If you have already started the ‘/etc/init.d/procallator’ script some time back, it should have collected enough information so as to allow you to plot a graph of the data. The default interval between readings is five minutes. So if at least five minutes have passed since you started Procallator, run the following command to create your first Orca performance graph:

# orca -o /opt/orca/etc/procallator.cfg

Replace ‘/opt/orca/etc/procallator.cfg’ with the correct path on the Procallator configuration file on your system. The above command initialises Orca, which then creates the basic files and creates the first report HTML file and places them under the directory ‘/var/www/html/orca’. Once you run the command, wait for Orca to return with a message saying ‘Orca has completed’. This means your reports are now ready. Launch your web browser and go to the URL of your web server, http://[yourserver.com]/orca/procallator. The index.html page would be a bare bones page listing a number of reports. There wouldn’t be too much data to show if you haven’t given Procallator long to collect system data, but you should be able to now view some reports.
The graphs in the reports contain the system’s host name in the title so that you can easily figure out which graph relates to which server. This is quite useful, particularly when you run Orca on a cluster of servers and collect reports across all the servers to one location. The graphs Orca generates are done so using the PNG format. You can modify the Orca code so that it exports these graphs to another location instead of generating HTML reports. This way you can insert the data and graphs generated by Orca in another web application if you like.
Now that your graphs have been generated correctly, it’s time to set up some scheduling so that these graphs are automatically generated at regular intervals of time. We’ll use cron jobs to set this up. Let’s look at how to get Orca to generate these reports for us on an hourly and daily basis. Launch a terminal window and create a new cron job using the command ‘# crontab -e’. Now add the following lines in the file:

@hourly orca -o /opt/orca/etc/procallator.cfg
@daily orca -o /opt/orca/etc/procallator.cfg

Replace ‘/opt/orca/etc/procallator.cfg’ with the correct path of your configuration file. Save and quit. Once these settings kick in at the beginning of the next hour and day, you should be able to view hourly and daily system performance reports in your web browser.

In Conclusion
Your system performance reports should now be in place and you should be able to view hourly, weekly, monthly and yearly reports of all the system parameters configured over your browser. This information can be quite useful, particularly if the system administrators look at the reports often and not only in times of crisis. On the other hand, the sheer amount of information displayed by tools such as Orca can be quite daunting to analyse. However, with some patience one can go through the piles of data and find information that can be of great use. Orca provides both system administrators as well as the management of companies with valuable information to help them make important decisions about upgrading IT infrastructure. That said, Orca is just another tool that can assist you with your decision. It is up to the administrators and the members of the management to properly interpret and act upon the information provided by it.

This article originally appeared in issue 82 of Linux User & Developer. You can find more tutorials from the magazine here.