Benchmark Factory for Database 8.3 - User Guide

Install Remote Agent on Linux

Benchmark Factory provides a Benchmark Factory Agent that you can deploy and use on a Linux platform.

Supported Operating Systems

The Benchmark Factory Agent for Linux supports the following platforms only:

• Red Hat 7.x (64-bit)
• CentOs 7.x (64-bit)
• Oracle Linux 7.x (64-bit)

Supported Databases

The Benchmark Factory Agent for Linux can be used to test workloads against the following database types:

• PostgreSQL
• Oracle
• MySQL

See the Benchmark Factory Release Notes for information about database versions supported by Benchmark Factory.

Prerequisites

Oracle Client. If you intend to use the Benchmark Factory Agent for Linux when testing against an Oracle database, ensure an Oracle Client is installed on the same Linux machine as the Agent. Review the following details:

• Oracle Client and Instant Client versions 10g R2 and later are supported.
• The Benchmark Factory Agent for Linux is compiled for Oracle Instant Client 10g R2. If you use a version of the Oracle Instant Client later than 10g R2, you must create a symbolic link to the shared library for the Oracle Instant Client you intend to use. See the instructions below.
• When installing an Oracle Instant Client, ensure you complete the necessary installation steps, such as setting the LD_LIBRARY_PATH in the .bash_profile, if necessary. See https://www.oracle.com/technetwork/topics/linuxx86-64soft-092277.html#ic_x64_inst for detailed instant client installation instructions.

(Oracle Instant Client 11g or later) Create symbolic link for Oracle shared library

1. Locate the directory where the Oracle Instant Client you intend to use is installed.

2. From this directory, execute the following command:

   ln -s libclntsh.so.<version> libclntsh.so.10.1

   Where <version> is the version of libclntsh.so for the Oracle client you intend to use.

   For example, if you intent to use Oracle Instant Client 11g R2, you would run the following:

   ln -s libclntsh.so.11.1 libclntsh.so.10.1

RPM Package

Beginning with Benchmark Factory 8.2, the Agent for Linux is provided as an RPM package. The RPM package can be downloaded or accessed from a quest repository at https://bintray.com/quest/bmfrepo. Benchmark Factory provides the following methods for installing the Agent on Linux:

• (Recommended method) Automatically install the agent through the Benchmark Factory Console. This method uses YUM and automatically accesses the RPM package from the quest bintray repository.
• Download the RPM package and use YUM or another RPM package management tool to install the agent.
• Use YUM or another RPM package management tool to install the agent from the quest bintray repository.

Installation

You can install the Benchmark Factory Agent on a remote Linux machine manually or through the Benchmark Factory Console. To install using the Benchmark Factory Console, you must be able to connect to the remote machine.

Install the agent on each Linux machine you wish to use as an agent machine.

To install a remote agent from the console

This is the recommended method for installing an agent on Linux.

1. Select Edit | Settings | Agent.
2. On the Agent Settings page, click Setup New User Agent.
3. Select Linux and click OK. The Setup Windows Agent dialog opens.

4. Enter the connection information for the remote machine. Review the following for additional information:

   IP Address / Name	Enter the IP address or name of the remote machine.
   User Name	Enter the user name to use to connect. You must use the root user when installing a remote agent on Linux through the Benchmark Factory Console. If you cannot log in as root, install the agent on Linux manually. See the instructions below.
   Password	Enter the password to use to connect.
   # of Agents	Specify the number of agents to deploy on this machine.

5. Click OK. Benchmark Factory checks the remote machine for an installed agent.

6. If no agent is found, Benchmark Factory prompts you to install the agent. Click Yes in the prompt window. Benchmark Factory then uses YUM to install the agent.

7. When the agent is successfully installed, the progress window closes and the new remote agent is displayed in the list.
8. After the remote agent is installed, you can double-click the agent name in the list to modify agent options.

Download and manually install agent on Linux

1. Download the Benchmark Factory Agent for Linux RPM package from https://bintray.com/quest/bmfrepo. This RPM package contains the files necessary to set up the Benchmark Factory Agent on a Linux platform and has the following naming format:

   bmfagent-<version>-<buildnumber>.el7.x86_64.rpm.

   Select the RPM package file with the same <version> and <buildnumber> as your Benchmark Factory Console.

2. Install the RPM package. For example, install the RPM package using YUM.

   sudo yum install <bmfagentpackage>.rpm

3. After installing the remote agent on the Linux machine, return to the Benchmark Factory Console to set up the remote agent. See Set Up New User Agent for more information.

Manually install agent from repository

1. In the YUM configuration file directory (/etc/yum.repos.d/) on the Linux machine, create a YUM Repository configuration file. Name the file bmfrel.el7.repo.

2. Add the following text to this file:

   [bmfrelrepo-el]

   name=Benchmark Factory Release Repo

   baseurl=https://quest.bintray.com/bmfrepo/7/$basearch

   gpgcheck=0

   repo_gpgcheck=0

   enabled=1

3. Use this file along with the appropriate YUM commands to install and manage the RPM package for the Benchmark Factory Agent on Linux.
4. After installing the remote agent on the Linux machine, return to the Benchmark Factory Console to set up the remote agent. See Set Up New User Agent for more information.

Additional Requirements

Edit the Hosts File

Benchmark Factory uses the name for the Benchmark Factory Console machine. If your network communication does not include host name resolution, you may need to add the name and IP address of the Benchmark Factory Console host machine to the hosts file on the Linux machine. This will allow the agent machine to communicate with the Benchmark Factory Console machine.

Additional Requirements for Running the Agent on Linux

After you install and set up the remote agent on a Linux platform, the Benchmark Factory Console starts the agent when needed. To successfully run the agent on Linux, ensure the following requirements are met.

• Ensure the at package is installed on the Linux Agent machine. The at package is required in order to start the Linux Agent from the Benchmark Factory Console. If the package is not installed, you can install it using your package manager, such as Yum.

  yum install at

• Ensure the atd service is running on the Linux Agent machine. Use the following commands:

  systemctl start atd

  systemctl enable atd

Run Benchmark Factory with Multiple Agents

You can install agents on multiple agent machines in order to run tests using multiple agents. To configure Benchmark Factory with multiple agents you must:

• Load the proper client libraries (i.e. Oracle, SQL server, etc.) on the Benchmark Factory Console and agent machines.
• Install Benchmark Factory on the local and agent machines.

Install Multiple Agents

For best results, install remote agents through your Benchmark Factory console. You can install remote agents from the Settings dialog (Edit | Settings | Agent) or from the Agent page of the New/Edit Job Wizard. See Install Remote Agent on Windows for more information.

You can also install agents manually on the agent machine. If you do, you must configure the agent on the agent machine (see below). Then, in your local Benchmark Factory, you must set up the agent to register it with the console. See Set Up New User Agent for more information.

Configure Agent Machine

If you set up an agent machine by installing Benchmark Factory manually, you must configure the agent on the agent machine.

Each Benchmark Factory Agent must be configured with the address of the Benchmark Factory Console. Each Agent sends load testing results back to the Benchmark Factory Console.

If you use only the agent installed locally on the console machine, make sure your local agent is configured with the IP address (name) of your local machine.

To configure Benchmark Factory Agents (Windows)

1. On each agent machine, go to Program Files\Quest Software\Benchmark Factory <version>\bin and double-click Agent.exe to open the Agent dialog.
2. Select Options | Settings.

3. In Machine Name/IP field, enter the address of the Benchmark Factory console machine. In the Console Port field, enter the port for the Benchmark Factory console.

   

4. Click OK.
5. Repeat this procedure for each Agent machine

Set Up the Agents

To run a job using multiple agent machines, after the agents are installed and configured, you must set up (or add) each remote agent to the Benchmark Factory console.

• If the agent was installed manually, you can set up the agent in Edit | Settings | Agent. See Set Up New User Agent for more information.
• If the agent was installed remotely through the Benchmark Factory console, the agent is automatically set up.

Running a Job With Multiple Agents

To run a job with multiple agents

1. Open the New Job Wizard and create a workload.
2. Click Agent in the left panel of the New Job Wizard.
   • For an existing job, open the job in the Edit Job Wizard and click Agent in the left pane.
3. From the list, select the agents to use. The list displays agents you have previously set up in Benchmark Factory.
   • To add an agent, click Setup New User Agent. See Set Up New User Agent for more information.
4. As the job runs, all connected Agents will display in the Agent view/pane of the Benchmark Factory console.

Note: When you run a job using one or more local agents, if Agent utilization of resources on the local machine is too high, errors could occur.

View Agents from Benchmark Factory Console

When you run multiple Agents for load testing, you can view the Agents results as they run from the Benchmark Factory Console.

To view the Agents from the Benchmark Factory Console

1. From the Benchmark Factory main menu, select View | Agent. The Agent view displays. Statistics from All Agents connected to the Benchmark Factory Console and running are displayed here.

Agents View

Select View | Agent to open the Agents pane where you can view information about agents that are currently running. The Agent view displays the status of all connected agents.

Use the Agent Component Console

The primary agent component installed with Benchmark Factory is Agent.exe. On your local machine, the Benchmark Factory console always attempts to use Agent.exe first. If Agent.exe is not found, the console uses BMFAgent.exe.

Agent.exe includes a GUI/console. You can use this console to view transaction and virtual user statistics during test execution.

The Benchmark Factory agent starts automatically when a job using the agent is executed. You can also start/open the agent manually.

To start or open Agent.exe

1. Locate the executable file Agent.exe in the bin sub-directory where Benchmark Factory is installed. The default installation path is

   Program Files\Quest Software\Benchmark Factory <version>\bin

2. Double-click Agent.exe to start/open it.
3. Configure Agent Settings. You can configure settings for the Benchmark Factory Agent from the Agent.exe console.

   • Select Options | Settings. Review the following options.

     Machine Name / IP	This is the name or IP address of the Benchmark Factory Console machine to which this agent connects. For a local agent, this is the local Benchmark Factory Console. For a remote agent, this is the Benchmark Factory Console used to run the benchmark tests. If the agent was installed remotely through the Benchmark Factory console, the agent is automatically configured. If installed manually, you must configure the agent. See Running Benchmark Factory with Multiple Agents for more information.
     Console Port	Enter the port for the Benchmark Factory Console.
     Max Virtual Users	Use this field to specify the maximum number of virtual users that this agent is allowed to spawn.
     Error Logs / Result Files	Use these fields to specify a location for storing error logs and result files generated for this agent.

4. When the agent is open, information about virtual user activity is displayed while a job is executing. Review the following for additional information.

   • Test Status Tab—Displays the current test phase and a summary of test information.

     

   • Virtual Users Tab—Displays the raw data from each virtual user. The grid shows each virtual user and its statistics. Right-click a column header to sort by that column.

     

   • Transactions Tab—Displays status information, including average, minimum, and maximum times on all active transactions.

     

   • Output Tab—Displays the same information as the Messages window, including messages, status, and results.