This chapter describes the two Storage Array management methods and the procedures necessary to prepare the SGI InfiniteStorage TP9300, SGI InfiniteStorage TP9400 and SGI InfiniteStorage TP9500 for installing the TPSSM software.
The installation process involves the following:
Deciding on either a host-managed or direct managed Storage Array
(“Storage Array Management Methods”).Preparing for the installation (“Preparing for a Network Installation”).
Installing the TPSSM software (“TPSSM Software Installation Procedure” in Chapter 3 and “Completing the Installation” in Chapter 3).
The storage management software provides two methods for managing Storage Arrays:
Host-managed via agent and client over Fibre Channel path(s)
Direct-managed via network (Ethernet) connections
Depending upon your specific Storage Array configurations, you can use either or both methods.
With this method, you manage Storage Arrays through a Fibre Channel connection to a host. The host-agent receives communication from the storage management software and passes it to the Storage Array controllers via a Fibre Channel I/O path.
Figure 2-1 shows a system in which Storage Arrays are managed through the host-agent.
The advantages of using the host-agent managed method include the following:
Ethernet cables are not required to connect each controller to the network.
Network configuration tasks for each controller are not required.
Only a host name or IP address for the host must be specified when you add Storage Arrays (rather than the information for each controller in a Storage Array). After you have added a specific name or IP address of a host, the host-agent software will automatically discover any Storage Arrays attached to that host.
The disadvantages of using the host-agent managed method include the following:
With this method, the Storage Array is managed directly over the network through each controller's Ethernet connection. To do this, you define each controller's IP address and host name, and attach a cable to each Ethernet connection on the controller enclosure.
The advantages of managing Storage Arrays directly include:
You can use an IRIX management station to manage Storage Arrays connected to a host with an operating system other than that supported by the storage management software. Contact your customer support representative for more information.
To manage the array, you can use an IRIX workstation (or Windows NT 4.0 or Windows 2000) to connect to an IRIX host.
The disadvantages of using the direct managed method include:
Each controller requires an Ethernet cable for connection (two per controller enclosure).
You will need to specify an IP address or host name for each controller whenever you add Storage Arrays.
You will need to complete several network preparation tasks (see “Preparing for a Network Installation” for a summary of the preparation tasks).
You are limited to one less logical unit number (LUN) than the maximum allowed by your operating system and host adapter. The host-agent requires a special Access Volume to communicate with the Storage Array controllers. The Access Volume uses LUN 31 as the default.
Figure 2-2 shows a system in which Storage Arrays are managed directly.
You can manage Storage Arrays over the network using either or both of the Storage Array management methods described in “Storage Array Management Methods”.
Decide how you will manage your Storage Arrays. (See Figure 2-3 for an example of a network that uses both types of management methods.)
Network A — Directly Managed Storage Array installation with the following components:
Network B — H ost-Agent Managed Storage Array installation with the following components:
Host connected to a Storage Array via a Fibre Channel I/O path
Management station connected through the Ethernet network to the host to manage the Storage Array controllers
Continue to “Preparing for a Network Installation”.
To ensure that you have installed and checked all network components and that you have obtained and recorded all required host and controller information (see the blank information record provided in Table 2-3), you must complete all of the applicable installation preparation tasks listed in Table 2-1. An example of a completed information record is provided in Table 2-2.
 | Note: Actual installation of network components can be performed using either of the following methods: “Network Setup Using the RS-232 Interface” (also see task 4 or task 5 in Table 2-1). |
Table 2-1. Summary of Network Preparation Tasks
Management Method | Task | Purpose | Go To: |
|---|---|---|---|
Direct and host-agent | 1. Install Ethernet connection. | To establish the Ethernet connection. | N/A |
Direct and host-agent | 2. Establish and record a naming scheme for your Storage Arrays. | To add the Storage Arrays to the management domain after installing the software. |
. |
Direct and host-agent | 3. Obtain IP addresses and host names from the network administrator. | Host-agent management requires host
IP addresses and host names. | “Obtaining IP Addresses and Host Names” . |
Direct | 4. Connect a VT100 terminal to the controller[a]. | To allow the a user to use the VT100 terminal to provide network configuration information to the controllers. | “Network Setup Using the RS-232 Interface” . |
Direct and host-agent | 5. Set up the host (or DNS) table. | To ensure that the management station can communicate with the host running host-agent software or with the controllers. | “Setting Up the DNS or Host Table” . |
Direct and host-agent | 6. Power up network devices | To ensure that all devices and links are operational. | The Installation Guide specific to each hardware component. |
[a] Use either task 5 or task 6 to provide network configuration information to the controllers. | |||
Table 2-2 shows a sample information record with entries for a directly managed Storage Array and a Host-Agent Managed Storage Array.
Table 2-2. Sample Information Record
Storage Array Name | Management Type | Controller B— | Host— | |
|---|---|---|---|---|
Storage Array Name: | Direct | Ethernet Address: IP Address: Host Name: | Ethernet Address: IP Address: Host Name: |
|
Storage Array Name: | Host-Agent |
|
| IP Address: Host Name: |
Table 2-3 provides a blank information record. Photocopy the blank information record and complete the information for each network Storage Array and controller. This will help you correctly add Storage Arrays after initial installation.
Table 2-3. Storage Array and Controller Information Record
Storage Array Name | Management Type (choose one) | Controller B— | Host— | |
|---|---|---|---|---|
Name: | ___ Direct ___ Host-Agent | Ethernet Address:
IP Address:
Host Name: | Ethernet Address:
IP Address:
Host Name: | IP Address:
Host Name: |
Name: | ___ Direct ___ Host-Agent | Ethernet Address:
IP Address:
Host Name: | Ethernet Address:
IP Address:
Host Name: | IP Address:
Host Name: |
Name: | ___ Direct ___ Host-Agent | Ethernet Address:
IP Address:
Host Name: | Ethernet Address:
IP Address:
Host Name: | IP Address:
Host Name: |
Name: | ___ Direct ___ Host-Agent | Ethernet Address:
IP Address:
Host Name: | Ethernet Address:
IP Address:
Host Name: | IP Address:
Host Name: |
When planning your network configuration, consider how you will name the Storage Arrays. When you start the storage management software for the first time, all Storage Arrays in the management domain will be displayed as <unnamed>. You then can use the Array Management window to rename individual Storage Arrays.
Use the following procedure to name Storage Arrays.
Decide on a Storage Array naming scheme that makes sense for your enterprise. Here are some tips on naming Storage Arrays:
The software allows a 30-character limit. All leading and trailing spaces will be deleted.
Use a unique, meaningful naming scheme that is easy to understand and remember.
Avoid arbitrary names or names that would quickly lose their meaning in the future.
The software displays Storage Array names with the prefix Storage Array. Therefore, if you rename a Storage Array Engineering, it will display as Storage Array Engineering.
Record the Storage Array names in the information record (Table 2-3).
Depending on which management method or methods you will use, you might need to record your controller's hardware Ethernet addresses.
Do you plan to manage your Storage Arrays directly or with the host-agent software?
Direct Managed | Go to “Identifying the Controller Hardware Ethernet Address” . |
Host-Agent Managed | Go to “Obtaining IP Addresses and Host Names” . |
Use the following procedure if you plan to directly manage Storage Arrays through Ethernet connections to each controller.
Remove the front panel from the controller tray (Figure 2-4 or Figure 2-5).
Look on the front of the controller canister for a label with the controller's hardware Ethernet address. The number will be in hexadecimal form (for example, 00.a0.b8.00.00.d8).Figure 2-6 shows the location of the controller's hardware Ethernet adress on the SGI InfiniteStorage TP9500 RAID Controller.
Figure 2-6. Identifying the Controller Ethernet Address on the SGI InfiniteStorage TP9500 RAID Controller
Figure 2-7 shows the location of the hardware Ethernet address label on the SGI InfiniteStorage TP9300 RAID Controller.
Figure 2-7. Identifying the Controller Ethernet address on the SGI InfiniteStorage TP9300 RAID Controller
Record each Ethernet address in the information record (Table 2-3).
Depending on which management method or methods you will use, perform the following procedures to obtain IP addresses and host names. Within your enterprise, you might choose to manage some Storage Arrays directly and others with the host-agent software.
Direct Management Method | Host-Agent Management Method |
|---|---|
1. Assign (or obtain from the network administrator) a unique IP address and associated host name for each Storage Array's controllers that you will manage directly. Record each controller IP address and host name in the information record (Table 2-3 ). . | 1. Assign (or obtain from the network administrator) a unique IP address and associated h ost name of each host that you will use to manage the Storage Array via the host-agent method. Record the IP address and host name in the information record (Table 2-3 ). . |
The SGI InfiniteStorage TP9300, SGI InfiniteStorage TP9400, and SGI InfiniteStorage TP9500 allow you to set up the network interface through the RS-232 port on each controller. To interface to the RS-232 port, you need:
A PC or device that can emulate a VT100 console.
A 9-pin null modem cable with female connectors to attach to the RS-232 port on the controller.
On IRIX systems, perform the following procedure:
Install the eoe.sw.uucp software subsystem.
Using a text editor, modify /ect/uucp/Devices to access the desired serial port. For example, change:
# Direct ttyd2 - 9600 direct
to:
Direct ttyd2 - 9600 direct
Use the cu(1c) command to access the serial port, as follows:
# cu -lttyd2
Enter -b% to transmit a CTRL-BREAK.
On Windows NT systems, perform the following procedure:
Install the version of Hyperterminal that supports the transmission of a CTRL-BREAK.
Select Start >> Programs >> Accessories >> Hyperterminal.
If this is the first time that HyperTerminal has been started, the Location Information dialog is displayed:
Select Cancel.
A Confirmation dialog is displayed.Select Yes.
An Informational dialog is displayed. Select OK.
The New Connection Dialog is displayed.
If this is not the firt time that HyperTerminal has been started, the New Connection dialog is displayed.
Enter the desired name in in the Name: text box and select OK. The Connect to dialog is displayed.
From the Connect using: spinner box, choose the COM port you will use to connect to the RAID controller serial port. Select OK. The Properties dialog is displayed.
From the Bits per second: spinner box, choose the baud rate youwill use to send/receive data (for example, 9600). Select OK.
At this point, the configuration setup has been completed and you are ready to send and receive data. Before shutting down this application, save your settings so you will not have to repeat this procedure.
On Windows 2000 systems, perform the following procedure:
Select Start >> Programs >> Accessories >>Communications>> Hyperterminal.
If this is the first time that HyperTerminal has been started, the Location Information dialog is displayed:
Select Cancel.
A Confirmation dialog is displayed.Select Yes.
An Informational dialog is displayed. Select OK.
The New Connection Dialog is displayed.
If this is not the firt time that HyperTerminal has been started, the New Connection dialog is displayed.
Enter the desired name in in the Name: text box and select OK. The Connect to dialog is displayed.
From the Connect using: spinner box, choose the COM port you will use to connect to the RAID controller serial port. Select OK. The Properties dialog is displayed.
From the Bits per second: spinner box, choose the baud rate youwill use to send/receive data (for example, 9600). Select OK.
At this point, the configuration setup has been completed and you are ready to send and receive data. Before shutting down this application, save your settings so you will not have to repeat this procedure.
After you have connected the cable and started the VT100 emulator, you must establish communication to the controller following these steps:
Enter a Ctrl+break key sequence.
When prompted, press the space bar to set the transfer baud rate.
Enter another Ctrl+break sequence.
Press Esc to get the login prompt.
Enter the password of infiniti.
Enter the following:
> netCfgSet
A list of network configuration options appear.
Set the configuration values as follows:
Network Configuration: Ethernet My MAC Address:00:a0:b8:06:6d:35 (Do not change) My Host Name:RAID5 (sets the name of the contoller) My IP Address:163.154.14.155 (sets the IP address of the contoller) Server Host Name:host (do not change) Server IP Address:0.0.0.0 (do not change) Gateway IP Address:163.154.12.2 (sets the gateway IP address) Subnet Mask:255.255.255.0 (sets the Subnet Mask) User Name:guest (do not change) User Password: (do not change) NFS Root Path (do not change) NFS Group ID Number:0 (do not change) NFS User ID Number:0 (do not change) value = 0 = 0x0
After these values are set, they are written to the controller's non-volatile memory (NVSRAM). You must now reboot the controller so that the changes will go into effect.
Enter the following:
> sysReboot
The controller now reboots itself.
Verify the changes by logging back into the controller and entering the following command:
netCfgShow
Verify that the following is displayed:
==== NETWORK CONFIGURATION: ALL INTERFACES ==== Network Init Flags : 0x00 Network Mgmt Timeout : 30 Startup Script : Shell Password : ==== NETWORK CONFIGURATION: Ethernet ==== Interface Name : dse0 My MAC Address : 00:a0:b8:06:6d:35 My Host Name : RAID5 My IP Address : 163.154.14.155 Server Host Name : host Server IP Address : 0.0.0.0 Gateway IP Address : 163.154.12.2 Subnet Mask : 255.255.255.0 User Name : guest User Password : NFS Root Path : NFS Group ID Number : 0 NFS User ID Number : 0 value = 0 = 0x0
Repeat steps 1–10 for each controller.
Use the following procedure to set up the Domain Name Server (DNS) or host table. Make sure the controller host names correspond to the appropriate controller IP addresses.
Edit either the DNS or the host table (/etc/hosts) to add the IP address and host name for each network controller.
For example, to set up the host table for the Network A controllers, (Figure 2-3), use a text editor to create the following controller IP address and name entries
Table 2-5. IP Address and Controller Host Name
IP Address
Controller Host Name
127.0.0.01
localhost
192.168.1.13
Denver_a
192.168.1.14
Denver_b
If you plan to manage Storage Arrays through a firewall, configure your firewall to open port 2463 to TCP data.