Learn about the planning tasks that you need to perform to get the performance that your business needs from your server running Linux®.
Before you begin your planning tasks, complete these tasks.
| Before you begin | |
|---|---|
| __ | See the eServer™ Linux  Web site for information on the various server platforms
that support Linux and
the three supported Linux distributions. |
| __ | Read the Optimizing Linux environments for performance and
scalability white paper to get an overview of the tasks
that are involved in planning for Linux performance. |
| Linux performance planning tasks | |||
|---|---|---|---|
| __ | Identify performance considerations
for Linux distribution
requirements Each Linux distribution has unique requirements that need to be met to get desired performance. Also, each distribution is tuned differently with different shell scripts or performance-enhancement tools. To get the most out of your distributions, see these resources: |
||
| __ | Plan to create a performance
baseline You need to assess the existing workloads that are
assigned to your Linux partitions.
There are several tools that you can use to measure these workloads. See the Linux on POWER® applications |
||
| __ | Plan to simulate the environment You need to model current and future capacity to ensure that you have the adequate resources for your workloads. There are several capacity simulation tools available. See the Linux on POWER applications |
||
| __ | Plan for Capacity on Demand If your simulations show that you need seasonal or permanent capacity upgrades, you need to be prepared to address these needs. For more information about planning for, activating, and ordering additional processors for your hardware system, see Planning for Capacity on Demand . |
||
| __ | Plan to test the performance
of your solution It is important to test the performance of your solution before running it in a production environment. Testing validates that your simulations take into account all of the variables and accurately represent production workloads. Testing also stresses the actual code with user transaction loads to capture performance bottlenecks that are inherent in the applications that you plan to run under Linux. For more information about planning to test the performance of your solution in a lab setting, see Planning for testing. |
||
| __ | Plan to sustain performance
through growth Ongoing monitoring of your Linux systems provides for updated data points that you can use to validate and enhance your system performance. |
||
When you have completed these tasks, you should have a plan for Linux performance that identifies these elements:
| After you finish | |
|---|---|
| __ | Identify and record any necessary changes to the operating environment of your server, such as adding a large number of users or a significant software product. |
| __ | Identify and record any considerations for upgrading to a new Linux release, such as requirements for disk storage, CPU, and memory. |
| __ | Identify and record communications performance requirements. |
| __ | Identify and record database accessibility requirements as they relate to disk and disk device requirements. |
| __ | Identify and record the components of the workload. |
| __ | Identify and record all performance requirements. |
| __ | Identify and record the estimated resource requirements of the workload. |
| __ | Record a complete strategy for meeting the performance requirements of projected workloads. |