SGI Linux™ Advanced Cluster Environment Administrator's Guide

Table of Contents

New Features in This Guide

About This Guide

Related Publications and Web Sites

Obtaining Publications

Conventions

Reader Comments

1. Introduction

Terminology

Assumptions

Contents of the Linux ACE Release

2. Topology

3. Preinstallation Requirements

TCP/IP Connectivity

Remote Execution for root

Installation of Additional Base OS Packages

Verification of Connectivity and Remote Access

4. Installing ACE Software

Install ACE using Interactive Mode

Install ACE using Batch Mode

5. Rebuilding the Kernel when using Myrinet without SGI ProPack

Standard Rebuild Procedure

Shortcut when the Source Matches the Running Kernel

6. Installing the Myrinet Driver (gm)

Steps on All Nodes

Steps on the Head Node Only

Monitoring Myrinet Packet Traffic

7. Configure gm-mpich

8. Verify the Installation

What confidence Verifies

Using confidence

Remove the CD-ROM

9. Upgrading an Existing Cluster

10. Using and Enhancing Performance Co-Pilot (PCP)

Configuring PCP for Remote Display

Configuring your X Server for 16-Bit Display

Using PCP Visualization Tools

Using PCP to Visualize MPI Jobs

Using PCP to Visualize Myrinet Switch Traffic

Using PCP to Visualize Cisco Switch Traffic

Monitoring Specific Applications and Processes

Customizing PCP Visualization Tools

Further Information About PCP

11. Lconsole Utility

Configure the Serial Console

Connect to Lconsole

Use Lconsole Features

Add Lconsole Users and Change Passwords

Delete Lconsole Users

12. Building Products from Source RPMs

13. Adding an Execution Node to the Cluster

Configuration Tasks

Interactive Mode Example

Batch Mode Example

14. Remove an Execution Node

15. Syncronizing Clocks in the Cluster

16. Troubleshooting

No rsh Access

.rhosts Errors in /var/log/messages

Problems with PAM

MPICH Problems

PBS Problems

PCP Errors

NFS Version Defaults to 2

A. Connecting the Digi EtherLite Serial Multiplexer

Hardware and Software Requirements

Connectivity

DHCP Configuration

EtherLite 16 Driver Installation

Index