Back to Contents Page
Overview
Partitioning array memory
Creating LUNs
Enabling or disabling array caching
Changing array caching parameters
Upgrading an array for caching
Updating Licensed
Internal Code (LIC) and PROM code
To configure an array initially, you perform the following tasks:
- Partition array memory if you will use read or write caching or create RAID 3 LUNs (see section
below)
- Create Luns (see "Creating LUNs")
After initial configuration, you may need to perform one or more of the following tasks:
 |
NOTE:
This chapter does not describe how to enable or
disable mixed mode or RAID3 write buffering because
- The SPs are always in mixed mode
- The design of the SPs make RAID3 write buffering unnecessary, and most importantly, enabling
RAID3 buffering may cause data loss when power fails or an array failure occurs.
|
To use read or write caching or bind RAID 3 LUNs, you must have the required hardware and specify
the array memory partitions. This section lists the hardware required for caching and RAID 3 LUNs,
describes how the SP memory architecture affects memory partitioning, and describes how to partition
memory.
An array supports read caching if it has an SP with at least 128 Mbytes of memory.
An array supports write caching if it has the following hardware:
- Two SPs.
- Two power supplies and two LCCs in the DPE and each DAE.
- Disk modules in slots 0 through 8
- Standby power supply (SPS) with a fully charged battery
Hardware requirements for RAID 3 LUNs
All arrays support RAID 3 LUNs. For maximum performance, we recommend that you do not bind
other types of LUNs with RAID 3 LUNs in the same array.
Before you partition array memory, you need a basic understanding of the SP memory architecture.
All the memory partitions reside in SP memory, as shown in the figure that follows.
Figure 6-1. SP memory architecture

The DIMMs on the SP make up the SP memory, so the size of the SP memory varies with the size and
number of the DIMMs.
The LIC system partition stores the Licensed Internal Code (LIC), and its size is fixed for a specific
LIC revision. A newer LIC revision may require more SP memory. The write cache partition is always the
same size on both SPs. If you resize it on one SP, it is automatically resized on the other SP.
The read cache, write cache and RAID 3 partitions contain the read cache, write cache, or RAID 3 data
and the partition's cache page control information. The size of the cache page control information varies
with the cache page size and the partition size. A small cache page size results in more cache pages,
and thus larger page tables and more space required for them in the partition. Conversely, a large cache
page size results in less cache pages, and therefore, smaller page tables and less space required for
them in the partition. You can set the cache page size using the following menu option in the Array
Configuration window: Array � Set Page Size.
The user free partition is the SP memory that is not allocated to the LIC system, read cache,
write cache, or RAID 3 partitions. When you decrease the size of a partition, the de-allocated memory
returns to the user free partition.
You can increase the size of the read cache, write cache, or RAID 3 partitions by the size of the
user free partition. If you do not have enough user free memory for the partition size you want, you
must decrease the size of one or more of the other partitions until you have enough user free memory.
 |
NOTE:
Changing the size of the RAID 3 memory partition
will reboot the SP. |
You partition memory for each SP using the Memory Partition window. The memory partitions are
read cache, write cache, RAID 3, extended, LIC system, and user free. The read cache, write cache, and RAID
3 partitions have a default size of 0. The extended partition always has a size of 0, and you cannot
change it. The LIC system partition size is the amount of memory required for the SPs, and its size
may change when you change the size of other partitions, as described in
"How SP memory
architecture affects memory partitioning".
To specify memory partitions
- Display the Array Configuration window
(see "To display the
Array Configuration window").
- In the Array Configuration window, disable array write and read caching for the array's SPs as
follows:
To disable array write caching for
both SPs
Either click the Write Cache Disable button on the array toolbar or select the menu option Array
� Write Cache State �
Disable.
To disable array read caching for SP A
If the array has an SP A, then on the array toolbar, either click the SP A Disable Read Cache Button
or select the menu option Array � Read Cache State �
SP A � Disable.
To disable array read caching for SP B
If the array has an SP B, then either click the SP B Disable Read Cache button on the array toolbar
or select the menu option Array � Read Cache State �
SP B � Disable.
- In the Array Configuration window, follow these steps to determine whether array
write and read caching is disabled:
- On the array toolbar, click the SP Information button for one SP in the array.
The SP Information window for the SP opens, similar to the sample that follows.
Figure 6-2. Sample SP Information window with configuration information

- In the SP Information window, click the Cache button.
SP cache information in displayed in the SP Information window, similar to that in the following
sample for SP A.
Figure 6-3. Sample SP Information window with cache information

- Check the Read Cache State and the Write Cache State entries.
If these entries are Disabled, the caches are disabled. It may take the array a while to disable
write caching if the SPs need to write data in the write cache to disks. As a result, the Write
Cache State entry in the SP Information may be Enabled for a while. You may want to poll the array
every few seconds to make sure you have the latest status. You can poll the array from the Array
Configuration window using either the Poll button on the array toolbar or the menu option Array
� Poll.
- Click the Close button.
- On the array toolbar, click the SP Information button for the other SP in the array.
The SP Information window for the SP opens.
- In the SP Information window, click the Cache button.
- Check the Read Cache State entry.
If the entry is Disabled, the caches are disabled.
- In the Array Configuration window, either click the Partition Memory button or select
the
menu option Array � Partition Memory.
A Memory Partition window for the array opens, similar to the following sample.
Figure 6-4. Memory Partition window

The meaning of the fields in the Memory Partition window are:
SP Total - Total amount of RAM memory on the SP, which consists of the SP memory
modules (DIMMs).
Read Cache - Amount of SP total memory allocated to the read cache.
Write Cache - Amount of SP total memory allocated to the write cache.
RAID 3 - Amount of SP total memory allocated to the RAID 3 memory partition.
Extended - Always 0. This partition is not currently supported.
LIC System - Amount of SP total memory required for the LIC, which is fixed for a specific
revision of LIC.
User Free - Amount of the total memory not already allocated to the read cache, write cache,
RAID 3, and LIC system partitions.
- If you want a different SP A read cache partition size, either enter the size in the SP
A Read Cache field or drag the slider to the right or left until the desired number of Mbytes (MB)
appears in the field.
 |
NOTE:
You can click the right side of the point on
the slider
to increase the number by one and click the left side to decrease the number by one. |
When you enter the size or as you move the slider, the pie chart changes to show the portion of
memory
allocated to the SP A read cache partition. The amount of memory you select for SP A is independent of
the amount you select for SP B.
 |
NOTE:
Array read caching for SP A is disabled if the
SP's read
cache partition is 0 MB, and it stays disabled until you allocate memory to the partition and enable
read caching for SP A as described in "To
enable
array read caching for SP A". |
Figure 6-5. Changing SP A read cache partition size

- If you want a different size SP B read cache partition, either enter the size in the SP
B Read Cache field, or drag the slider to the right or left until the desired number of Mbytes (MB)
appears in the field.
 |
NOTE:
You can click the right side of the point on
the slider
to increase the number by one and click the left side to decrease the number by one. |
When you enter the size or as you move the slider, the pie chart changes to show the portion of
memory
allocated to the SP B read cache partition. The amount of memory you select for SP B is independent of
the amount you select for the SP A.
 |
NOTE:
Array read caching for SP B is disabled if the
SP's
read cache partition is 0 MB, and it stays disabled until you allocate memory to the partition and
enable read caching for SP B as described in "To
enable array read caching for SP B". |
Figure 6-6. Changing SP B read cache partition size

- If you want a different size for the write cache partition, either enter the size in
the SP A write cache field, or drag the slider to the right or left until the desired number of
MB (Mbytes) appears in the field.
 |
NOTE:
You can click the right side of the point on
the slider
to increase the number by one and click the left side to decrease the number by one. |
The write cache partition on both SPs must be the same size. As a result, when you enter the size
or move the slider for the write cache partition on one SP, the slider for the write cache partition
for the other SP also moves. Both pie charts change to show the portion of memory allocated to the
write cache partition on both SPs.
 |
NOTE:
Array write caching is disabled if the write
cache partition
is 0 MB, and it stays disabled until you allocate memory to the partition, enable array write caching, as
described in "To enable array write
caching", and
the array has all the hardware components required for write caching. |
Figure 6-7. Changing write cache partition size

 |
CAUTION:
If you change the RAID 3 partition
size, Data
Supervisor reboots the array. Rebooting restarts the SPs in the array, which terminates all
outstanding I/O to the array. Do not complete this procedure for an array on a Windows NT server.
During reboot, LUNs are inaccessible, and you may receive error messages if the array is polled.
You must wait until the reboot is complete before changing any other settings.
If you set the RAID 3 partition to 0, any existing RAID 3 LUNs and new RAID 3 LUNs you create
become
unowned and you lose access to them. |
- If you want a different size for the RAID 3 partition, either enter the size in the RAID 3 field
for one SP, or drag the slider to the right or left until the desired number of Mbytes (MB) appears
in the field.
 |
NOTE:
You can click the right side of the point on
the slider
to increase the number by one and click the left side to decrease the number by one. |
We recommend a RAID 3 partition size of 16 Mbytes. The RAID 3 partition on both SPs must be the
same size. As a result, when you enter the size or move the slider for the RAID 3 partition on one
SP, the slider for the RAID 3 partition on the other SP also moves. Both pie charts change to show
the portion of memory allocated to the RAID 3 partition on both SPs.
- Click the OK button to change the memory partitions to the sizes you
specified.
A confirmation window opens.
- Click the Yes button in the window to confirm the new sizes.
- If you want to use read or write caching, and you have allocated memory to the Read and/or Write
Cache
partition, enable array read caching for SP A or SP B or array write caching as follows:
To enable array read caching for SP
A
If the array has an SP A, then on the array toolbar, either click the SP A Enable Read Cache button
or select the menu option Array � Read Cache State �
SP A � Enable.
To enable array read caching for SP
B
If the array has an SP B, then on the array toolbar, either click the SP B Enable Read Cache button
or select the menu option Array � Read Cache State �
SP B � Enable.
To enable array write caching
On the array toolbar, either click the Write Cache Enable button or select the menu option Array
� Write Cache State �
Enable.
 |
NOTE:
If the array does not have all the required
components
for array write caching, Data Supervisor tells you that the write cache was successfully enabled when
in fact it was not enabled. As soon as all the required components are installed in the array, the
SPs enable write caching. (The required components are listed in
"Hardware requirements for
caching".) |
You have set up the array to perform read and/or write caching. The array will use the cache page
size you specified as the I/O size and the default values for the low and high watermarks. For
information on the low and high watermark parameters and how to change them, see the section
"Changing array caching
parameters".
What next?
If you are setting up a new storage system or want to create new LUNs, continue to next section
"Creating LUNs."
You must bind disk modules into LUNs so that the server's operating system will recognize them.
This
section describes the bind parameters that you specify when you create a LUN, and describes how to bind
disk modules into LUNs. It also describes how to edit the device information in the Data Agent
configuration file after binding LUNs.
Number of disks in LUNs
The number of disk modules for the different types of LUNs are
| LUN Type |
Number of Disk Modules |
LUN Type |
Number of Disk Modules |
| RAID 0 |
3 minimum 16
maximum |
RAID 5 |
3 minimum 16
maximum |
| RAID 1 |
2 |
DISK |
1 |
| RAID 1/0 |
Even number 4 minimum 16
maximum |
HOT SPARE |
1 |
| RAID 3 |
5 or 9 |
|
|
The bind parameters are rebuild time, verify time, element size, read cache state, write cache state,
auto assignment state, minimal latency reads state, and default SP. The bind parameters available for
the different types of LUNs (that is, different RAID types) are
| LUN type |
Parameters |
| Rebuild time |
Verify time |
Element size |
Read cache state |
Write cache state |
Auto assignment state |
Minimal latency reads
state |
Default SP |
| RAID 0 |
No |
No |
Yes |
Yes |
Yes |
Yes |
No |
Yes |
| RAID 1 |
Yes |
Yes |
No |
Yes |
Yes |
Yes |
No |
Yes |
| RAID 1/0 |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
No |
Yes |
| RAID 3 |
Yes |
Yes |
No |
No |
No |
Yes |
Yes |
Yes |
| RAID 5 |
Yes |
Yes |
Yes |
Yes |
Yes |
Yes |
No |
Yes |
| DISK |
No |
No |
No |
Yes |
Yes |
Yes |
No |
Yes |
| HOT SPARE |
No |
No |
No |
No |
No |
No |
No |
No |
Rebuild time
The rebuild time is the amount of time that the array allots to reconstruct the data on either a
hot spare or a new disk module that replaces a failed disk module in a LUN. The time you specify
determines the amount of resources the SP devotes to rebuilding instead of to normal I/O activity.
Verify time
The verify time is the amount of time that the array allots to checking parity. If an SP detects
parity inconsistencies, it starts a background process to check all the parity sectors in the LUN.
The time you specify determines the amount of resources the SP devotes to verifying instead of to
normal I/O activity.
Element size
The stripe element size is the number of disk sectors that the array can read or write to a single
disk module without requiring access to another disk module (assuming that the transfer starts at the
first sector in the stripe). The stripe element size can affect the performance of a RAID 5 or RAID
1/0 LUN. A RAID 3 LUN has a fixed stripe element size of one sector.
The smaller the stripe element size, the more efficient the distribution of data read or written.
However, if the stripe size is too small for a single I/O operation, the operation requires access
to two stripes, which causes the hardware to read and/or write from two disk modules instead of one.
Read cache state
You can enable or disable the read cache for a LUN. The SP A read cache memory that you allocated
when you partitioned memory is shared by all LUNs for which the read cache is enabled and that are
assigned to (owned by) SP A. Likewise, the SP B read cache memory that you allocated when you
partitioned memory is shared by all LUNs for which the read cache is enabled and that are assigned
to (owned by) SP B.
Write cache state
You can enable or disable the write cache for a LUN. The array write cache memory that you
allocated
when you partitioned memory is shared by all LUNs for which the write cache is enabled.
Auto assignment state
You can enable or disable auto assignment for a LUN. Auto assignment controls the ownership of the
LUN when an SP fails in a array with two SPs. With auto assignment enabled, if the SP that owns a LUN
fails and the server tries to access that LUN through the second SP, the second SP assumes ownership
of the LUN so the access can occur. The second SP continues to own the LUN until the failed SP is
replaced and the array's power is turned off and on again. Then, ownership of each LUN returns to its
default SP. If auto assignment is disabled in the previous situation, the other SP does not
assume ownership of the LUN, so the access to the LUN does not occur.
Minimal latency reads state
You can enable or disable minimal latency reads for a LUN. Minimal latency reads provide a more
constant bandwidth between the SP and the server when the bandwidth between the disk modules in a
RAID 3 LUN and the SP decreases because of a slow response from a disk module. Minimal latency
reads change how the SP responds to a request from the server to read a RAID 3 LUN.
When minimal latency reads are disabled, the SP responds to a read request to a RAID 3 LUN with
n disk modules as follows:
- It issues a read request to the n-1 disk modules in the RAID 3 LUN that contain data.
- It performs a checksum on the data as it receives the data from the n-1 disk modules.
- If the checksum is correct, it sends the data to the server; if the checksum is incorrect, it reads
the parity from the remaining disk module and uses it to reconstruct the data before sending the data
to the server.
When minimal latency reads are enabled, the SP responds to a read request to a RAID 3 LUN with n
disk modules as follows:
- It issues a read request to all n disk modules in the LUN.
- It performs a checksum on the data as it receives it from the first n-1 drives to respond.
- If the checksum is correct, it sends the data to the server; if the checksum is incorrect, it uses
the parity it has received to reconstruct the data before sending the data to the server.
Because minimal latency reads always require access to all the disk modules in the RAID 3 LUN, they
nominally decrease the bandwidth between the disk modules and the SPs.
Default SP
The default SP is the SP that assumes ownership of the LUN after the array's power is turned off
and then on again. If the array has two SPs, you can choose to bind some LUNs using one SP as the
default, and the rest using the other SP as the default. You would do this to balance the load
across the SPs or to establish the primary route for the LUN in a dual-server configuration. The
primary route to a LUN is the route through the default SP, and the secondary route is through the
other SP.
What next?
Proceed to next section "Binding disk modules into LUNs."
Before you create LUNs, you need to know which disk modules you want to use for each LUN. Make
sure
that the array has disk modules in each of the required slots, and that these disk modules are not
already bound into LUNs.
You may need to move disk modules. Generally, modules should not be moved from one slot to
another;
but if moving one is absolutely necessary, the system operator or service person can move it with the
following cautions:
- The disk module must be unbound. Moving a module that is part of a LUN to another slot makes all
information on the LUN inaccessible. The "Unbinding a LUN"
section explains how to unbind LUNs.
- You must remove and install the disk module while the array is powered up, and use the procedure
explained in the Dell PowerVault 650F Rackmount Storage System Installation and Service Guide
(P/N 5867C) or the Dell PowerVault 651F Deskside Storage System Installation and Service
Guide
(P/N 3867C).
In a Fibre Channel array with Fibre Channel disks, you can choose any disk modules for a LUN of
any
RAID type without affecting the performance or high availability of the LUN. For simplicity, however,
we recommend that you choose consecutive disk modules.
Before you start binding disk modules into LUNs, read the restrictions and recommendations in the
table that follows.
Table 6-1. Restrictions and recommendations for binding disk modules into LUNs
| Disk module |
Restrictions and recommendations for
binding |
| Any LUN |
You can bind only unbound disk modules. All
disk modules
in a LUN must have the same capacity to fully use the modules' storage space. |
| RAID 5 |
You must bind a minimum of three
disk modules and
no more than sixteen disk modules. We recommend you bind five modules for more efficient use
of disk space. You can bind one less module per LUN than you will eventually use by selecting an
empty slot icon. However, the LUN will operate in a degraded mode until a module is installed
in the empty slot, and the array integrates it into the LUN. You can select the modules in any
order. |
| RAID 3 |
You must bind exactly five or nine
disk modules.
You can bind one less module per LUN than you will eventually use by selecting an empty slot
icon. However, the LUN operates in a degraded mode until a module is installed in the empty slot,
and the array integrates it into the LUN. You can select the modules in any order. You cannot bind a
custom RAID 3 LUN until you have allocated array memory for the LUN. |
| RAID 1 |
You must bind exactly two disk
modules. You can
select the modules in any order. |
| RAID 0 |
You must bind a minimum of three
disk modules,
and no more than sixteen disk modules. You can select the modules in any order. |
| RAID 1/0 |
You must bind a minimum of four disk
modules,
and an even number of modules, but no more than sixteen modules. Data Supervisor pairs
modules into mirrored images in the order in which you select them. The first and second modules
you select are a pair of mirrored images; the third and fourth modules you select are another
pair of mirrored images; and so on. The first module you select in each pair is the primary image,
and the second module is the secondary image. |
| Individual disk unit |
None |
| Hot spare |
You cannot bind disk modules 0:0
through 0:8 as
hot spares. The capacity of a disk module bound as a hot spare must be at least as great as the
capacity of the largest disk module that it might replace. |
To bind disk modules into LUNs
 |
NOTE:
If you bind disk modules into a RAID 3 LUN
without allocating
adequate memory to the RAID 3 partition, the LUN will be unowned. You will not be able to access the
LUN
until you allocate memory to the RAID 3 partition and rebind the LUN. |
- If no LUNs exist on any array on which you want to create LUNs, edit the
device
information in the Data Agent configuration file on each server connected to the array, as described
in "Editing
or checking device information in the Data Agent configuration file"
You can tell if an array has LUNs by displaying the Array Configuration window for it (see
"To display the Array
Configuration window").
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, either click the Bind LUN button on the array toolbar or select
the menu option Array � Bind LUN.
The Bind LUNs windows for the selected arrays open, similar to the following sample.
Figure 6-8. Sample of Bind LUNs window

- If the array has more than one enclosure, select the enclosures
containing the disk modules
that you want to bind as follows:
To select disk modules from all enclosures
If the Unbound Disks field contains "All Chassis," continue to step 5. If it contains
the name of an enclosure, pull down its selection list and select "All Chassis."
To select disk modules from one enclosure
If the Unbound Disks field contains the name of the desired enclosure, continue to step 5.
Otherwise, pull down its selection list and select the name of the desired enclosure.
- Move the disk modules that you want to bind into a LUN from the Unbound Disks area
to the
Bind Disks area as follows:
Either
Use both mouse buttons on a two-button mouse or the middle button on a three-button mouse to drag
and drop the desired disk modules from the Unbound Disk area to the Bind Disks area.
Or
Select the desired disk modules, and then click the right arrow button.
If you move a wrong disk module into the Bind Disks area, simply select it and click the left arrow
button to move it back into the Unbound Disks area. Or you can drag and drop it into the Unbound Disks
area.
 |
NOTE:
Do not select any bind parameters for the
LUN until you
have selected all the disk modules for it. The number of disk modules you select determines which RAID
types are available, and the RAID type determines availability of the other bind
parameters. |
If you do not want to change any bind parameters, go to step 16.
Figure 6-9. Sample of Bind LUNs window with disk modules selected for binding

- If you want the LUN to be a different RAID type than the one in the RAID Type field,
click
the field list button and select the desired type from the list that opens.
For example,

Only the RAID types that are available for the number of selected disk modules appear in the list.
- If you want the LUN to have a different hexadecimal identifier (ID) than the number in the LUN ID
field, click the field list button and select the desired number from the list that opens.
For example,

The default LUN ID is the next hexadecimal number available, starting with 0 and ending with 1f.
Only available numbers appear in the list. The default number is 0 for the first LUN that you bind,
regardless of the number of SPs or servers attached to the array. The default number for the second
LUN you bind is 1; for the third LUN, it is 2; for the fourth LUN, it is 3, and so on. Depending on
the operating system on the server and type of host bus adapter, you can have a maximum of 8, 16,
or 32 LUNs. If you want to specify a nondefault number you can do so. After you bind a LUN with a
nondefault number, the default number for the next LUN is the lowest number you skipped.
- If the LUN is not a RAID 0 LUN, individual disk, or hot spare, and you want it to have a
different rebuild time than the time in the Rebuild Time field, either enter the number of hours in
the field or click the field list button and select the desired number of hours from the list that
opens.
For example,

The rebuild time is the amount of time that the array allots to reconstruct the data on either a
hot spare or a new disk module that replaces a failed disk module in a LUN. The time you specify
determines the amount of resource the SP devotes to rebuilding instead of to normal I/O activity.
The default time of 4 hours is adequate for most situations. A rebuild time of 2 hours rebuilds the
disk more quickly, but slightly degrades response time. A rebuild time of ASAP (as soon as possible)
rebuilds the disk module as quickly as possible, but significantly degrades response time.
The actual rebuild time can differ significantly from the time you specify, especially for a RAID
1/0 LUN, or a LUN containing 9-Gbyte disk modules. Since a RAID 1/0 with n disk modules can continue
functioning with up to as many as n/2 failed disk modules and only one disk module at a time is
rebuilt, the actual rebuild time for such a LUN is the time you specify multiplied by the number of
failed disk modules.
- If the LUN is not a RAID 0 LUN, individual disk, or hot spare, and you want it
to
have a different verify time than the time in the Verify Time field, either enter the number of
hours in the field or click the field list button and select the desired number of hours from the
list that opens.
For example,

The verify time is the amount of time that the array allots to checking parity. If an SP detects
parity inconsistencies, it starts a background process to check all the parity sectors in the LUN.
The time you specify determines the amount of resource the SP devotes to verifying instead of to
normal I/O activity. The default time of 4 hours is adequate for most situations. A verify time of
2 hours checks the parity sectors more quickly, but slightly degrades response time. A verify time
of ASAP (as soon as possible) checks the parity sectors as quickly as possible, but significantly
degrades response time. If your site requires fast response time and you want to minimize
degradation to normal I/O activity, you can extend the rebuilding process over the maximum time of
4 hours.
- For a RAID 0, RAID 1/0, or RAID 5 LUN, if you want the LUN to have a stripe element
size
with a different number of sectors than the number in the Element Size field, click the field list
button and select the desired number of sectors from the list that opens.
For example,

The stripe element size is the number of disk sectors that the array can read or write to a single
disk module without requiring access to another disk module (assuming that the transfer starts at the
first sector in the stripe). The stripe element size can affect the performance of a RAID 5 or RAID
1/0 LUN. A RAID 3 LUN has a fixed stripe element size of one sector.
The smaller the stripe element size, the more efficient the distribution of data read or written.
However, if the stripe size is too small for a single I/O operation, the operation requires access
to two stripes, which causes the hardware to read and/or write from two disk modules instead of one.
Generally, we recommend the smallest stripe element size that rarely forces access to another
stripe. The default stripe element size is 128 sectors.
- If the LUN is not a RAID 3 LUN or a hot spare, and you want to change the
read cache
state for the LUN, click the Read Cache box.
For example,

A check in the box tells you that the read cache is enabled for the LUN.
We recommend that you enable the read cache for any LUN for which the option is available.
If you enable the read cache for a LUN, caching occurs only when the array read cache is enabled
for the default SP. You enable the array read cache for SP A from the Array Configuration window
using either the SP A Enable Read Cache button on the array toolbar or the menu path Array
� Read Cache State � SP
A
� Enable. The SP read cache memory that you allocated when you
partitioned memory is shared by all LUNs for which the read cache is enabled.
- If the LUN is not a RAID 3 LUN or a hot spare, and you want to change the
write
cache state for the LUN, click the Write Cache box.
For example,

A check in the box tells you that the write cache is enabled for the LUN.
We highly recommend that you enable the write cache for a RAID 5 LUN and recommend that you
enable
it for any other type of LUN for which the option is available.
If you enable the write cache for a LUN, caching occurs only when the array write cache is enabled.
You enable the array write cache from the Array Configuration window using either the Write Cache
Enable button on the array toolbar or the menu path Array �
Write Cache
State � Enable. The write cache memory that you allocated when
you
partitioned memory is shared by all LUNs for which the write cache is enabled.
- If the LUN is not a hot spare and you want to change the auto assignment state for the
LUN,
click the Auto Assign box.
For example,

A check in the box tells you that auto assignment is enabled for the LUN.
You can enable auto assignment for arrays with two SPs only. Data Supervisor disables this option
for arrays with one SP. As a general rule, you should disable auto assignment; however, some
applications require that it be enabled, as explained in the application's documentation.
Auto assignment controls the ownership of the LUN when an SP fails in a array with two SPs. With
auto assignment enabled, if the SP that owns a LUN fails and the server tries to access that LUN
through the second SP, the second SP assumes ownership of the LUN so the access can occur. The
second SP continues to own the LUN until the failed SP is replaced and the array's power is turned
off and then on again. Then, ownership of each LUN returns to its default SP. If auto assignment is
disabled in the previous situation, the other SP does not assume ownership of the LUN, so the
access to the LUN does not occur.
- If you want to change the minimal latency reads state for a RAID 3 LUN, click the
Minimal
Latency Reads box.
For example,

A check in the box tells you that minimal latency reads are enabled for the LUN.
You can enable minimal latency reads for a RAID 3 LUN only.
Minimal latency reads provide a more constant bandwidth between the SP and the server when the
bandwidth
between the disk modules in a RAID 3 LUN and the SP decreases because of a slow response from a disk
module.
Minimal latency reads change how the SP responds to a request from the server to read a RAID 3
LUN.
When minimal latency reads are disabled, the SP responds to a read request to a RAID 3 LUN with n
disk modules as follows:
- It issues a read request to the n-1 disk modules in the RAID 3 LUN that contain data.
- It performs a checksum on the data as it receives the data from the n-1 disk modules.
- If the checksum is correct, it sends the data to the server; if the checksum is incorrect, it reads
the parity from the remaining disk module and uses it to reconstruct the data before sending the data
to the server.
When minimal latency reads are enabled, the SP responds to a read request to a RAID 3 LUN with n
disk modules as follows:
- It issues a read request to all n disk modules in the LUN.
- It performs a checksum on the data as it receives it from the first n-1 drives to respond.
- If the checksum is correct, it sends the data to the server; if the checksum is incorrect, it uses
the parity it has received to reconstruct the data before sending the data to the server.
Because minimal latency reads always require access to all the disk modules in the RAID 3 LUN, they
nominally decrease the bandwidth between the disk modules and the SPs. For a RAID 3 LUN
with
five disk modules, the bandwidth decreases by about 25%; for one with nine disk modules, it decreases
about 12%.
- If the LUN is not a hot spare, and you want the SP without the dot in its button
to be the default owner of the LUN, click the button for that SP.
For example,

You can change the default SP in arrays with two SPs only. Data Supervisor disables this option for
arrays with one SP. When you select one SP, the other SP is automatically deselected.
The default SP is the one that assumes ownership of the LUN after the array's power is turned off and
then on again. If the array has two SPs, you can choose to bind some LUNs using one SP as the default,
and the rest using the other SP as the default. You would do this to balance the load across the SPs or
to establish the primary route for the LUN in a dual-server configuration. The primary default SP, and
the secondary route is through the other SP.
- When all bind parameters for the LUN are set as you want, click
the
Bind button.
A confirmation window opens.
- In the window, click the Yes button to initiate the bind operation.
A window opens informing you whether the bind operation was initiated successfully or not.
- In the window, click the OK button.
A blue icon for the LUN appears in the Unowned LUNs area of the Array Configuration window.
Binding
takes a while; just how long varies with the type of SP and size of the disk modules. When the bind
operation is completed, the icon moves to the area for its default SP and becomes grey.
- If you want to create another LUN, repeat step 4 through step
17.
If you want to configure another array on the server, you need to exit.
 |
NOTE:
If you have other arrays on the server that
need configuring,
you may want to configure them before completing the next step. To configure another array on the
server,
you must exit this Data Supervisor session and start a session for that array. |
- Reboot the operating system on the server for it to recognize the newly-created
LUNs.
What next?
After you create LUNs on an array, you need to edit or check the device information in the Data
Agent
configuration file on the servers connected to the array, as described in the next section.
Whenever you create one or more LUNs on an array, you need to edit the Data Agent configuration
file
on the servers connected to the array, as described below.
 |
NOTE:
You also need to edit the device information
before
binding any LUNs on an array that does not have a bound LUN. |
You may want to refer to the Dell OpenManage Data Agent for NT and Integrator
Installation
and Operation Guide (P/N 3967C).
To edit device information in the Data Agent configuration file on an array server
- On the server, start the Data Agent Configurator.
The Agent Configuration window opens.
- On the window's toolbar, click the Clear Device List button.
- On the windows' toolbar, click the Auto Detect Array button.
- Save the Data Agent configuration file.
- When you are asked if you want to restart the Agent, click the Yes button.
What next?
From viewpoint of a server's operating system, the LUNs in the array are identical to standard single
disk drives.
On powerup, an array enables the read and write caches on each SP if the prerequisite hardware is
working and the Read and Write Cache partitions have nonzero sizes. Array read and write
caching is enabled when the SP Read and Write Caches are enabled.
You can disable or enable array read or write caching without affecting the information stored on
the LUNs. You must enable the array read and/or write caching for the array to use caching for LUNs
with their read or write caches enabled. You must disable the array read and/or write caches before
changing the memory partitions and most of the LUN caching parameters. We also recommend that you
disable the array write caching before replacing an SP and before downloading LIC.
You disable or enable array write caching by setting the state of each SP's Write Cache to Disabled
or Enabled, respectively. Since the state of the Write Cache on each SP is always the same, you set
the state of each cache together. You disable or enable read caching for an SP by setting the state
of the SP's read cache to Disabled or Enabled, respectively. Since the state of the read cache on one
SP is independent of the state of the read cache state on the other SP, you must set the state of each
cache separately. This section describes how to disable or enable write caching for the array (both
SPs) and how to disable and enable read caching for each SP.
The rest of this section describes how to
- Determine the array cache state
- Disable or enable array write caching
- Disable or enable array read caching for an SP
To determine if array write caching is enabled or disabled
Look at the Write Cache State in the cache information part of the SP's Information window for
either
SP (see "Displaying SP cache
information").
To determine if array read caching for an SP is enabled or disabled
Look at the Read Cache State entry in the cache view of the SP's Information window (see
"Displaying SP cache
information").
To disable array write caching
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, either click the Write Cache Disable button on the array toolbar
or select the menu option Array � Write Cache State �
Disable.
- Check that array write caching is disabled by looking at the Write Cache State entry on the cache
view of the SP Information window (see
"Displaying SP cache
Information").
It may take the array a while to disable write caching if the SPs need to write data in the write
cache to disks. As a result, the Write Cache State entry in the cache view of the SP Information
window may be Enabled for a while. You may want to poll the array every few seconds to make sure you
have the latest status. You can poll the array using either the Poll button on the array toolbar or
the menu option Array � Poll.
To enable array write caching
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, either click the Write Cache Enable button on the array toolbar
or select the menu option Array � Write Cache State �
Enable.
- Check that array write caching is enabled by looking at the Write Cache State entry on the cache
information part of the SP Information window (see
"Displaying SP cache
information").
You may want to poll the array to make sure you have the latest status. You can poll the array using
either the Poll button on the array toolbar or the menu option Array � Poll.
If write caching is not enabled, make sure that the array has two SPs, disk modules in DPE slots 0
through 8, a standby power supply (SPS), and a write cache partition of at least 2 Mbytes.
 |
NOTE:
If the array does not have all the required
components
for array write caching, Data Supervisor tells you that the write cache was successfully enabled when
in fact it was not enabled. As soon as all the required components are installed in the array, the SPs
enable write caching. (The required components are listed in
"Hardware requirements for
caching".) |
To disable array read caching for an
SP
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, disable read caching for the array's SP A or SP B as
follows:
For SP A
If the array has an SP A, then on the array toolbar, either click the SP A Disable Read Cache Button
or select the menu option Array � Read Cache State �
SP A � Disable.
For SP B
If the array has an SP B, then on the array toolbar, either click the SP B Disable Read Cache button
or select the menu option Array � State � SP B
� Disable.
- Check that array read caching is disabled for the appropriate SP by looking at the Read
Cache State entry on the cache information part of the SP Information window (see
"Displaying SP cache
information").
You may want to poll the array to make sure you have the latest status. You can poll the array using
either the Poll button on the array toolbar or the menu option Array � Poll.
To enable array read caching for an
SP
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, enable read caching for the array's SP A or SP B as
follows:
For SP A
If the array has an SP A, then on the array toolbar, either click the SP A Enable Read Cache Button
or select the menu option Array � Read Cache State �
SP A � Enable.
For SP B
If the array has an SP B, then on the array toolbar, either click the SP B Enable Read Cache button
or select the menu option Array � Read Cache State �
SP B � Enable.
- Check that array read caching is enabled for the appropriate SP by looking at the Write
Cache State entry on the cache information part of the SP Information window (see
"Displaying SP cache
information").
You may want to poll the array to make sure you have the latest status. You can poll the array using
either the Poll button on the array toolbar or the menu option Array � Poll.
If read caching is not enabled, make sure that the read cache partition for the SP is of at least 1
Mbytes.
This section describes how to change the following caching parameters:
- Page size
- Write cache high and low watermarks
The cache is managed by pages instead of sectors. The page size specifies the number of Kbytes stored
in one cache page.
| Available page size values: |
2, 4, 8, or 16 Kbytes |
| Default page size value: |
2 Kbytes |
The larger the page size, the more continuous sectors the cache stores in a single page. As a general
guideline, the page size should be 8 Kbytes for general file server applications and 2 to 4 Kbytes for
database applications.
You can determine the cache page size by looking at the cache information in the SP Information
window for either SP (see
"Displaying SP cache
information").
To change the cache page size for the selected arrays
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- Disable array write caching (see
"To disable array write caching").
- Disable array read caching for each SP (see
"To disable array read caching for an
SP").
- In the Array Configuration window, follow the menu path Array � Set Page Size.
- In the menu that appears, select the value for the page size.
- Re-enable array write caching (see
"To enable array write caching").
- Re-enable read caching for each SP (see
"To enable array read caching for an
SP").
Changing the write cache high or low watermarks
The write cache high and low watermarks determine when the SPs start and stop flushing their write
caches, respectively. When an SP flushes its write cache, it writes its dirty pages to disk. A dirty
page is a write cache page with modified data that has not been written to disk.
The high watermark is the percentage of dirty pages in the write cache, which when reached causes
the SPs to begin flushing their write cache. The default value is 96 percent. If you specify a lower
value, the SPs start flushing the write cache sooner.
The low watermark is the percentage of dirty pages in the write cache that determines when write
cache flushing stops. When the low watermark is reached during a flush operation, the SPs stop
flushing the write cache.
 |
NOTE:
The high watermark cannot be less than the
low watermark.
To turn off watermark processing, set both the low and high watermarks to 100. |
| Available watermark values: |
0 through 100% |
| Default watermark values: |
96% for high watermark 80% for low watermark |
The rest of this section describes how to change the high or low watermarks.
To change the high watermark
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, follow the menu path Array � Set
Watermark � High Watermark.
- In the window that opens, enter the high watermark value.
To change the low watermark
- Display the Array Configuration window (see
"To display the Array
Configuration window").
- In the Array Configuration window, follow the menu path Array � Set
Watermark � Low Watermark.
- In the window that opens, enter the low watermark value.
To upgrade an array to support caching, the system operator or service person must install the
necessary
hardware components, and then you must set up array caching. This section describes how to perform
each of
these tasks.
Installing the hardware components for caching
To support read caching, an array must have an SP with at least 128 Mbytes of memory.
To support write caching, an array must have the following hardware:
- Two SPs.
- Two power supplies and two LCCs in the DPE and each DAE.
- Disk modules in DPE slots 0 through 8.
- Standby power supply (SPS) with a fully charged battery.
The system operator or service person can install memory modules, disk modules, a second SP, LCC,
and power supply without powering down the array. If you add disk modules, you need to bind them
into new LUNs (see "Creating LUNs") or change the
existing
LUNs to include the new disk modules (see
"Changing a LUN
RAID type or its number of disk modules").
If you add a second SP, you may want it to own some of the LUNs. You can switch the ownership of
a LUN from one SP to the new SP (see
"Transferring
default ownership of a LUN from one SP to another").
Setting up caching
Once you have installed the necessary hardware components, follow these steps to set up
caching:
- Enable array write caching and read caching for the desired SPs (see
"Enabling or disabling array
caching").
- Enable read or write caching for the LUNs that you want to use read or write caching (see
"Changing any LUN parameter except for the default SP and prefetch
parameters").
The Licensed Internal Code (LIC) media may also include an update to the SP programmable read-
only
memory (PROM) code. When you install LIC, the SP tries to copy it to reserved areas outside operating
system control on several disk modules, which are called the database disks. Having multiple copies
of code offers higher availability if a disk module fails. The database disk modules have disk IDs
00, 01, and 02, as shown in the figure that follows.
Figure 6-10. Disk modules that store LIC

When you install LIC, at least two of the database disk modules must be on line, and ideally, all
of them should be on line. A disk module is on line if it is fully powered up and not faulted; that
is, if its state in the Disk Information window is Normal. If you try to power up the array without
two of these disk modules in place, the powerup fails.
To update LIC and PROM code
The following procedure assumes that the system operator or service person has loaded the new
revision
of the LIC onto the server to which the array is connected. If the LIC media also includes new PROM
code,
the PROM code is loaded along with the LIC.
 |
NOTE:
Before updating LIC and PROM code, you
must disable
array write caching. |
- Disable array write caching (see
"Enabling or disabling array
caching").
- Display the Array Configuration window
(see "To display the
Array Configuration window").
- In the Array Configuration window, select the menu option Array � Firmware
Download.
A Firmware Download window opens, similar to the following.
The Array List gives the name of the array whose LIC will be updated.
Figure 6-11. Firmware Download window

- In the Enter Full Path Name of File field, enter the complete path name of the file with
the revision of LIC that you want to download.
 |
NOTE:
The LIC file must be on a server connected to
the array
in the Array list. |
- In the Select Host field, click the field list button, and from the list that appears, select
the name of the server where the LIC file resides.
The SPs must be rebooted for the new revision of LIC to take effect. If you select automatic reboot,
the SPs are rebooted automatically when the LIC download operation is complete. When you reboot the
array's SPs you may have to restart the Data Agent on the array's server.
 |
CAUTION:
If the new LIC revision you are
downloading includes
a more recent revision of PROM code than the one running in the SPs, you must select the automatic
reboot option because the first time you load a new LIC revision is the only time that the SPs will
update PROM code. Rebooting the SPs in any other way after the download does not update PROM code.
A revision of LIC often requires a specific PROM code revision to work properly.
Rebooting restarts the SPs in the array, which terminates all outstanding I/O to the array. If you
enable automatic reboot for an array, do not complete this procedure until you have stopped all I/O
to the array. |
- If you want to automatically reboot the SPs when the download is completed, click the Automatic
Reboot button.
- In the Firmware Download window, click the OK button.
A confirmation window containing the LIC license agreement opens.
- In the window, click the Yes button to accept the license agreement and start
downloading
LIC.
A window opens saying that the firmware was loaded to the array successfully.
- When Data Supervisor tells you that the firmware was loaded successfully, re-enable
array
write caching (see
"Enabling or disabling array
caching").
 |
NOTE:
The size of the new LIC revision may cause
memory sizing
constraints that prevent you from re-enabling the array write cache or that disable the array read cache
on an SP and prevent you from re-enabling it. If this happens, you should resize the memory partitions
either after the SPs reboot automatically (if you enabled automatic reboot) or after you reboot them
in step 10 (see next step). (if you did not enable it). |
- If you did not enable automatic reboot, reboot the SPs.
You reboot the SPs in the array either by powering it down and then up again or by using the Data
CLI command rebootSP. If you use the CLI command, you may have to stop and restart the Data
Agent on the server after the SPs reboot. For information on the rebootSP command, see the
Dell OpenManage Data CLI for NT Installation and Operation Guide (P/N
4967C).
Back to Contents Page