Suggestions for EduGeek Cacti Guide

[Geoff]

Ric has asked me to put together a guide on Cacti. So I'm wondering what my intended audience would like to see in such a guide beyond the basics?

[tosca925]

What is Cacti?

[john]

I was thinking it was a new geek thing to have Cacti in the office

[Geoff]

http://cacti.net/what_is_cacti.php

[localzuk]

Err... I use it but really found it the simplest piece of software I've come across to set up. Why would a guide be needed? *confused*

 

But if you want suggestions:

 

How to install on a normal system (ie. Ubuntu )

How to add your various devices

How to add graphs to your various devices

How to structure your various graphs into a useful structure (including picking out individual, important, ones to place under a single page - such as server traffic etc...)

[Ric_]

What I had in mind was pretty much what localzuk has said.

 

I would say that it should be compiled from source to be distro specific (particularly as Debian-based distros often use their own file locations).

 

I was also thinking of some notes on what particular metrics should be added - i.e. what is useful and any stumbling-blocks that should be avoided. For instance the best way to manage a trunk on a HP Procurve as well as non-standard metrics such as switch temperatures and CPU utilisation.

 

To give people an idea of the bigger picture... The idea is to eventually have a number of these documents so that people (who don't necessarily have a lot of experience with these things) can run through them and set up a number of open source solutions for performing daily tasks.

[Geoff]

For instance the best way to manage a trunk on a HP Procurve as well as non-standard metrics such as switch temperatures and CPU utilisation.

 

That's great but I don't have any clever HP Procuvre kit like that.

[Ric_]

@Geoff: If you check http://www.nagiosexchange.org you will find the relevant OIDs for CPU utilisation, temperature, etc. on the Procurve kit

 

The trunk ports simply come up as port numbers. For example, if you have a 24-port switch and ports 23 and 24 are combined in a trunk, the metrics for ports 23 and 24 will show the amount of data flowing through the individual ports whilst port 25 will show metrics for the trunk.

 

Since we will wiki-fy this, members will be able to add config examples to the end so that different brands of kit and different metrics can be monitored.

[NetworkGeezer]

Ric has asked me to put together a guide on Cacti. So I'm wondering what my intended audience would like to see in such a guide beyond the basics?

 

Extended paragraphs

[Geoff]

If you check http://www.nagiosexchange.org you will find the relevant OIDs for CPU utilisation, temperature, etc. on the Procurve kit

 

Indeed, I had managed to forget last night that we do actually have a 2650 in Room 17. So I've gone and plugged in the SNMP checks into Nagios.

 

Have you already cooked up the relevant XML configuration for Cacti to graph memory/cpu loads of ProCurves?

[Ric_]

@Geoff: Nope... you can have a quick look at my Nagios ones though...

 

##### Checks for HP Procurve Switches
define command{
       command_name    rmc_check_hpmemoryfree
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.5.1.1.2.1.1.1.6.1 -t 5 -w $ARG2$ -c $ARG3$ -u bytes -l free
       }
define command{
       command_name    rmc_check_hp_cpu
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.5.1.9.6.1.0 -t 5  -w $ARG2$ -c $ARG3$ -u % -l "5min cpu"
       }
define command{
       command_name    rmc_check_hpfan
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.1 -w $ARG2$ -c $ARG3$ -l 'Fan status'
       }
define command{
       command_name    rmc_check_hppower
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.2 -w $ARG2$ -c $ARG3$ -l 'Power Supply status'
       }
define command{
       command_name    rmc_check_hptemp
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.4 -w $ARG2$ -c $ARG3$ -l 'Temprature status'
       }
define command{
       command_name    rmc_check_hppower_4100
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.3 -w $ARG2$ -c $ARG3$ -l 'Power Supply status'
       }

[ICTNUT]

Graphs are nice and indeed I use Cacti, although probobly not to is't fullest potential however I tend to rely on NAGIOS more than anything for segment, hardware, and service failures including notification on and off site.

 

I would be more than happy to hear how to fully utilise cacti in addition to nagios.

[TechMonkey]

Difference between Cacti and Nagios?

 

The metrics would be a fantastic idea. I know here when we set up a passler monitoring station it took about 3 attempts before we were happy and plenty of maths to get to there!

 

Only thing I would realllllly ask for is assume nothing. I really get narked at these beginner guides that forget that some trying this aren't Linux and are possibly using this as a way into Linux. Then they write something like:

"Then exchange the partition codes and mount points in the usual way".

At that point I scream go green and stomp around abit. Possibly the reason I have loved finding the apt-get in Ubuntu but I also have it in the back of my mind that I'm not sure where things are going and what they are doing.

[Geoff]

Nagios gives you an overview of whats happening now. Cacti lets you see what happened yesterday, last week, or further.

[Ric_]

@TechMonkey: I realise tht not everyone knows Linux... this is why the Nagios and Cacti guides will assume only that the base system is installed. They will then be written in a step-by-step fashion to the point where a simple configuration is up and running.

 

I intend the guides to be extended by the community with the addition of example configs and uses.

 

BTW... Nagios also lets you know when it all goes tits up!

[ChrisH]

Wanting to run systems like nagios, web filters etc is what gets people to learn Linux and they never look back. I certainly didnt. I learnt most of my Linux stuff playing with squid and Dansguardian a couple of years ago. Last year I extended that further by learning Nagios with guidance from Ric and as a result of that I started writing scripts in Perl and learnt alot about SNMP.

Recently I have been extending that skillset further with Pykota but I have been quiete lazy with this one and have had Ric tell me most of it over the phone as I did it

[Ric_]

Anybody else wanting to learn more should go to http://edugeek.net/index.php?name=Forums&file=viewtopic&t=8491 and ask me to add whatever you need to my document... I should have a first draft some time tomorrow

[plexer]

Did the Cacti guide ever get produced?

 

Ben

[DanW]

@Geoff: Nope... you can have a quick look at my Nagios ones though...

 

##### Checks for HP Procurve Switches
define command{
       command_name    rmc_check_hpmemoryfree
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.5.1.1.2.1.1.1.6.1 -t 5 -w $ARG2$ -c $ARG3$ -u bytes -l free
       }
define command{
       command_name    rmc_check_hp_cpu
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.5.1.9.6.1.0 -t 5  -w $ARG2$ -c $ARG3$ -u % -l "5min cpu"
       }
define command{
       command_name    rmc_check_hpfan
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.1 -w $ARG2$ -c $ARG3$ -l 'Fan status'
       }
define command{
       command_name    rmc_check_hppower
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.2 -w $ARG2$ -c $ARG3$ -l 'Power Supply status'
       }
define command{
       command_name    rmc_check_hptemp
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.4 -w $ARG2$ -c $ARG3$ -l 'Temprature status'
       }
define command{
       command_name    rmc_check_hppower_4100
       command_line    $USER1$/check_snmp -H $HOSTADDRESS$ -C $ARG1$ -o .1.3.6.1.4.1.11.2.14.11.1.2.6.1.4.3 -w $ARG2$ -c $ARG3$ -l 'Power Supply status'
       }

 

Ric, i'd be interested to know where you put those commands.... as i'm looking to do the same

 

 

Dan