Table of Contents

Search

  1. Preface
  2. Performance Tuning Overview
  3. Bottlenecks
  4. Optimizing the Target
  5. Optimizing the Source
  6. Optimizing Mappings
  7. Optimizing Transformations
  8. Optimizing Sessions
  9. Optimizing Grid Deployments
  10. Optimizing the PowerCenter Components
  11. Optimizing the System
  12. Using Pipeline Partitions
  13. Performance Counters

Performance Tuning Guide

Performance Tuning Guide

Balancing CPU and Memory Usage

Balancing CPU and Memory Usage

Unlike local file systems, a shared file system server can take extra CPU cycles to access files. If you use one of the computation nodes as the shared file system server for the rest of the nodes, it might become overloaded and become a bottleneck for the entire grid. When the shared file system server is overloaded, CPU cycles can increase, along with repeated transmissions and time-out requests.
To avoid this, use one or more machines as dedicated shared file system servers for your PowerCenter grid nodes. Each machine should have enough storage, CPUs, and network bandwidth for required tasks.
Alternatively, you can cross-mount the shared file system server to distribute the file server load across the nodes of the grid. When PowerCenter mappings and sessions are configured to use an even balance of I/O and CPU usage, cross-mounting shared file system servers can optimize performance. If the number of nodes in the grid is small and you have a balanced mix of I/O and CPU usage, you might not need a dedicated shared file system server.
When you use more than one shared file system server, dedicated or cross-mounted, try to distribute shared files across the servers.