Transition from PSSP to cluster systems management (CSM) / [Dino Quintero ... et al.].

Format:

Book

Contributor:

Quintero, Dino.

International Business Machines Corporation. International Technical Support Organization.

Series:

IBM redbooks.

IBM redbooks

Language:

English

Subjects (All):

Parallel processing (Electronic computers).

IBM Cluster systems management.

Physical Description:

xiv, 268 p. : ill.

Edition:

1st ed.

Other Title:

Transition from Parallel System Support Programs to cluster systems management

Place of Publication:

[S.l.] : IBM, International Technical Support Organization, 2004.

Language Note:

English

System Details:

text file

Summary:

This IBM Redbooks publication covers the process of converting a PSSP cluster to a CSM cluster. It examines the different tools, utilities, documentation and other resources available to help the system administrator move a system from PSSP to CSM. It also examines the different paths a customer can take to achieve this transition and illustrates these different paths with step-by-step transition scenarios of actual clustered systems. This book also makes recommendations about which procedures are most suitable for different types of customer environments (high performance technical computing, server consolidation, business intelligence, etc.) and points out the relative advantages and disadvantages of the different procedures, tools, and methods. Customers will use the working knowledge presented in this book to plan and accomplish the transition of their own clustered systems from PSSP to CSM. Are you using GPFS? This transition book covers a number of GPFS transition methods, each with its own recommendations, GPFS cluster types, advantages versus disadvantages, and step-by-step scenarios. It also shows a method of moving GPFS storage between two clusters, which could be used during hardware transition or upgrades.

Contents:

Front cover

Contents

Notices

Trademarks

Preface

The team that wrote this redbook

Become a published author

Comments welcome

Chapter 1. Introduction

1.1 Summary of PSSP

1.2 Summary of CSM

1.3 Summary of PSSP to CSM transition

1.3.1 Transition types

1.3.2 Transition tools

1.4 Scope of this redbook

1.4.1 What we cover

1.4.2 What we do not cover

1.5 Chapter summaries

1.6 Additional resources

Chapter 2. Planning and considerations

2.1 Why transition to Cluster Systems Management (CSM)?

2.2 Introduction to the PSSP to CSM transition tools

2.2.1 Analysis tools

2.2.2 Configuration recording tool: mkcfgutils

2.2.3 PSSP removal tools

2.2.4 File collections to CFM conversion tool: cfgcfmutil

2.3 Procedural steps for each of the transition types

2.3.1 All-at-once transition

2.3.2 Frame-at-a-time (phased) transition

2.3.3 New hardware transition

2.4 Risks and time scales

2.5 Hardware and software prerequisites

2.6 Contingency plans

2.6.1 Back-up

2.6.2 Points of no return

2.7 Cluster verification

2.7.1 CSM verification

2.8 High Performance Computing (HPC) software stack

2.8.1 LoadLeveler

2.8.2 General Parallel File System (GPFS)

2.8.3 Virtual Shared Disk (VSD)

2.8.4 Parallel Environment (PE)

2.8.5 Parallel Engineering and Scientific Software Library (ESSL)

2.8.6 High Availability Cluster Multi-Processing (HACMP)

Chapter 3. Process steps for transition

3.1 Introduction

3.2 Initial points to check

3.3 Process steps for various transition techniques

3.3.1 AIX 5.2/PSSP 3.5 to AIX 5.2/CSM 1.3.3 transition

3.3.2 AIX 5.1/PSSP 3.4 on an existing system to AIX 5.2/CSM 1.3.3 on a new system with elements of the HPC software stack

3.3.3 AIX 5.2/PSSP 3.5 to AIX 5.2/CSM 1.3.3 transition with HPC.

3.4 Detailed transition steps

3.4.1 Run the PSSP cluster analysis report tool

3.4.2 Run the PSSP node analysis report tool

3.4.3 Run the PSSP configuration recording tool

3.4.4 Review and check the reports

3.4.5 Take system and application back-ups

3.4.6 Stop production

3.4.7 Shut down HACMP

3.4.8 Shut down LoadLeveler

3.4.9 Back up /spdata

3.4.10 Shut down GPFS

3.4.11 Shut down VSD

3.4.12 Uninstall PSSP from the nodes

3.4.13 Uninstall PSSP from CWS

3.4.14 Remove the SP switch adapter from the nodes

3.4.15 Install and configure the NIM master on the CWS (MS)

3.4.16 Install the CSM management server (MS)

3.4.17 Define the nodes to the CSM cluster

3.4.18 Create a NIM machine definition

3.4.19 Update CSM filesets to nodes

3.4.20 Adding nodes to the CSM cluster

3.4.21 Set up conserver

3.4.22 AIX migration using NIM

3.4.23 Post installation tasks

3.4.24 Verify the installation

3.4.25 Set up a peer domain

3.4.26 Set up HACMP

3.4.27 Configure VSD

3.4.28 Configure GPFS

3.4.29 Configure LoadLeveler

3.4.30 Change Hostname and IP address of MS or nodes (optional)

3.4.31 Store hardware control point user ID and password (Hardware Management Console (HMC) only)

3.4.32 alt_disk_install procedure

Chapter 4. Implications of transition

4.1 Items to check before and after the transition

4.1.1 Data saved by the transition tools

4.2 RRA (Restricted Root Access)

4.3 BIS (Boot Install Server)

4.4 Kerberos

4.5 Security on PSSP versus CSM

4.6 File collections

4.7 User management

4.8 Monitoring

4.8.1 Differences in monitoring between PSSP to CSM

4.8.2 Sample of monitoring item

4.9 Custom SDR classes

4.10 Customer scripts for specific tasks

4.11 Node-to-node access within CSM, RPD

4.12 sysctl

4.13 NTP

4.14 Automount.

4.15 Accounting

4.16 Node groups

4.17 NIM (Network Install Manager)

4.17.1 Adding the secondary adapters

4.18 HACWS

4.19 Tivoli T/EC Adapter

4.20 DCE

Chapter 5. High Performance Computing (HPC) recommendations

5.1 High Performance Computing (HPC) transition

5.2 GPFS and VSD transition

5.2.1 GPFS and VSD transition tools

5.2.2 GPFS and VSD transition scenarios

5.2.3 Staged GPFS and VSD move: shrink and grow

5.2.4 GPFS and VSD in-place migration

5.3 High Availability Cluster Multi-Processing (HACMP)

5.4 LoadLeveler

5.5 PE and PESSL

Appendix A. How to configure monitoring in CSM from pman subscriptions

Switch responds check

/ (root) fs full check

Page space usage check

hardmon monitor on CWS check

Appendix B. Hints and tips

Here are our recommendations

General recommendations

Appendix C. Transition tools sample outputs

Map of logs and saved files for each tool

Example output from running mkpssprpt

Example log file from running mkpssprpt

Example output from running mknoderpt

Example log from running mknoderpt

Example PSSP analysis report from mkpssprpt

Example output of the mknoderpt report

Example output files from the mkcfgutils command

Example log file showing activity of rmpsspnode

Example log file from rmpssp

Files generated and saved by rmpssp and mkcfgutils

List of files generated and saved by rmpssp on the node

Output from the cfgcfmutil command

Examples of logs generated by cfgcfmutil

Appendix D. Additional GPFS/VSD sample scenarios

Shrink and grow

VSD scripts, in conjunction with GPFS commands

Exporting a GPFS filesystem using mmexportfs

Using the VSD scripts to clone VSDs

Importing a GPFS file system using mmimportfs

How to define VSDs manually

Abbreviations and acronyms

Related publications.

IBM Redbooks

Other publications

Online resources

How to get IBM Redbooks

Help from IBM

Index

Back cover.

Notes:

"August 2004."

"SG24-6967-00."

Includes bibliographical references and index.

OCLC:

137342310