1 Installing from a Remote Installation Services (RIS) Server

This chapter contains the following information:

The prerequisites for installing from a RIS server (Section 1.1)

A simple test to make sure the network is configured before you begin (Section 1.2)

How to start an Update Installation from a RIS server (Section 1.3)

How to start a Full Installation from a RIS server (Section 1.4)

1.1 Prerequisites to Installing from a RIS Server

If your system is located within a networked environment, you may be able to install the operating system from a server on the network if the following items are in place:

Your site has a system that has been configured to be a Remote Installation Services (RIS) server

Your system is a registered client of a RIS server that is serving Version 5.1B of the operating system

Note to RIS Administrators

The RIS server must be running Version 4.0 or later of the operating system to be able to serve Version 5.1B to clients.

See Sharing Software on a Local Area Network if you are a network administrator and need more information about setting up a RIS server, creating software environments to serve, and registering a client system to the right software environment.

1.2 Is Your System Connected to the Network?

If your system already is running a version of the operating system and you want to perform an Update Installation, ensure that your system can communicate with the RIS server by executing the /sbin/ping command to verify the network connection.

To test the network connection, enter the following command and replace ris_server_name with the name of your local RIS server:

# /sbin/ping -c2 ris_server_name

Successful output shows a 0 (zero) percent packet loss, which indicates that your system can communicate with the RIS server over the network. In Example 1-1, the RIS server name is server1.

Example 1-1: Output of the /sbin/ping Command

# ping -c2 server1
PING server1 (16.59.l24.96): 56 data bytes
64 bytes from 16.59.l24.96: icmp_seq=0 ttl=255 time=1 ms
64 bytes from 16.59.l24.96: icmp_seq=1 ttl=255 time=0 ms
 
----server1 PING Statistics----
2 packets transmitted, 2 packets received, 0% packet loss
round-trip (ms)  min/avg/max = 0/0/1 ms

Your system is not able to communicate with the RIS server if you obtain the following results from the /sbin/ping command:

Only the system name and IP address is displayed in the command output if the name server recognizes the system name but the network connection is not set up.

The message Unknown host is displayed if your system does not recognize the RIS server name.

If you experience any of these results, ask your network administrator to troubleshoot the problem. The Network Administration: Connections and Network Administration: Services manuals contain extensive network troubleshooting information.

1.3 Starting an Update Installation from a RIS Server

Follow this procedure to start an Update Installation from a RIS server:

Ensure that your system meets the prerequisites to installing over the network as described in Section 1.1 and the system is configured on the network as described in Section 1.2. See your network administrator or Sharing Software on a Local Area Network if you encounter problems.

Complete all Update Installation prerequisite tasks that are described in the Installation Guide. These tasks include backing up your current operating system, reading the Release Notes, and updating system firmware.

As superuser or root, place the system in single-user mode by using the shutdown command.
- The following example shows how to switch to superuser and then shut down the system to single-user mode:
```
% su - 
password: 
# shutdown +10 Please log out
```
  In the previous example, +10 shuts down the system in ten minutes and sends the message Please log out to all logged in users.
After your system is in single-user mode, the screen looks similar to the following:
```
.
.
.
Halting processes ...
 
INIT: SINGLE-USER MODE
#
```

Mount the local file systems:
```
# /sbin/bcheckrc
```
The bcheckrc command invokes the mount -a command and mounts all file systems in the /etc/fstab file, not just the standard UNIX file systems (/, usr, and var). The bcheckrc command also runs the file system check command, fsck, on UNIX file systems (UFS) and starts the Logical Storage Manager (LSM) if necessary. If fsck finds a problem with the / (root) partition, the system shuts down and requires a reboot to fix the file system.

Delete the table of Internet addresses to ensure that the routed and gated daemons do not start up during the Update Installation:
```
# route flush
```

Enter the /sbin/installupdate command with the following syntax:
/sbin/installupdate [-u] [-nogui] {ris_server_name:}

The following describes each option:
- The optional -u flag runs the Update Installation without any user intervention. This option automatically removes blocking layered products, deletes obsolete files, and installs all kernel components thereby eliminating the need for you to respond to any questions during the update.
- The optional -nogui (no graphical user interface) flag invokes the text-based interface on systems with graphics capabilities.
- The RIS server name must be appended with a colon ( : ).
For example, to start an Update Installation from a RIS server named server1, enter the following command:
```
# /sbin/installupdate server1:
```
As another example, you would enter the following command to run the Update Installation in unattended mode using the text-based interface from a RIS server named server2:
```
# /sbin/installupdate -u -nogui server2:
```
Note

If you have problems starting the Update from a RIS server, see the Sharing Software on a Local Area Network manual for troubleshooting information.

After the Update has started, you can follow along with the remainder of the Update Installation process as documented in the Installation Guide.

1.4 Starting a Full Installation from a RIS Server

Follow this procedure to boot your system over the network to start a Full Installation from a RIS server:

Ensure that your system meets the prerequisites to installing over the network as described in Section 1.1 and the system is configured on the network as described in Section 1.2. See your network administrator or Sharing Software on a Local Area Network if you encounter problems.

Complete all Full Installation prerequisite tasks that are described in the Installation Guide. These tasks are in a self-contained chapter and include backing up your current operating system, reading the Release Notes, and updating system firmware.

Bring your system down to console mode (the >>> prompt). Do one of the following depending upon the current state of your system:
- If your system is up and already running a version of the operating system, shut down and halt the processor using a command similar to the following :
```
# shutdown -h +10 Please log out
```
  In the previous example, the system is shut down and halted in 10 minutes and sends the message Please log out to all logged in users. See the System Administration manual and shutdown(8) if you need more information about shutting down a system to console mode.
- If the system is turned off, turn on the processor. The console subsystem prints various start-up and diagnostic messages and ends with the console mode prompt ( >>> ). However, if your system is configured for automatic reboots, it will not stop at the console mode but will boot to multiuser mode. To bring the system to the console mode, use the shutdown -h command as shown in the previous bullet list item.

Certain processors require one or more console environment variables to be set in addition to the standard console variables that are set for all processors. Look for the corporate logo on the front panel of your hardware to determine what kind of processor you have. Next, go to the Full Installation Procedures chapter in the Installation Guide to find out if your system requires special console variables to be set before the system is booted over the network. Then, continue with Step 5 in this procedure.

Clear and reset the boot_osflags console variable:
```
>>> set boot_osflags ""
```

To ensure that your system returns to the console mode in the event of a system crash or power failure during the installation, set the auto_action console variable:
```
>>> set auto_action halt
^{[Footnote 1]}
 
```

Determine the network adapter device name:
```
>>> show device
```
A device information table similar to the following is displayed:
```
dka400.4.0.6.0          DKA400      RRD43  2893
dva0.0.0.0.1            DVA0
ewa0.0.0.13.0           EWA0        08-00-2B-3E-B6-C8
pka0.7.0.6.0            PKA0        SCSI Bus ID 7
```
The network boot device is shown in the middle column next to the hardware Ethernet address in the third column. In this example, the hardware Ethernet address is 08-00-2B-3E-B6-C8, and the boot device is EWA0.

Initialize the bootp boot request protocol using the following command syntax:
set network-device_protocolsbootp

set network-device_inet_initbootp

The network boot device name is case insensitive and can be entered in lower case or upper case letters. Using the network adapter device name obtained in Step 7 as an example, enter the following bootp initialization commands:
```
>>> set ewa0_protocols bootp
>>> set ewa0_inet_init bootp
 
 
```

Before you boot the system in the next step, check Table 1-1 to make sure there are no special boot commands for your processor type.

Note

Every attempt was made to make the information in Table 1-1 complete and accurate for every supported system. However, it is recommended that you also check the hardware documentation for your processor to make sure it does not require any other console variables or boot flags to be set. Hardware documentation is customized for each system type, and it is the definitive source for required console variables and boot flags.

Table 1-1: Processor-Specific Network Boot Commands

Processor or Option	Boot Commands
AlphaStation 500, 600, 600A AlphaServer 800, 1000, 1000A Personal Workstation au Series Ultimate Workstation au Series Professional Workstation XP1000	>>> boot -fl "" ewa
AlphaServer 4000, 4100 Alphaserver 8200, 8400 ES40, GS60, and GS140 Servers	>>> boot -fl "" ewa0
Systems with FDDI devices	See Section 1.4.1

Reset the boot file:
```
>>> set boot_file ""
```

Enter the boot command with the following syntax:
boot network-device
Using the information obtained in Step 7 as an example, the boot command is:
```
>>> boot ewa0
```

This completes booting the system over the network to invoke a Full Installation. The Installation Guide documents the remainder of the Full Installation.

If you encounter trouble booting your system over the network, see the Sharing Software on a Local Area Network manual.

1.4.1 Network Boot Option: Fiber Distributed Data Interface (FDDI)

Booting over an FDDI network interface is supported for DEFTA devices.

To boot from an FDDI network interface device, find the device names by entering the show config command when the system is at console level. For example, a DEFTA device appears as PMAF-FA when you use the show config command. Identify the slot number and enter the boot command as noted in Table 1-2.

Table 1-2 shows the required boot devices for booting over the network if your system uses FDDI. Follow the instructions in your hardware owner's guide to update the FDDI firmware before booting over the network.

Table 1-2: FDDI Boot Devices by Bus Type

Bus Type	Boot Device
EISA (Extended Integrated System Architecture)	`fra0` ^{[Footnote 2]}
PCI (Peripheral Component Interconnect)	`fwa0` ^{[Footnote 2]}
XMI (Extended Memory Interface)	`fxa0` ^{[Footnote 2]}

1.4.2 Troubleshooting Network Boot Failures

If your initial network boot fails, enter init at the console prompt and attempt to boot over the network again. If you encounter other problems during the remote server installation, see the Sharing Software on a Local Area Network manual, which contains troubleshooting information for network boot failures.

1.4.3 Network Reboot Considerations for Systems with Graphics Devices on an ISA Bus

If your system does not reboot automatically after the software subsets are loaded, and your system has a graphics device on its ISA bus that requires a kernel device driver, you must modify the isacfg entry to match the kernel device driver before rebooting the system.

When you issue the following command, let the input line wrap; do not press the [Return ]key in the middle of the line. The backslash character (\) represents line continuation; do not enter a backslash in the command line.

>>> isacfg -mod -slot slot_number -dev device_number    \
    -handle vendor_handle -etyp 1 -enadev 1

In the previous example, replace vendor_handle with the handle supplied in the vendor's installation documentation.

If you performed a RIS installation from a RIS area that already has a kernel graphics device driver built into the generic kernel on the RIS server, and you already set the handle to the one specified in the vendor's installation documentation, you do not need to execute this command. If your system does not support the automatic reboot feature, the boot commands will be displayed on the screen.