Notice: Undefined index: order_next_posts in /nas/content/live/gadgetmag/wp-content/plugins/smart-scroll-posts/smart-scroll-posts.php on line 194

Notice: Undefined index: post_link_target in /nas/content/live/gadgetmag/wp-content/plugins/smart-scroll-posts/smart-scroll-posts.php on line 195

Notice: Undefined index: posts_featured_size in /nas/content/live/gadgetmag/wp-content/plugins/smart-scroll-posts/smart-scroll-posts.php on line 196
News

Use Orca to monitor system performance

Getting the performance-monitoring data for a server can be invaluable for a person handling a production environment. There are a number of ways to collect this information. Once you collect this data it can be quite a challenge to interpret it accurately and can be even more difficult to make

Sukrit Dhandhania
Advisor
Sukrit Dhandhania
Sukrit has spent over seven years working with GNU/Linux servers, during which he set up and managed a number of  database and web servers, running live production environments.

Requirements:
Perl 5.005_03
Math::IntervalSearch 1.05
Digest::MD5 2.33
Devel::DProf 19990108
Date::Parse 2.24
RRDtool’s RRDs Perl module
Storable 2.15


Getting the performance-monitoring data for a server can be invaluable for a person handling a production environment. There are a number of ways to collect this information. Once you collect this data it can be quite a challenge to interpret it accurately and can be even more difficult to make alterations to your system configuration based on the data. To do this effectively you need to run several rounds of performance testing and fixing. Sometimes it can be quite helpful to get a reading of your system’s performance so as to know where you stand. We recently covered two tools, Dstat and gnuplot, that together allow you to collect performance monitoring data and plot into performance graphs. In this article we’ll look at another tool – Orca – that both collects performance data and also plots graphs which helps in interpreting the system performance data. The important difference between Orca and Dstat/gnuplot is that, unlike the latter, Orca plots live reports on the fly and allows you to view these ports using your web server, which can be very useful. This allows you to be able to view live performance graphs of your servers pretty much from anywhere in the world.

Orca was written by Dr Blair Zajac, who has received a PhD Caltech in Geophysics. This wonderful tool is a free download and is relatively simple to get going. Once installed and configured properly it reads system performance data and plots the data into graphs such as the one shown here. The graphs are made available to the user in HTML files which can be accessed in your web browser over the network. Orca reads and plots a performance graph for a number of parameters by default. These include memory, CPU, network, disk IO and so on. The Orca project is hosted here. You can download the source code of the project, read some documentation on it and also view the archives of the Orca support mailing list on this website. If you are serious about using Orca, we would encourage you to sign up for the orca-users mailing list here.

On the Orca website, Dr Blair Zajac has a note requesting a donation as a token of appreciation. We would encourage you to do so, particularly if you are using this wonderful tool as part of a commercial venture. He also has an Amazon wish list from which you can buy him something if you like. The tool is quite useful and we’ve found that it can greatly enhance your ability to assess the load on your servers.

Let us now proceed to the installation. We will then make some configuration changes, then set up a cron job so that Orca can provide you with up-to-date information on the state of your server. Note that Orca is a very flexible tool and there are several configuration parameters that we will not cover in this article. We will leave many of these parameters to the default settings. Once you are comfortable with the configuration of Orca, you can tweak it further.

Orca has been written using the Perl scripting language. You need to have Perl 5.005_03 or better installed on your computer. Most modern Linux distributions ship with Perl 5.005_03 or better. Run a quick check to see if you already have perl installed. Run the command ‘# perl-v’ in terminal window to check. If this command returns with a message like ‘This is perl, v5.8.8 built for i386-linux-thread-multi’, you are good to proceed. If not, check the documentation of your Linux distribution to see how to get Perl installed on your computer. There are also a few Perl modules that Orca requires to function fully. (See the box titled ‘Orca Prerequisites’). The Orca project has been kind enough to bundle these Perl modules along with the source code of the project. You can either download it off the project’s website or find it on the DVD accompanying this issue of Linux User & Developer. You will also need to have a web server if you want to be able to view the reports it generates over the web. Again, most modern distributions of Linux ship with Apache installed by default. If you have landed a machine that does not have Apache, please refer to your distribution’s documentation to see how to get it installed. Without Apache you will not be able to view the Orca reports properly.

×