 |
» |
|
|
|
|
|
|
|
|
HP StorageWorks Disk Array XP - Operating System Configuration Guide IBM AIX
This Document describes the requirements and procedures for connecting the XP family of disk
arrays to an IBM AIX system and configuring the new disk array for operation with AIX.
Features and requirements
The disk array and host have the following features and requirements.
HP StorageWorks disk arrays:
XP48: Up to 48 drives from 72 GB to 8.7 TB, 24 FC ports
XP128: From 8 to 128 drives for up to 18 TB, 48 FC ports
XP512: Up to 512 drives from 72 GB to 93 TB, 48 FC ports
XP1024: From 8 to 1024 drives for up to 149 TB, 64 FC ports
XP12000: Up to 1152 drives for up to 165 TB, 128 FC ports
IBM RS/6000 series, POWERstation, POWERserver, or SP series
IBM AIX operating system with current OS patches
superuser (root) login access to the system
Host Bus Adapters (HBAs): Install adapters and all utilities and
drivers. Refer to the adapter documentation for installation details.
(Recommended) HP StorageWorks Command View XP with LUN
management feature or Remote Control with the LUN Configuration
Manager XP option for configuring disk array ports and paths.
(Recommended) HP StorageWorks Secure Manager XP:
Allows the
host to access only array devices for which it is authorized.
Other available XP Software (some may not apply to your system):
HP StorageWorks Business Copy XP
HP StorageWorks Continuous Access XP
HP StorageWorks Continuous Access Extension XP
HP StorageWorks Auto LUN XP
HP StorageWorks Data Exchange XP
HP StorageWorks Resource Manager XP
HP StorageWorks RAID Manager XP
HP StorageWorks Cache LUN XP
HP StorageWorks Auto Path XP
HP StorageWorks Cluster Extension XP
HP StorageWorks Performance Advisor XP software
Fibre Channel interface
The XP48, XP128, XP512, XP1024, and XP12000 disk arrays support
these 1 Gbps and 2 Gbps Fibre Channel interfaces:
Short-wave non-OFC (open fiber control) optical interface
Multimode optical cables with SC or LC connectors
Public or private arbitrated loop (FC-AL) or fabric direct attach
Fibre Channel switches
Even though the interface is Fibre Channel, this guide uses the term "SCSI
disk" because disk array devices are defined to the host as SCSI disks.
Device types
The disk arrays support the following device types: OPEN-x devices:
OPEN-x logical units represent disk devices.
Except for OPEN-V, these devices are based on fixed sizes. OPEN-V
is a user-defined size. Supported emulations include OPEN-3,
OPEN-8, OPEN-9, OPEN-E, OPEN-L, and OPEN-V devices.
LUSE devices (OPEN-x*n):
Logical Unit Size Expansion (LUSE)
allows you to combine 2 to 36 OPEN-x devices to create expanded
LDEVs larger than standard OPEN-x disk devices. For example, an
OPEN-x LUSE volume created from ten OPEN-x CVS volumes is
designated as OPEN-x*10.
CVS devices (OPEN-x CVS):
Volume Size Configuration (VSC)
defines custom volumes (CVS) that are smaller than normal fixed-sized logical disk devices (volumes). (OPEN-V is a CVS-based
custom disk size that you determine. OPEN-L does not support CVS.)
LUSE (expanded) CVS devices (OPEN-x*n CVS):
LUSE CVS
combines CVS devices to create an expanded device. This is done by
first creating CVS custom-sized devices and then using LUSE to
combine from 2 to 36 CVS devices. For example, if three OPEN-9
CVS volumes are combined to create an expanded device, this device
is designated as OPEN-9*3-CVS.
Failover
The disk arrays support many standard software products that provide host,
application, or I/O path failover and logical volume (storage) management.
SNMP configuration
The disk arrays support standard Simple Network Management Protocol
(SNMP) for remotely managing the disk array from the host. The SNMP
agent on the remote console PC or Command View can provide status and
Remote Service Information Message (R-SIM) reporting to the SNMP
manager on the host for up to eight disk arrays. To configure the SNMP
manager on the host, refer to the operating system documentation. 
1 - Disk array
2 - SIM
3 - SNMP
4 - Remote Console
5 - Error info
6 - Public
7 - SNMP manager
8 - Open system host
RAID Manager command devices
RAID Manager manages Business Copy (BC) and/or Continuous
Access (CA) operations from a server host. To use RAID Manager
with BC or CA, you must use Command View or LUN Configuration
Manager to designate at least one LDEV as a command device. Refer
to the Command View or LUN Configuration Manager user guide
for information about how to designate a command device.
Installation procedures
Install and configure the disk array
The HP service representative performs the following tasks:
Assembling hardware and installing software
Loading the microcode updates
Installing the channel adapters (CHAs) and cabling
Installing and formatting devices
You perform the additional tasks below. If you do not have Command View
or LUN Configuration Manager, your HP service representative can
perform these tasks for you.
Setting the System Option Modes
The HP representative sets the System Option Mode(s) based on the
operating system and software configuration of the host.
Configuring the Fibre Channel ports
Configure the disk array Fibre Channel ports by using Command View
or the Fibre Parameter window in LUN Configuration Manager.
Select the settings for each port based on your storage area network
topology. Use switch zoning if you connect different types of hosts to the
array through the same switch.
Fibre Address
In fabric environments, the port addresses are assigned automatically. In
arbitrated loop environments, you set the port addresses by selecting a
unique arbitrated loop physical address (AL-PA) or loop ID for each port.
Fabric and Connection parameter settings
You can set each array port to FABRIC ON or OFF with connections of
POINT-TO-POINT or FC-AL as shown in the following table and figures.
For detailed topology information, refer to the HP StorageWorks SAN
Design Reference Guide
on the
http://www.hp.com
Web site. Figure 1: Simple Point-to-Point Fabric Topology Example
1 - Server
2 - N Port
3 - F Port
4 - Fabric Switch
5 - F Port
6 - N Port
7 - Disk Array
Figure 2: Arbitrated Loop Fabric Topology Example
1 - Server
2 - NL Port
3 - FL Port
4 - Fabric Switch
5 - FL Port
6 - NL Port
7 - Disk Array
| Fabric
Parameter
|
Connection
Parameter
|
Provides
|
|---|
| ON |
FC-AL |
NL-port (SAN public arbitrated loop)
|
| ON |
Point-to-Point
|
N-port (SAN fabric port) |
| OFF
|
FC-AL |
NL-port (private arbitrated loop; direct
connect without a SAN) |
| OFF |
Point-to-Point |
Not supported |
Setting the Host Mode for the disk array ports
The disk array ports have Host Modes that you must set depending on the
host you use. After the disk array is installed, use Command View (shown)
or LUN Configuration Manager to set the Host Mode for each port. Figure 3: The host mode for AIX is OF
Install and configure the host
Install and configure the host and host bus adapters (HBAs) that connect
the host to the disk array.
Loading the OS and software
Follow the manufacturer’s instructions to load the operating system and
software onto the host. Load all OS patches and configuration utilities
supported by HP and the HBA manufacturer.
Installing and configuring the HBAs
Install and configure the host bus adapters using the HBA manufacturer’s
instructions.
Supported HBAs:
Supported HBAs include the IBM FC6227, IBM FC6228, and IBM
FC6239.
To check whether the drivers are installed:
Example (FC 6227)
Use the lslpp
command to display the drivers currently installed on the
system.
# lslpp -l | grep df1000f7
Check the list for the required two drivers:
For the IBM FC 6227 HBA, the following drivers are required:
devices.pci.df1000f7
devices.fcp.disk
If the drivers are displayed, you do not need to install the drivers.
If the drivers are not displayed, install the drivers by using the installp
command or SMIT.
Example (FC 6228)
Use the lslpp
command to display the drivers currently installed on the
system.
# lslpp -l | grep df1000f9
Check the list for the required two drivers:
For the IBM FC 6227 HBA, the following drivers are required:
devices.pci.df1000f9
devices.fcp.disk
If the drivers are displayed, you do not need to install the drivers.
If the drivers are not displayed, install the drivers by using the installp
command or SMIT.
To check the firmware level:
The HBA should have the proper version of firmware installed.
Use the lsdev
command to display the device object.
# lsdev -Cc adapter
Use the lscfg
command to display the firmware level.
# lscgfg -vl fcsX
fcsX is the fiber devids object (typically fcs0). The field “Devices
Specific.(Z9)�? shows the installed firmware revision of the HBA.
To install the drivers using the AIX command line:
Insert the IBM drivers CD.
Use the installp
command to install the drivers.
Example
# installp -a -dlpfc.installp all
Use the lslpp
command to verify that the drivers are installed on the
system.
Configure the new devices by rebooting the system with the
shutdown -r
command or using the cfgmgr
command to run
Configuration Manager.
To install the drivers using SMIT:
Insert the IBM drivers CD.
Start SMIT
.
Example
# smit
The System Management
screen appears.
Select Install Additional Software
.
Select Input Device / Directory For Software
and press F4
to select
device.
Select Software To Install
and press F4
to display a list of software.
Select the drivers needed using F7
.
You can use the slash (“/�?) to search for the components in the list.
HBA FC 6227 requires these drivers:
devices.pci.df1000f7
devices.fcp.disk
Press Enter
.
Wait for status to change from RUNNING
to OK
.
Check installation summary result (SUCCESS).
Press F10
.
Use the smit devinst
command to configure the devices.
Clustering and Fabric zoning
If you plan to use clustering, install and configure the clustering software
on the servers. Clustering is the organization of multiple servers into
groups. Within a cluster, each server is a node. Multiple clusters compose a
multi-cluster environment. The following example shows a multi-cluster
environment with three clusters, each containing two nodes. The nodes
share access to the disk array. Figure 4: Multi-cluster
environment with three clusters
Within the Storage Area Network (SAN), the clusters may be homogeneous
(all the same operating system) or they may be heterogeneous (mixed
operating systems). How you configure LUN Security and fabric zoning
depends on the operating system mix and the SAN configuration.
Fabric zoning and LUN security for multiple operating systems
By using appropriate zoning and LUN security, you can connect multiple
clusters of various operating systems to the same switch and fabric: Host zones must contain only homogeneous operating systems.
Storage port zones may overlap if more than one operating system
needs to share an array port.
Heterogeneous operating systems may share an XP array port if you
use Secure Manager and set the appropriate host group and mode; all
others must connect to a dedicated XP array port.
Use Secure Manager for LUN isolation when multiple hosts connect
through a shared array port. Secure Manager provides LUN security
by allowing you to restrict which LUNs each host can access.
| Environment
|
OS Mix
|
Fabric Zoning
|
LUN Security
|
|---|
| Standalone SAN
(non-clustered)
|
homogeneous (a single OS
type present in the SAN)
|
Not required |
Must be used when multiple
hosts connect through a
shared port |
| heterogeneous (more than one
OS type present in the SAN)
|
Required |
| Clustered SAN |
homogeneous (a single OS
type present in the SAN)
|
Not required |
Must be used when multiple
cluster nodes connect through
a shared port |
| heterogeneous (more than one
OS type present in the SAN)
|
Required |
| Multi-Cluster
SAN
|
homogeneous (a single OS
type present in the SAN)
|
Not required |
Must be used when multiple
cluster nodes connect through
a shared port |
| heterogeneous (more than one
OS type present in the SAN)
|
Required |
Connect the disk array
Connect the disk array to the host as follows:
The HP service representative verifies operational status of the disk
array channel adapters, LDEVs, and paths.
The HP representative connects the Fibre Channel cables between the
disk array and the host.
Verify the ready status of the disk array and peripherals.
Defining the paths
Use Command View (shown) or LUN Configuration Manager to create
paths (LUNs) between hosts and volumes in the disk array, also called LUN
mapping. LUN mapping includes these tasks: Configuring ports
Setting LUN security
Creating host groups by operating system and setting their host modes
Assigning host bus adapter WWNs to host groups.
Mapping volumes to host groups (by assigning LUNs).
For details, see the Command View
or LUN Configuration Manager guide
.
HP recommends that you note LUNS and their ports, WWNs, nicknames,
and LDEVs for later use in verifying host and device configuration.
Verifying disk array device recognition
Log into the host as an administrator (root).
If the disk array LUNs are defined after the IBM system is powered on,
issue a cfgmgr
command to recognize the new devices.
Use the lsdev
command to display system device data and verify that
the system recognizes the newly installed devices.
# lsdev -Cc disk
The devices are listed by device file name. All new devices should be
listed as “Available.�? If they are listed as “Define,�? you must do more
configuration before they can be used.
Figure 5: Example
(Fibre Channel)
The example shows that Device hdisk0 is installed on bus 60 and has
TID=5 and LUN=0.
Record the device file names for the new devices. You will use this
information in changing the device parameters.
Use the lscfg
command to find out the AIX disk device’s corresponding
array LDEV designation.
Example
# lscfg -vl hdisk3
In this example, the emulation type, LDEV number, CU number and
array port designation should all be displayed for disk device hdisk3.
Configure disk array devices
Configure the disk array devices in much the same way you would
configure any new disk on the host. Creating scripts to configure all devices
at once may save you considerable time.
Changing the device parameters
When the device files are created, the system sets the device parameters to
the system default values. You may need to change a few of those values
for each new OPEN-x device:
The recommended queue depth settings may not provide the best I/O
performance for your system. You can adjust the queue depth setting to
optimize the I/O performance of the disk array.
| Type
|
Parameter Name
|
Default Value
|
Required Value for Disk Array
|
|---|
| SCSI |
Read/write
time-out |
30 |
60 |
| Queue depth |
1 |
2 (For LUSE devices use 2 for each LUN. For
example, if one LUSE device contains 8 LUNs,
use 2 × 8 = 16 for the queue depth.) |
| Queue type |
None |
Simple |
| Fibre
Channel |
Read/write timeout |
30 |
60 |
| Queue depth |
Before
52-38-xx |
Use 2 if exclusively OPEN-x volumes are
mapped to the SCSI/FC port.
Use 8 if exclusively LUSE volumes are
mapped to the SCSI/FC port.
Use 2 if an intermix of LUSE and
OPEN-x volumes is mapped to the
SCSI/FC port.
Use 8 if an intermix of LUSE and
OPEN-x volumes is mapped for dummy LU
(I-7135-Emu). |
| 52-40-xx to
52-44-xx |
Number of volumes × queue-depth ≤ 256 AND
queue-depth ≤ 8 |
| 52-45-xx or
later |
Number of volumes × queue-depth ≤ 256
and queue-depth ≤ 32 |
| Queue type |
None |
Simple |
To show the device parameters using the AIX command line:
At the command line prompt, enter lsattr -E -l hdiskx
, where hdiskx is the
device file name.
Example
# lsattr -E -l hdisk2
To change the device parameters using the AIX command line:
Change the parameters as follows:
To change the R/W timeout parameter, enter:
chdev -1 hdiskx -a rw_timeout='60'
To change the queue depth parameter, enter:
chdev -l hdiskx -a queue_depth='x'
where x is a value from the above table.
To change the queue type parameter, enter:
chdev -l hdiskx -a q_type='simple'
Example This example changes the queue depth for device hdisk3:
# chdev -l hdisk3 -a queue_depth='2'
Verify that the parameters for all devices were successfully changed.
Example
# lsattr -E -l hdisk3
Repeat these steps for each OPEN-x device on the disk array.
|
NOTE:
|
The lsattr
command also shows other useful information, such as LUN ID
of the mapped LDEV, worldwide name of the disk array FC port, and
N-Port ID. |
Another useful command for determining the slot position and port
worldwide name of the HBA is the lscfg -v -l hdiskx
command.
To change the device parameters using SMIT:
Start SMIT. (Optional) For an ASCII session, use the smit -C
command.
Example # smit
The System Management screen appears.
Select Devices
.
Example Figure 6: The System Management screen
The Devices screen appears.
Select Fixed Disk
.
The Fixed Disk screen appears.
.
Select Change/Show Characteristics of a Disk
.
The Disk screen appears
Select the desired device from the Disk menu.
The Change/Show Characteristics of a Disk screen for that device is
displayed.
Example Figure 7: Change/Show Characteristics of a Disk screen
Enter the correct values for the read/write timeout value, queue depth,
and queue type parameters. Press Enter
to complete the parameter
changes.
Repeat these steps for each OPEN-x device on the disk array.
Assigning the new devices to volume groups
Assign the new devices to volume groups, using the AIX system’s Logical
Volume Manager (accessed from within SMIT). This operation is not
required when the volumes are used as raw devices.
To assign a device to a volume group: Start SMIT. (Optional) For an ASCII session, use the smit -C
command.
Example # smit
The System Management screen appears.
Select System Storage Management (Physical & Logical Storage)
.
Example Figure 8: System Managment screen
Select Logical Volume Manager
.
Select Volume Groups
.
Select Add a Volume Group
.
Enter or select values for the following fields: Volume Group name (the volume group can contain multiple hdisk
devices)
Physical partition size in megabytes, see “Physical partition size table�?
(page 65).
Physical Volume names
To enter values, place the cursor in the field and type the value.
To select values, place the cursor in the field and press F4
.
Enter yes
or no
in the Activate volume group AUTOMATICALLY at
system restart?
field.
If you are not using HACMP (High Availability Cluster
Multi-Processing) or HAGEO (High Availability Geographic), enter
yes
.
If you are using HACMP and/or HAGEO, enter no
.
Press Enter
when you have entered the values. The confirmation screen
appears.
Press Enter
again.
The Command Status screen will appear. To ensure the devices have
been assigned to a volume group, wait for OK
to appear on the
Command Status line. Figure 9: Cammand startus screen
Repeat these steps for each volume group needed.
Creating the Journaled File Systems
Create the Journaled File Systems using the System Manager Information
Tool (SMIT). This operation is not required when the volumes are used as
raw devices. The largest file system permitted in AIX is 64 GB.
To create the Journaled File Systems:
Start SMIT.
# smit -C
Select System Storage Management (Physical & Logical Storage)
.
Select File Systems
.
Select Add/Change/Show/Delete File Systems
.
Select Journaled File Systems
Select Add a Journaled File System
.
Select Add a Standard Journaled File System
.
Select a volume group, and press Enter
.
Enter values for the following four fields:
SIZE of file system (in 512-byte blocks)
: Enter the lsvg command to
display the number of free physical partitions and physical partition
size. Calculate the maximum size of the file system as follows:
(FREE PPs - 1) x (PP SIZE) x 2048
Mount Point
: Enter mount point name. (Make a list of the mount
point names for reference.)
Mount AUTOMATICALLY at system restart?
: Enter yes
. | CAUTION:
|
In high availability systems (HACMP and/or HAGEO), enter no. |
Number of bytes per inode
: Enter the number of bytes appropriate
for the application, or use the default value.
Press Enter
to create the Journaled File System.
The Command Status screen appears.
To ensure that the Journaled File System has been created, wait for OK
to appear on the Command Status line.
To continue creating Journaled File Systems, press the F3
key until you
return to the Add a Journaled File System
screen.
Repeat steps b
through k for each Journaled File System to be created.
To exit SMIT, press the F10
key.
Mounting and verifying the file systems.
Mount the file systems and verify that the file systems were created
correctly and are functioning properly.
- To mount and verify the file systems:
Mount the file system. Enter:
mount mount_point_name
Example # mount /vg01
Repeat step a for each new file system.
Use the df
command to verify the size of the file systems. The capacity
is listed in 512-byte blocks. To list capacity in 1024-byte blocks, use the
df -k
command. Figure 10: df command screen output
Verify that the file system is usable by performing some basic operations
(for example, file creation, copying, and deletion) on each logical
device.
Use the df
command to verify that the file systems have successfully
automounted after a reboot. Any file systems that were not automounted
can be set to automount using the SMIT Change a Journaled File
System
screen.
If you are using HACMP or HAGEO, do not set the file systems to
automount. |
NOTE:
|
HACMP and HAGEO do not provide a complete disaster recovery or
backup solution, and are not a replacement for standard disaster recovery
planning and backup/recovery methodology. |
|
|
Printable version
