute can only be used at creation time, and cannot be
subsequently be changed. This attribute also insists that the first component
has less than or an equal number of data blocks as the second. With the Initial Rebuild bit set to 'On' and user data has yet to be written to the array,
the same data will be returned by whichever member disk is read. If the attribute
is 'Off', it is not guaranteed to always return the same data for the same
block before that block contains user data. It is recommended that the Initial Rebuild attribute is set to 'On', to
avoid possible problems if applications depend on
the same data returned before data is written. Note that this will incur
some performance overhead while the array is being
rebuilt.
IBM SSA RAID Adapter or the
IBM SSA RAID Cluster Adapter
Default: Off Supports: On, Off
Note:
Data Scrub Enabled:
Default: | Off |
Supports: | On, Off |
When Data Scrub Enable is set to 'On', the array will, when it is non-degraded, verify in the background that all the data for the array on the member disks can be read without error. If an error is detected during this verification, the failing block on the disk is repaired by rewriting the data from the mirrored array member disk.
Data Scrub Rate:
Default: | 24 (hours) |
Supports: | 6 to 48 (hours) in steps of 1 |
The Data Scrub rate sets the time in hours for the period of a single pass of data scrubbing the array. The default time is 24 hours. The shorter time incurs more interference with processing requests but detects latent failures earlier.
Rebuild Priority:
Default: | 50% |
Supports: | 1% to 100% |
A higher value will result in faster array rebuilds, but may reduce the performance of applications during the rebuild period. A lower value will result in slower array rebuilds, but may increase the performance of applications during the rebuild period.
Split Confirm:
IBM SSA RAID Cluster Adapter Only
Default: | Off |
Supports: | On, Off |
When Split Confirm is set to 'On', an array with only one member disk, and previously both member disks were not attached to this adapter, is not brought online and an SSA event is generated. This provides protection from a member disk being moved to another system and then being used whilst the other disk of a mirrored pair is still being used on the original system.
![]() |
Additional 'RAID 1' Attributes: when created with the |
RAID 1 Arrays created with this adapter will have the following attributes. These are in addition to the Common and Hot-spare attributes outlined above. For further information see the help which explains Multiple Domains and Component Ordering.
Split Array Resolution:
This attribute controls how the subsystem will react in the event
of a split between the domains. When set to 'Primary', the normal setting,
ONLY the Primary Domain will continue to operate when a split occurs. When
set to 'Secondary' ONLY the Secondary Domain will continue when a split
occurs. In use, if a split occurs due to the loss of the Primary Domain, all
operation will stop. Changing the setting of this attribute to 'Secondary'
will allow the secondary domain to continue to operate. This should only be
done when you are sure that the other using system has lost
access to and is not using the Primary Domain. If the fault has
occurred in the SSA loop, for example a cable fault, and the other
using system is still using the Primary Domain, both systems will be
able to access the two domains independently which will result in
unsynchronised data. Cluster Mode Only Initial Rebuild:
When Initial Rebuild is set to 'On', the array will initially be
rebuilding and will copy the data from the first component to the second.
If this bit set to 'Off', the array starts life with its contents undefined,
and will immediately become non-degraded. This attribute can only be used at creation time, and cannot be
subsequently be changed. This attribute also insists that the first component
has less than or an equal number of data blocks as the second. With the Initial Rebuild bit set to 'On' and user data has yet to be written to the array,
the same data will be returned by whichever member disk is read. If the attribute
is 'Off', it is not guaranteed to always return the same data for the same
block before that block contains user data. It is recommended that the Initial Rebuild attribute is set to 'On', to
avoid possible problems if applications depend on
the same data returned before data is written. Note that this will incur
some performance overhead while the array is being
rebuilt.
Default: Primary Supports: Primary, Secondary
Default: Off Supports: On, Off
Note:
Rebuild Priority:
Default: | 50% |
Supports: | 1% to 100% |
A higher value will result in faster array rebuilds, but may reduce the performance of applications during the rebuild period. A lower value will result in slower array rebuilds, but may increase the performance of applications during the rebuild period.
Invalid Data Strips:
The number of data strips which are invalid.
Unbuilt Data Strips:
The number of data strips remaining to be built.
![]() |
Additional 'RAID 5' Attributes: |
![]() |
Note: | RAID 5 arrays can consist of three to sixteen components. The larger the number of components, the worse the performance may be for long transfers. However, the larger the number of components, the greater the reduction in the overhead of the parity function. Therefore the available capacity is maximised by using a larger number of components. |
RAID 5 attributes depend on the adapter that created the array. Select your adapter from this list to view the relevant attributes.
![]() |
Additional 'RAID 5' Attributes: when created with the |
RAID 5 Arrays created with this adapter will have the following attributes. These are in addition to the common and the 'Hot-spare' attributes outlined above.
Strip Size:
This defines the number of contiguous blocks held on a disk before
switching to the next disk. Larger values may result in worse performance
for long transfers as the data may be striped across fewer disks. Lower
values can result in worse performance for writes but incur a longer
rebuild time when a failed component is replaced.. Stretch Size:
This is the number of strips before the disk that holds the
parity data is changed. The default value is 4 and there is nothing to be
gained by using any other value. Rebuild Priority:
A higher value will result in faster array rebuilds, but may
reduce the performance of applications during the rebuild period.
A lower value will result in slower array rebuilds, but may increase
the performance of applications during the rebuild period. Invalid Data Strips: The number of data strips which are invalid. Unbuilt Data Strips:
The number of data strips remaining to be built. Unbuilt Parity Strips: The number of parity strips remaining to be built.
Default: 128 Supports: 128 Only
Default: 4 Supports: 1 to 9999 in steps of 1.
Default: 50% Supports: 1% to 100%
![]() |
Additional 'RAID 5' Attributes: when created with the |
RAID 5 Arrays created with either of these adapters will have the following attributes. These are in addition to the common and 'Hot-spare' attributes outlined above.
Strip Size:
This defines the number of contiguous blocks held on a disk
before switching to the next disk. Larger values may result in worse
performance for long transfers. Lower values may result in worse performance
for writes but incur a longer rebuild time when a failed component is
replaced. Stretch Size:
This is the number of strips before the disk that holds the
parity data is changed. The default value is 4 and there is nothing to be
gained by using any other value. Rebuild Priority:
A higher value will result in faster array rebuilds, but may
reduce the performance of applications during the rebuild period.
A lower value will result in slower array rebuilds, but may increase
the performance of applications during the rebuild period. Invalid Data Strips: The number of data strips which are invalid. Unbuilt Data Strips:
The number of data strips remaining to be built. Unbuilt Parity Strips: The number of parity strips remaining to be built.
Default: 128 Supports: 32, 64, 128
Default: 4 Supports: 1 to 9999 in steps of 1.
Default: 50% Supports: 1% to 100%
![]() |
Additional 'RAID 10' Attributes: when created with the |
![]() |
RAID 10 Arrays are only supported by the IBM Advanced SerialRAID/X Adapter. Such arrays have the following attributes in addition to the Common and the 'Hot-spare' attributes outlined above. For further information see the help which explains Multiple Domains and Component Ordering.
Note: | RAID 10 arrays can consist of four to sixteen components. The larger the number of components, the more the data is spread across disks and performance is increased. With a finite number of disks available for RAID 10 arrays, best performance is achieved with an equal number of RAID 10 arrays on each SSA loop. |
Strip Size:
This defines the number of contiguous blocks held on a disk before
switching to the next disk. Larger values may result in worse performance
for long transfers as the data may be striped across fewer disks. Lower
values can result in worse performance for writes but incur a longer
rebuild time when a failed component is replaced. Split Array Resolution:
This attribute controls how the subsystem will react in the event
of a split between the domains. When set to 'Primary', the normal setting,
ONLY the Primary Domain will continue to operate when a split occurs. When
set to 'Secondary' ONLY the Secondary Domain will continue when a split
occurs. In use, if a split occurs due to the loss of the Primary Domain, all
operation will stop. Changing the setting of this attribute to 'Secondary'
will allow the secondary domain to continue to operate. This should only be
done when you are sure that the other using system has lost
access to and is not using the Primary Domain. If the fault has
occurred in the SSA loop, for example a cable fault, and the other
using system is still using the Primary Domain, both systems will be
able to access the two domains independently which will result in
unsynchronised data. Cluster Mode Only Initial Rebuild:
When Initial Rebuild is set to 'On', the array will initially be
rebuilding and will copy the data from the first component to the second.
If this bit set to 'Off', the array starts life with its contents undefined,
and will immediately become non-degraded. This attribute can only be used at creation time, and cannot be
subsequently be changed. This attribute also insists that the first component
has less than or an equal number of data blocks as the second. With the Initial Rebuild bit set to 'On' and user data has yet to be written to the array,
the same data will be returned by whichever member disk is read. If the attribute
is 'Off', it is not guaranteed to always return the same data for the same
block before that block contains user data. It is recommended that the Initial Rebuild attribute is set to 'On', to
avoid possible problems if applications depend on
the same data returned before data is written. Note that this will incur
some performance overhead while the array is being
rebuilt.
Default: 32 Supports: 32, 64, 128
Default: Primary Supports: Primary, Secondary
Default: Off Supports: On, Off
Note:
Rebuild Priority:
Default: | 50% |
Supports: | 1% to 100% |
A higher value will result in faster array rebuilds, but may reduce the performance of applications during the rebuild period. A lower value will result in slower array rebuilds, but may increase the performance of applications during the rebuild period.
Invalid Data Strips:
The number of data strips which are invalid.
Unbuilt Data Strips:
The number of data strips remaining to be built.
![]() |
Additional 'Fast Write' Attributes: when created with the |
![]() |
Fast Write resources have the following Fast Write attributes in addition to the Common array, or Common disk attributes.
Minimum LBA: The Minimum Logical Block Address (LBA) is the address of the lowest
block of the array or disk that should be captured in the fast write cache.
Maximum LBA: The Maximum Logical Block Address (LBA) is the address of the highest
block of the array or disk that should be captured in the fast write cache.
Maximum Write Length Cached: Data will be held in the fast write cache for transfer length up to
the Maximum Write Length Cached. Transfers of lengths longer than this are
considered to be applications where data is being written to be saved on
disk without needing to be read back soon afterwards, so it is better not
to use the cache for these applications as there is little benefit from
reduced response time for future reads and it is better to not reduce the
available bandwidth by using the fast write cache.
Allow Data Delete:
When set to 'Off', this prevents a disk that is configured for fast write
being deleted if any data for that disk is held in the fast write cache
and that data cannot be destaged to the disk. When set to 'On', deletion of
disks is allowed even if data for that disk is in the cache and cannot be
destaged.Default: Off Supports: On, Off
Snoop Data into Battery Backed SRAM:
Default: | On |
Supports: | On, Off |
This should never be changed from its default value of 'on'. If it is set to 'off', fast write data is not held in the non-volatile cache and can be lost if power fails.
Cache Full Stride Writes:
Default: | On |
Supports: | On, Off |
When set to 'Off', data is not cached in the fast write cache if the length of the write is equal to or greater than the size of the stride. This results is improved performance of the writing and typically data written with full stride writes are not read back again soon afterwards, so there is little value in caching the data.
Bypass cache if 1-way:
Default: | Off |
Supports: | On, Off |
When set to 'Off', caching is still used when one adapter is missing in a cluster configuration. If it is required to maintain protection of data and avoid possible loss due to a second failure when one adapter is missing, this should be set to 'On'. When set to 'On', caching is disabled, so the performance benefit of fast write cache is not realised.
Actions:
Select any serial number to display the relevant
Disk View. This action updates the current page. This action starts the flashing indicator light on all of
the array components.
A message will be displayed in the Information Area
indicating the resource is flashing. This message
will appear only once. You must remember to Flash Off the
resource, or the light will continue to flash. To clear all indicators,
use the Flash All Off button on the
Adapter View to Flash Off the indicator
lights on all disks attached to the selected adapter. This action stops the indicator light flashing on all of
the array components. This action changes the use of an array from
'Free' to
'System (PC)'.
The array will then be available to the operating system. This action
links to the Attach page, where a resource
number is chosen. This number dictates the order in which resources
are presented to the operating system. 'Free' Resources Only. This action changes the use of an array from 'System (PC)' to
Free. The resource will no longer be available to the operating
system. 'System (PC)' Resources Only.
This action destroys the array and returns its component disks to
the 'Free' usage class. This action will take you to the Create Array
page. The default attributes will be assigned to the same values as the
current array. This action will take you to the Modify Array
page where you can modify the relevant array's attributes. This action will take you to the
Create Fast Write Array page. When complete,
the array will use the Fast Write Cache for data writes. 'Free' Resources, and
Fast Write Adapters Only. This action will take you to the
Modify Fast Write Array page where you can
modify some of the current array attributes. 'Free' Resources, and
Fast Write Adapters Only. This action will stop the array using the Fast Write Cache Feature
for data writes. 'Free' Resources, and
Fast Write Adapters Only. This action allows you to assign or modify the Hot-spare P that ALL of the array components will be protected by. A JavaScript text entry pop-up will be displayed. Enter the new Pool number. IBM Advanced SerialRAID/X Adapter Only
Serial Number
Refresh
Flash On
Note:
Flash Off
Attach
Detach
Delete
Create
Modify Attributes
Create FW
Modify FW
Delete FW
Assign Hot-spare pool
Related Information