1 - C H A P T E R -

C H A P T E R 1

Introduction

Overview

This manual explains how to use the Sun HPC ClusterTools software installation utilities to install, activate, deactivate, and remove Sun HPC ClusterTools software on one or more cluster nodes. You can access these utilities through either a graphical user interface (GUI) or a set of command-line interface (CLI) tools. See TABLE 1-1 for a summary of these interfaces.


Interface	Role
GUI Utility
ctgui	Graphical interface to a set of Java-based wizards, which are used to install, remove, activate, and deactivate the software on cluster nodes.
CLI Utilities
ctinstall	Installs software on cluster nodes.
ctremove	Removes software from cluster nodes.
ctact	Activates software.
ctdeact	Deactivates software.
ctstartd	Starts all Sun HPC ClusterTools software daemons.
ctstopd	Stops all Sun HPC ClusterTools software daemons.
ctnfssvr	Sets up Sun HPC ClusterTools software on an NFS server.

Note - The Sun HPC ClusterTools 5 software installation utilities offer the same capabilities as the Sun HPC ClusterTools 4 software installation tools, including support for NFS client-server configurations and initiating parallel installation on multiple nodes from a central host. However, the version 5 methods differ from version 4 in several important ways.

Sun HPC ClusterTools 5 Software Packaging Note

Sun HPC ClusterTools 5 software packages are organized into two categories to facilitate installing the software in an NFS client/server configuration. The categories are:

Root packages - These packages contain daemon startup scripts, which are placed in/etc/opt of a cluster node. Each cluster node must have the root packages installed locally.

Nonroot packages - These packages contain all other Sun HPC ClusterTools 5 software.

The next section explains how the root and non-root packages are handled in NFS and Non-NFS cluster configurations.

NFS and Non-NFS Configurations

You can install Sun HPC ClusterTools software in an NFS client/server configuration or locally on the cluster nodes.

NFS Configuration

In an NFS-based cluster, the tools install the root packages locally on each cluster node using either ctgui or ctinstall and the non-root packages on the NFS server using the CLI's ctnfssvr command. The non-root packages are remotely mounted on the NFS client nodes in the cluster.

The NFS server can, but need not be one of the NFS client nodes in the cluster. When a Sun HPC ClusterTools software NFS server is also a cluster node, both root and non-root packages are installed on the NFS server by means of their respective installation tools. In other words, the non-root packages are installed on the NFS server with ctnfssvr and the root packages are installed with either ctgui or ctinstall -c.

Non-NFS Configuration

In a non-NFS cluster configuration, the tools install a complete copy of the Sun HPC ClusterTools software, both root and non-root packages, locally on each node in the cluster.

Centralized and Local Command Initiation

You can choose between two methods of initiating operations on the cluster nodes:

Local - Initiate commands directly on the node you are logged in to. The effects of the command are restricted to the local node.

Centralized - Initiate commands from a central host, specifying the nodes on which the command is to take effect. The initiating host establishes remote connections to the target nodes and broadcasts the commands to them over an rsh, ssh, or telnet connection. The central (initiating) host can be part of the cluster or it can be an administrative system external to the cluster.

Note - Centralized operations are performed on the specified nodes in parallel. That is, when a command is specified on the central host, the operation is initiated on all the specified nodes at the same time.

Support for centralized command initiation is built into the Sun HPC ClusterTools software installation utilities. Issuing these commands from a central host has the equivalent effect as invoking the commands locally using one of the Cluster Console tools, cconsole, ctelnet, or crlogin.

Cluster Console is part of the Sun Cluster software distribution. It is also included in the Sun HPC ClusterTools software distribution as an auxiliary convenience.

Note - Although you do not need Cluster Console for centralized installation of Sun HPC ClusterTools software, it can be a convenient tool for various administrative operations, such as booting cluster nodes in a Custom JumpStart installation or modifying the cluster configuration file, hpc.conf.

Centralized command initiation provides a command execution summary for each initiated activity. This summary reports the results of the operation, identifying the nodes on which the operation was successful and the nodes on which it failed. Such reports are not available when using the Cluster Console tools.

The Sun HPC ClusterTools software CLI utilities provide several options that are specific to the centralized command initiation mode and are intended to simplify management of parallel installation of the software from a central host. These options support:

Creating corresponding versions of local log files on the central host for easier access

Generating a list of nodes that had successful operations and another list of nodes that were unsuccessful. These pass/fail node lists can then be used in subsequent operations, such as software activation, deactivation, and removal.

The initiating system can be one of the cluster nodes or it can be external to the cluster. It must be a Sun system running the Solaris 8 or 9 operating environment.

Support for Alternative Sun HPC ClusterTools Software Versions

The Sun HPC ClusterTools 5 software can be installed on systems that have Sun HPC ClusterTools 4 software installed and active. This is referred to as alternative version support. It is equivalent to the Sun HPC ClusterTools 4 software feature known as coexistence, which allowed Sun HPC ClusterTools 4 software to share space on cluster nodes with Sun HPC ClusterTools 3.1 software.

As with coexistence, alternate versions support means that both Sun HPC ClusterTools software versions 5 and 4 can reside on the same cluster nodes, but only one version can be active (available for use) at a given time.

Note - You must use install_gui or the other installation tools supplied with the Sun HPC ClusterTools 4 software distribution to perform operations on Sun HPC ClusterTools 4 software. These tools are available on the Sun HPC ClusterTools 4 software distribution CD or in /opt/SUNWhpc/HPC4.0/bin/ Install_Utilities/bin on an installed system.

Support for Other Installation Contexts

The Sun HPC ClusterTools software installation utilities are completely self-contained and are fully capable of scaling from single node installations to installing very large clusters. However, if you customarily use Custom JumpStart trademark and/or Solaris Web Start Flash methods for installing software on your servers, you can easily integrate the CLI installation tool ctinstall into those contexts.

The following variations on a basic Sun HPC ClusterTools software installation are described briefly below. More detailed descriptions are provided later in the manual.

Custom JumpStart installation

Custom JumpStart installation, using Web Start Flash archives

Custom JumpStart Installation

For this context, you must set up a Custom JumpStart environment in advance of the installation. Then you invoke local installation of Sun HPC ClusterTools software on the cluster nodes by integrating the ctinstall command in the Custom JumpStart finish script, using the -l and -R switches.

Custom JumpStart installations are initiated from the console of each Custom JumpStart client. The cconsole tool that is included in the Cluster Console software allows you to access multiple consoles through a single common window.

Web Start Flash Installations

The first step for this type of installation is to perform a local installation of the Sun HPC ClusterTools software on a node that will serve as the Flash master. You can use either ctgui or ctinstall -l for this step. Once the Flash master is fully installed and activated, you create a flash archive and apply it to the target nodes, usually in a Custom JumpStart environment.

This flash archive-based approach creates clones of the Flash master, which includes reinstalling the Solaris operating environment on each clone.

Note - Web Start Flash installations are restricted to cluster environments where all the systems have identical hardware and software configurations.

Sun HPC ClusterTools 5 Installation Log Files

The Sun HPC ClusterTools 5 installation tools log information about installation-related tasks locally on the nodes where installation tasks are performed. The default location for the log files is /var/sadm/system/logs/hpc. If installation tasks are initiated from a central host, a summary log file is also created on the central host.

Local, Node-Specific Log Files

Two types of log files are created locally on each cluster node where installation operations take place.

Task-specific logs - Separate log files are created for each installation-related task. They are: ct_install.log, ct_remove.log, ct_act.log, ct_deact.log, ct_startd.log, ct_stopd.log, and ct_nfssvr.log. These log files contain detailed logging information for the most recent associated task. Each time a task is repeated, its log file is overwritten.

History log - A ct_history.log file is created to store all installation-related tasks performed on the local node. This provides a convenient record of the Sun HPC ClusterTools 5 software installation history on the local node. Each time a new installation task is performed on the node, a new log entry is appended to the history log.

These node specific installation log files are created regardless of the installation method used, local or centralized.

Central Node Summary Log

When installation tasks are initiated from a central host, a summary log file named ct_summary.log is created on the central host. This log file records the final summary report that is generated by the CLI and GUI. The ct_summary.log is not overwritten when a new task is performed. As with the ct_history.log file, new entries are appended to the summary log file.

Summary of Tasks Performed by Sun HPC ClusterTools Software Utilities

FIGURE 1-1 shows an overview of the installation-related tasks you can perform using either the graphical installation tool, ctgui, or the corresponding command-line interface (CLI) tools. The CLI also supports starting and stopping all Sun HPC ClusterTools software daemons, two operations that are not available through ctgui.

Both ctgui and the CLI tools require superuser privileges to execute.

FIGURE 1-1 Sun HPC ClusterTools Software Installation Tasks

Figure depicting a flow diagram of installation activities

Select an Activity

Whether you use ctgui or the CLI tools, the various installation-related operations are independent of each other. With ctgui, you are given a wizard screen that asks you to select one of the four activities shown in FIGURE 1-1. With the CLI, you simply start the applicable utility: ctinstall, ctact, ctdeact, or ctremove. The CLI also provides a pair of utilities, ctstartd and ctstopd, for starting and stopping all Sun HPC ClusterTools software daemons on the cluster nodes. The GUI does not offer an equivalent feature.

The operations these tools control are described below.

Install the Software

The installation activity loads the Sun HPC ClusterTools software onto cluster nodes. You have the following choices in how the installation will be carried out:

Install in an NFS client-server cluster configuration.

Install in a non-NFS cluster configuration.

Activate the software on the cluster nodes automatically when the installation completes.

Activate the software on the cluster nodes manually a later time.

These choices are described more fully below.

If HPC ClusterTools 4 software is present on the cluster nodes, you can install the Sun HPC ClusterTools 5 software even if the version 4 software is active. However, version 4 must be inactive before you can activate version 5. Use the version 4 installation/configuration tools provided with the Sun HPC ClusterTools 4 software to change the version 4 software.

With the CLI command ctinstall, you can install individual Sun HPC ClusterTools 5 software packages instead of the entire software suite. This capability is not available with ctgui.

NFS Installation

In an NFS installation, all Sun HPC ClusterTools software packages, except for a subset containing Sun HPC ClusterTools software daemon startup scripts, are installed on a Sun NFS server and remotely mounted on the NFS client nodes in the cluster. The packages containing the daemon startup scripts must be installed locally on each cluster node.

The packages that contain the daemon startup scripts are placed in the root file system of the cluster nodes and are referred to as the root packages.

The NFS server can be one of the NFS client nodes in the cluster, but need not be. When a Sun HPC ClusterTools software NFS server is also one of the cluster nodes, both root and non-root packages are installed on the server.

Non-NFS Installation

In non-NFS configurations, a complete copy of the Sun HPC ClusterTools software--both root and non-root packages--is installed locally on each node in the cluster.

Install With Automatic Activation

When you initiate a software installation operation you can specify to have the nodes activated automatically as soon as the installation process completes.

When you use the automatic activation feature, you must also specify one of the cluster nodes to serve as master node for the cluster. This node will have special daemons running on it. It will also contain cluster configuration and security information not present on the other cluster nodes.

Install Without Activating

You can install the software without activating it. In such cases, you do not specify a master node until you activate the node manually at a later time.

Activate the Sun HPC ClusterTools Software

The node activation step prepares the cluster nodes so that the installed Sun HPC ClusterTools software can be used. Specifically, node activation:

Sets up symbolic links that point to the Sun HPC ClusterTools software.

Starts a set of Sun HPC ClusterTools software daemons.

If a node contains both versions 4 and 5 of the Sun HPC ClusterTools software, only one version can be active at a time.

Deactivate the Sun HPC ClusterTools Software

Node deactivation stops the Sun Cluster Runtime Environment (CRE) daemons and removes the symbolic links that point to the Sun HPC ClusterTools software. This step makes the software unusable until the node is activated again.

Remove the Sun HPC ClusterTools Software

This operation deletes Sun HPC ClusterTools software packages from the cluster nodes on which it is executed. If a node is active at the time you initiate the removal operation, it will be deactivated automatically before the software is removed.

With the CLI command ctremove, you can remove individual Sun HPC ClusterTools software packages instead of the entire software suite.

Start and Stopping Sun HPC ClusterTools Software Daemons

Starting and stopping the Sun HPC ClusterTools software daemons are subtasks in software activation and deactivation. You can use the commands ctstartd and ctstopd to start and stop all the Sun HPC ClusterTools software daemons without performing the full range of actions associated with activating and deactivating the software.

Recommendations for Installing Sun HPC ClusterTools 5 Software on Large Cluster Installations

The following are tips for installing Sun HPC ClusterTools 5 software on clusters containing hundreds of nodes using the centralized method:

Minimize other system activity during installation - Invoking installation of Sun HPC ClusterTools 5 software on hundreds of nodes from a central host imposes high demands on system resources. Avoid system resource exhaustion by keeping the cluster nodes as quiescent as possible during the installation.

Use a node list file - For various centralized installation tasks, you specify the nodes on which the task is to be invoked. You have the choice of specifying the nodes either on the command line, using the -n option or by referencing a node list file using the -N option. If you reference a node list file, you only enter the node names once when you create the file.

Reduce system resource consumption on the central host - You can avoid overtaxing system resources on a single central host by using more than one central host. Simply divide the total list of nodes to be installed on into separate node lists, and initiate the installation commands on the various central hosts, with each host using a different node list.

Use the -g option with CLI-initiated tasks - Use the -g option with CLI commands to obtain a list of nodes that successfully executed the command and a separate list of nodes that failed. You can then reference the list of failed nodes with the -N option in a later retry of the command.

Use the -k option with CLI-initiated tasks - Use the -k option with CLI commands to have all logs saved on the central node where the command was initiated. This option makes it unnecessary to go to each node to examine local logs. The Save Logs feature on the GUI provides the same capability.

Note - When using the telnet connection method, all nodes specified in a given centralized invocation must share the same password. If you have more than one password for the nodes in the cluster, separate the nodes into installation groups in which all nodes share the same password, and perform separate command invocations for each node group.