Starting the eG Agent

To start the agent, first, login as the eG user, and then, run the command /opt/egurkha/bin/start_agent. The following message would be displayed

Starting the eG agent. . .

The eG agent 5.1 has been started...

Please check the file: /opt/egurkha/agent/logs/error_log

for any errors while executing the agent.

The following error message would appear if the agent fails to start

The eG agent failed to start ...

This is probably because the eG manager’s license does not permit this agent to communicate with it.

Please check the file: /opt/egurkha/agent/logs/error_log for details.

An eG agent may fail to start if:

  • the eG manager cannot be contacted (or)
  • there is a license mismatch

Note:

The following processes will run when the agent starts:

  • A Java process that executes EgMainAgent - this is the core agent process
  • A script named eGAgentMon that periodically monitors the agent and restarts it if the agent ever fails

Then, the eG administrative interface (described in the Administering eG Enterprise document) can be used to enable/disable authentication during manager/agent communication.

In Linux, AIX, and HPUX systems, error and output logging for the eG agent can be triggered by editing the start_agent script in the /opt/egurkha/bin directory. The steps involved in this process are discussed hereunder:

  1. Open the start_agent file.
  2. Edit the line that begins with nohup /opt/egurkha/jre/bin/java–Xrs EgMainAgent . . . (see Figure 1).

    Figure 1 : The start_agent script

  3. At the end of the line indicated by Figure 1, you can find an entry that reads as follows: . . . /dev/null 2>/dev/null.
  4. This entry is appended to the nohup /opt/egurkha/jre/bin/java line by default, and indicates that both output and error logging is not enabled for the eG agent in question, by default.
  5. To enable output logging, replace the first occurrence of /dev/null in the line with the full path to an output log file (see Figure 2). Similarly, to enable error logging, replace the second occurrence of /dev/null with the full path to the error log file (see Figure 2).

    Figure 2 : The edited start_agent script

  6. Finally, save the start_agent script.
  7. Restart the agent.

In Solaris environments, error and output logging for the eG agent can be triggered by editing the starta script in the /opt/egurkha/bin directory. The steps involved in this process are discussed hereunder:

  1. Open the starta file.
  2. Edit the line that begins with nohup java –client -Xrs . . . (see Figure 3).

    Figure 3 : The starta script

  3. At the end of the line indicated by Figure 3, you can find an entry that reads as follows: . . . /dev/null 2>&1.
  4. This entry is appended to the nohup java line by default, and indicates that both output and error logging is not enabled for the eG agent in question, by default.
  5. Unlike Linux, HPUX, and AIX agents, which can be configured with two separate log files for error and output logging respectively, the agent on Solaris can only be configured with a single log file; both errors and output will be captured by this log file only. Therefore, to enable error and output logging, replace the /dev/null entry in the nohup line with the full path to the log file (see Figure 4).

  6. Figure 4 : The edited starta script

    Finally, save the starta script.
  7. Restart the agent.

Note:

Before attempting to administer the eG Enterprise system, ensure that sysstat package is installed on the component to be monitored (check for the existence of the iostat command on the target system). The DiskActivity Test will not function on Unix environments without the sysstat package.