The information assume that you have installed nav in /usr/local/nav.

The information here is under construction. It is possibly erroneous and for a fact incomplete.

The default installation of Cricket in NAV-v3 gives you statistics for your routers and switches and their interfaces. However, it may be desirable to add local statistics to the installation of Cricket. This document will lead you through a detailed way to do this without installing a new instance of Cricket. This document will not help you if you do not know anything about Cricket. For information about how to configure Cricket, see the documentation at SourceForge

Adding your local statistics to Cricket may be done in three (or four) steps:

• Create a directory to store the configuration.
• Create the configuration you wish in the newly created directory.
• Edit the subtree-sets so that Cricket knows that this directory is active.
• Alternatively: Add a new cron-job to collect data in another interval than 5 minutes.

The Cricket config-tree is located at /usr/local/nav/cricket/cricket-config.

The basic structure of a NAV-v3 install is the following directories:

• giga-router-interfaces
• giga-switch-ports
• router-interfaces
• routers
• servers
• switch-ports
• switches

These directories are configured automatically, and the files here should not be altered.

However, any other directories you create will not be affected by the automatic processes in NAV-v3. This is thus where you start when making your local additions.

Create the directory you want to store your configuration in here.

	> cd /usr/local/nav/cricket/cricket-config
	> mkdir localadditions

Make sure that directory has correct permissions. At least it should be a member of the nav-group and that group should have read and execute-permissions.

This document does not cover how to configure Cricket. The Cricket-documentation is located at SourceForge and should give you all the information you need. Remember to compile the new config-tree when you are done making the configuration.

	> /usr/local/nav/cricket/cricket/compile

Cricket gathers data using cron. The cron-daemon executes the collector-script located in /usr/local/nav/cricket/cricket. To ease the configuration of the cron-daemon, the file “subtree-sets” is used to group directories into specific “sets”. To collect data from a “set” the collect-subtree script is called with the name of the set as argument.

NAV-v3 uses two different sets when gathering data - “normal” and “giga”. The “normal” set runs every five minute (which is default in Cricket) and the “giga” set runs every minute (see also “Coming updates in NAV” below). The reason giga runs every minute is to reduce the chance that the 32-bits counters wrap on giga-interfaces. This method is not perfect, but mitigates the threat of a wrap.

To collect data from your local configuration you only need to edit the subtree-sets file and add your directory in the “normal” set. This will ensure that data is gathered every five minute.

	set normal:
            /routers
            /switches
            /switch-ports
            /router-interfaces
            /servers
            /localadditions

If you want to collect data in another interval you will have to make another set. In addition you will need to make a new cron-job so that Cricket gathers data (covered in the next section).

Adding a new cron-job may be done in several ways:

1. Adding a cron-job to an existing user. If the user is a member of the nav-group, this user can run the collect-subtrees script with the intended interval. Remember to use the set you made as argument to collect-subtrees. The cron-job should look something like this (assuming 10 minute intervals):

	CRICKETDIR=/usr/local/nav/cricket
	*/10 * * * * $CRICKETDIR/cricket/collect-subtrees local

2. Adding a new cron-job in /usr/local/nav/etc/cron.d/cricket and restart cricket using /usr/local/nav/bin/nav. This is not recommended because this file may be either overwritten or not changed when upgrading to a new version of NAV.

3. Make a new cron-fragment in /usr/local/nav/etc/cron.d. This will ensure that you can start/stop collecting using the nav-script. Make sure that you create a fragment that is not likely to be overwritten, i.e. the name should reflect that this is a local addition. This because if a future release of NAV has a fragment with the same name, complications may occur.

The current directory-structure with giga-switch-ports and giga-router-interfaces in addition to the other directories is not an optimal structure. When you search for a graph you do not want to have to search in multiple places to find what you are looking for.

A coming update in NAV plans to solve this problem. It will get rid of the giga-directories and put all information in one directory for switch-ports and one for router-interfaces.

The reason this has not been done before is that you need to use SNMP-v2 to fulfill this, and the current set of network-equipment does not always update the SNMP-v2 counters according to the standard. We plan to develop a test for this and put the ones that follow the standard in the “normal” directories, and the others in another directory. Hopefully the number of ports/interfaces that go in the other directory will decrease as time goes by and they start to update the counters correctly.