Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Image

Cacti 0.8 Network Monitoring 45

GJdeBoer writes "The book is aimed at people who are managing a network and would like to get insight into the performance of that network. It covers the installation and configuration of the Cacti application. In the preface the book states that it's not necessary to be a Linux Guru to use the book and that exactly is the case. The book builds up your knowledge about Cacti and the necessary steps to configure it for your network, and it teaches you about Net-SNMP and RRDTool, the building blocks of Cacti." Read on for the rest of GJdeBoer's review.
Cacti 0.8 Network Monitoring
author Dinangkur Kundu, S. M. Ibrahim Lavlu
pages 132 pages
publisher Packt publishing
rating 9/10
reviewer Gert-Jan de Boer
ISBN 1847195962
summary This book teaches you to monitor your network, customize the output graph and input source, and take backups
As I've been working with Cacti for several years now, my aim was to get a book that describes the best practices for Cacti installations and to get a reference guide for myself. My hope was to get some more knowledge about the inner workings of Cacti and I think although meant for Cacti beginners, the book did a good job at that. I got a more clear idea about the architecture of Cacti which helps me with the integration of Cacti in my client's networks.

The book starts off with an introduction to Cacti. It explains what Cacti is, how the global architecture is and for what purposes it can be used. It also explains the basics of the prerequisite RRDTool. In the next chapter the book explains the installation of the prerequisites. The book then progresses on the installation, configuration and tasks like authentication and authorization of users. We then learn to add devices and assign templates to them.

The last chapters end the book with advanced topics for Cacti users such as Data Management and Cacti Management. It explains how to create your own data and snmp queries to be able to monitor custom devices. Personally, I found these chapters to be the most educational part of the book.

As for this book no advanced knowledge of Linux is needed. It explains the installation steps of Cacti and its prerequisites clearly and with a lot of exemplary screenshots. As Cacti is managed by means of an web interface it is the most clear way to make a point in a book about Cacti. The book is easy to read and I think the book covers the theory needed to install and operate a Cacti server perfectly. As it explains the use of Templates in Cacti and why you should use them, the book helps people build scalable and neat Cacti setups.

As a downside of the book I have found the clear focus being on the Debian side of Linux distributions. All the installation done in the book is by using apt-get, Debian and Ubuntu's package management system, but in the professional Linux world you are seeing more RedHat based distributions then Debian. I would have liked a couple of tooltips on how to install the prerequisites on RedHat or CentOS with the yum package manager or maybe by using source packages for installation. It's not a big downside for more advanced users but for the Linux novices, at who the book targets on, it could be a bit hard to find out the right way to install Cacti on a RedHat or CentOS box. Since the configuration of Cacti is the same on every platform this is only applicable for the installation chapters.

In general the book does exactly what the cover says: "Monitor your network with ease" although I found it a bit short. The book consists of a hundred and ten pages, but since there are a lot of screenshots on the pages there is less text. The book doesn't dive very deep into the inner workings of Cacti. One could argue that is exactly the point of the book: most people don't use that kind of knowledge. I would have liked a bit more insight into the MySQL database behind Cacti and troubleshooting steps for when your graphs stop working.

I think the book is great for people who want to start with Cacti because they want to monitor their network. They can install and operate a Cacti instance very quickly with help of this book without having previous knowledge of Linux. In my field of work I often come in contact with customers who have problems in their network. I always advice them to install a network monitoring appliance like Cacti. Since most of them use Windows networks they often have no experience in configuring a Linux server for Cacti. I think I will recommend this book in the future to these people.

Gert-Jan de Boer ia a self-employed IT Consultant with a company that specializes in Networking, Voice over IP, Storage and Virtualization.

You can purchase Cacti 0.8 Network Monitoring from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.

*

This discussion has been archived. No new comments can be posted.

Cacti 0.8 Network Monitoring

Comments Filter:
  • I have been using cacti for approximately two years now and can say that the installation time and configuration are well worth every second spent learning its in and outs. Even if this book is not endorsed by the cacti group, a comprehensive guide is an extrodinarily nice thing to have. When I first began using cacti, documentation was haphazard and scattered all over the place. This book is a good resource for those who are new to linux and Cacti in general.

    I myself have been developing a plugin for Cacti

    • by ascari ( 1400977 )

      I have been using cacti for approximately two years now

      As TFA points out cacti can be a bit harsh for a newb. Not even Carlos Castaneda went straight to cacti. One suggestion is to start out with something milder, like 'shrooms and work your way up to the real deal.

    • by jhoegl ( 638955 )
      Yes, Cacti+Nagios = win
      • Re: (Score:2, Informative)

        by inKubus ( 199753 )

        This is off-topic and possibly flamebait, but I'm a little tired of comments that have some stupid "equation" that equals "win" or "fail". It literally sounds like you are mentally disabled. This manner of speaking is cliched and was never that funny to begin with. Please give it a rest with the "win", "fail", "teh internets", "FTW", etc. It's old, and you can do better.

        • Re: (Score:2, Informative)

          by draxbear ( 735156 )

          I was wondering how Cacti relates to Nagios. Do the both do the same job or compliment one-another?
          In four words the post you're complaining about answered that question nicely.

          • I was wondering how Cacti relates to Nagios. Do the both do the same job or compliment one-another? In four words the post you're complaining about answered that question nicely.

            I am a longtime user of cacti, and dealt with MRTG before then. Nagios is designed for simple on-off monitoring. It does this very well. If you need pager rotation schedules, nagios is prolly your best bet. I like wowing the executives with pretty color graphs, and that is a job for Cacti. You can visually see the impact on th
  • What do you think that this says about version numbers? I'm not really taking any point of view here, other than that version numbers don't mean squat.

    Here we have a book that's been produced on a less than 1.0 version, alpha? beta? what?

    Google seems to keep beta on their products for a very long time, largely so they can't be held responsible for bugs.

    Yall chime in.

    • For some reason OSS people seem to have version numbers >= 1.0. I don't know why that is, but there are an amazing amount of OSS projects that seem stuck perpetually in the "below 1.0" area. Hell look at OpenSSL. They are in 1.0.0 beta, but the current stable release is 0.9.8. Ok well first there's the fact that one would think "stable" would imply above 1.0 and then there's the fact that OpenSSL has been deployed for YEARS in production environments and is in every way release code.

      I don't know what the

      • ... "maybe they figure that if they aren't 1.0 they aren't responsible for bugs or something" ...

        My guess would be that it's more of a corporate culture thing that leads Google to label their apps as "beta" for so long; their exemption from "responsibility" is specifically covered in their Terms of Service [google.com] as so:

        14.6 GOOGLE FURTHER EXPRESSLY DISCLAIMS ALL WARRANTIES AND CONDITIONS OF ANY KIND, WHETHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO THE IMPLIED WARRANTIES AND CONDITIONS OF MERCHANTAB
    • Re: (Score:3, Insightful)

      by pclminion ( 145572 )

      I came to post the same thing. I'm really tired of open source developers being so spineless in this regard. Come on -- your software is clearly successful, used by thousands, people are writing books about it -- surely you can make the commitment and say hey, we've created a real, useful product here, let's roll the version to 1.0. Sure, you might need to break your API at some future time, you might discover some terrible, data-destroying flaw... But at some point you ought to just pinch your nose and jum

      • Normally 1.0 implies that all features they expect/want to have in the "final version" of the product are present.

        The question is, which features do you decide are going to be in "version 1.0" and what goes in "version 2.0"? I think a lot of OSS developers have a vision of what their project will eventually become, and those features all need to work and be stable before it'll be "1.0".

        More commercially-focused developers may have a list of things they want in a "final" product, but then cull it down to a m

  • Comment removed based on user account deletion
  • by bofar ( 902274 ) on Monday February 08, 2010 @03:40PM (#31064590)
    Love the Amazon review, "My initial inclination is to say that this book is worthless. Given that I spent $35 on it, it's worse that worthless. At around 100 pages in length, the first 40 are dedicated toward understanding what a network is, a general overview of RRD and Cacti and a very poorly written install guide...."
  • Minutely graphs? (Score:1, Interesting)

    by Anonymous Coward

    Anyone here got Cacti working with minutely instead of five-minutely updated graphs for network interface statistics? It is essential to spot traffic spikes. But after wasting an afternoon trying to get it right, I presume I just do not have what it takes (voodoo skills).

    • if you mean the polling time, you just change the cacti cron job to run every 1 minute (default is 5), and in your data source / data template set Step to 60 (secs). usually you also set Heartbeat to 2x that of Step (so 120).

    • To get a minute by minute graph of your network traffic, you will have to sample at least twice that fast. i.e. every 30 seconds or preferrably faster.
       

  • Cacti is great for graphing performance, capacity planning and spotting anomalies while Nagios [nagios.org] is tops for monitoring / alerting. I have worked with many different monitoring tools and suites both commercial and open source. A well configured Nagios / Cacti solution is hard to beat for stability and usability.
  • by Anonymous Coward

    ..You should be aware of some default-setup 'flaws'. All of them correctable with some custom templates.

    * The default templates for load-monitoring is additive, they stack 1min/5min/15min average onto each other.
    Thus rendering the graph plot exponential rather than flat accurate.

    * Almost all graphs are averaged, round down, as they age.
    If you had a peak netload for half a day in December, you don't wanna see a graph that tells you that your 24h average was just like a regular day.

    * The templates for networ

  • I wonders if the author tried to mention how to integrated Cacti and Nagios in his book? Nagios is good at monitoring and sysadmin and manager might use the Cacti for trends and capacity plaining. At my work we are using Nagios for monitoring and pnp4nagios for trending the Nagios performance data and its work great so far.
  • Fair point kokoko1

"Being against torture ought to be sort of a multipartisan thing." -- Karl Lehenbauer, as amended by Jeff Daiell, a Libertarian

Working...