C H A P T E R  3

Installing the Software by Using the nhinstall Tool

The nhinstall tool installs the Solaris Operating System and the Foundation Services on the cluster according to the options that are set in the configuration files.

To use nhinstall to install the software, see the following sections:


Stages of the Installation Process

You can configure the nhinstall tool to install both the Solaris Operating System and the Foundation Services on the cluster. You can also install the Solaris Operating System manually and then configure the nhinstall tool to install only the Foundation Services on the cluster. For more information, see Configuring the nhinstall Tool.

After you have configured the installation server, you are ready to install the software on the nodes of the cluster. TABLE 3-1 describes the stages of the installation process in an example scenario. In this scenario, the nhinstall tool is configured with the following parameters defined in the env_installation.conf file:

The following parameters are defined in the cluster_definition.conf file:

Diskless and dataless nodes are specified in the NODE parameter or by the DISKLESS and DATALESS parameters.

For information about installation parameters, see the env_installation.conf(4) and cluster_definition.conf(4) man pages.

The following table lists the stages of an automatic installation using the nhinstall tool. The times quoted are a guideline only.


TABLE 3-1 Stages of the Installation Process

Stage

Action Carried Out By

Related Section

1.

Launch the nhinstall tool.

You

To Launch the nhinstall Tool

2.

Configure the network interfaces.

nhinstall

3.

Prepare the Solaris JumpStart environment for the master-eligible nodes and the dataless nodes (5 minutes).

nhinstall

4.

Boot the master-eligible nodes and the dataless nodes to install the Solaris Operating System (30 minutes - 1 hour).

You

To Boot the Master-Eligible Nodes and Dataless Nodes to Install the Solaris Operating System

5.

Install the Foundation Services and the CGTP patches on the master-eligible nodes (20 minutes).

nhinstall

6.

Reboot the master-eligible nodes.

nhinstall

To Boot the Master-Eligible Nodes

7.

Configure the Solaris services for diskless nodes on the master-eligible nodes (20 - 45 minutes).

nhinstall

8.

Install the Solaris packages, the Foundation Services packages, and the CGTP patches for diskless nodes (5 minutes) and the Foundation Services packages, and the CGTP patches for dataless nodes.

nhinstall

9.

Reboot the master-eligible and dataless nodes.

You

To Boot the Master-Eligible Nodes

10.

Boot the diskless nodes.

You

To Boot the Diskless Nodes

11.

Run the nhadm tool on the master-eligible nodes to ensure that installation was successful.

You

Verifying the Installation




Note - The time guidelines are estimates only. The actual times that are required depend on the type of installation server, your configuration options, and the number of nodes in your cluster.




Launching the nhinstall Tool

Before running the nhinstall tool, make sure that you have configured the nhinstall tool and prepared the installation environment as described in
Chapter 2.

The following procedures are based on Stages of the Installation Process. These procedures must be carried out in the documented sequence.


procedure icon  To Launch the nhinstall Tool

1. Log in to the installation server as superuser.

2. Start the nhinstall tool on the installation server:

# cd /opt/SUNWcgha/sbin
# ./nhinstall -r config-file-directory -l logfile

For details, see the nhinstall(1M) man page.

If you have chosen a two-node cluster configuration in the cluster_definition.conf file, you are prompted to install the Solaris services for diskless nodes:


Do you want to install the Solaris services for
diskless nodes anyway [y/n]:

If you type y, the nhinstall tool installs the Solaris services for diskless nodes, which gives you the option of adding diskless nodes to the cluster at a later stage.

After you have made this choice, the nhinstall tool configures the network interfaces, prepares the master-eligible nodes, and prepares the Solaris JumpStart environment. This process takes approximately 5 minutes.



Note - If the Solaris JumpStart procedure stops, the utility might not restart because the /tmp/.install_client.lck file prevents two Solaris JumpStart commands from being executed simultaneously. If Solaris JumpStart stops, make sure that there are no other Solaris JumpStart processes running. Then delete the lock file.



After the Solaris JumpStart environment has been prepared, the nhinstall tool displays instructions. Follow these instructions by performing the following procedure.


procedure icon  To Boot the Master-Eligible Nodes and Dataless Nodes to Install the Solaris Operating System

Perform the following steps on both master eligible nodes and on each dataless node listed in the nhinstall output.

1. At the console window of the node, get the ok prompt.

    a. To get the telnet prompt, press Control-].

    b. To get the ok prompt, type send brk at the telnet prompt.

    # Control-]
    
    telnet> send brk
    
    Type  'go' to resume
    
    ok>
    

The first master eligible-node is the first master-eligible node that is defined in the cluster_definition.conf file with the NODE definition or with the MEN parameter. For details, see Configuring the nhinstall Tool.

2. Set the boot variables by using the setenv command.

    a. Configure the processors to use local Ethernet (MAC) addresses:

    ok> setenv local-mac-address? true
    

    b. Configure the processors to retry booting when autoboot fails:

    ok> setenv auto-boot-retry? true
    

3. Make a note of the network device aliases:

ok> devalias

Note the device alias of the network adapter of the master-eligible node that is connected to the same switch as the installation server's network interface. If there is no device alias, define one using the nvalias command. In the case of Netra 20 servers, the network adapter used is not the native adapter. In this case, you need to define the device alias.

For more information, see the OpenBoot 4.x Command Reference Manual.

4. Start the installation of the Solaris Operating System on the node:

ok> boot net - install

where net is the device alias.

5. When you have started the installation of the Solaris Operating System on the node, continue the installation by typing y at this prompt:

Answer 'y' after the command has been entered on all nodes: y

The installation takes approximately 30 minutes for the master-eligible nodes and dataless nodes. The progress of the installation and the packages that are being installed are displayed on the console window of each master-eligible node.

When the Solaris packages have been installed, each node is automatically rebooted. Several status messages are displayed regarding service startup and connectivity.

The nhinstall tool checks which version of the Solaris Operating System has been installed. The nhinstall tool also checks whether the Foundation Services are already installed on the master-eligible nodes. If the Foundation Services are already installed, the nhinstall tool exits and displays an error message. If you want to upgrade to the new version of the Foundation Services, see Chapter 9 through Chapter 12. Otherwise, remove the existing Foundation Services packages and restart the installation as described in Resetting the nhinstall Tool for a New Installation.

The nhinstall tool then installs the Foundation Services and the CGTP patches on the master-eligible nodes.

Wait 15 to 20 minutes while the Foundation Services packages and the CGTP patches are installed on all nodes. After installing packages and patches on the master-eligible and dataless nodes, these nodes are rebooted and nhinstall waits for the data to be ready.


procedure icon  To Boot the Diskless Nodes

After installation, the nhinstall tool displays the following instructions:


The software installation is complete.
Setup the eeprom boot parameters on your diskless nodes:
At the ok prompt, type:
  ok> setenv local-mac-address? true
  ok> setenv auto-boot-retry? true
  ok> setenv boot-device net:dhcp,,,,,5 net2:dhcp,,,,,5
  ok> setenv diag-switch? false
You can now boot your diskless nodes.

To configure the boot parameters, perform the following procedure.

1. Get the ok prompt on each diskless node.

2. Execute the following commands at the ok prompt on each diskless node:

ok> setenv local-mac-address? true
ok> setenv auto-boot-retry? true
ok> setenv boot-device net:dhcp,,,,,5 net2:dhcp,,,,,5
ok> setenv diag-switch? false



Note - If you are going to use client_id on a Netra CT diskless node, set the Boot_Devices environment variable. For more information, see the Netra CT Server System Administration Guide.



3. Boot each diskless node:

ok> reset


Verifying the Installation

After you have completed the installation and configuration, check that the cluster nodes have connectivity.


procedure icon  To Verify the Installation

1. Become superuser on all nodes.

2. Run the nhadm tool:

# /opt/SUNWcgha/sbin/nhadm check starting

If all checks pass the validation, the installation of the Foundation Services on the nodes of your cluster has been successful. For more information, see the nhadm(1M) and the cluster_nodes_table(4) man pages.


Troubleshooting and Restarting the nhinstall Tool

Most warnings displayed by the nhinstall tool do not require you to take any action.

When the nhinstall tool is launched, the tool parses the configuration files. If the nhinstall tool encounters errors in the files, it exits and provides a list of error messages. A progress indicator monitors the progress of the installation so you can continue the installation from the point of failure when you restart the nhinstall tool.

If an error occurs during the installation process, fix the error and then restart the nhinstall tool:


# cd /opt/SUNWcgha/sbin
# ./nhinstall -r config_file_directory -l logfile



Note - If you modify the configuration files to correct an error, the nhinstall tool displays a warning that the configuration has changed. The tool then prompts you to reset the installation. For information, see Resetting the nhinstall Tool for a New Installation.




Resetting the nhinstall Tool for a New Installation

If you are installing a new version of the software on the cluster or a new cluster configuration, you must start the installation from the beginning. You can reset the nhinstall tool in one of two ways:


procedure icon  To Remove the Progress Indicator and Reset the nhinstall Tool

To reset the nhinstall tool by removing the progress indicator, perform the following steps:

1. Log in to the installation server as superuser.

2. Reset the nhinstall tool and remove the progress indicator:

# cd /opt/SUNWcgha/sbin
# ./nhinstall -r config_file_directory reset

The next time that you launch the nhinstall tool, the installation starts from the beginning.


procedure icon  To Clear the Installation Environment and Reset the nhinstall Tool

To reset the nhinstall tool and clear the installation directories and files, perform the following steps:

1. Log in to the installation server as superuser.

2. Unshare all shared and exported directories, and remove the temporary files, including the progress indicator:

# cd /opt/SUNWcgha/sbin
# ./nhinstall -r config_file_directory clear

The shared and exported directories are no longer shared or exported. All temporary files are deleted. The installation environment is cleared. The next time that you launch the nhinstall tool, the installation starts from the beginning.