Introducing ndo2fs

You never heard of ndo2fs? I’m sure you know ndo2db which is part of the Nagios addon ndoutils. In short ndo2fs is a lightweight alternative to ndo2db. Both of them, ndo2fs and ndo2db, gather information which are present in Nagios core and save them in an ordered way. Ndo2fs receives information from ndomod.o via TCP- or UNIX-socket and stores the data formated in JSON as plain text files in a hierarchical folder structure.

The big benefits to ndo2fs are:

• Ndo2fs is extremely smart. Ndo2fs is a perl script with about 1200 lines of code.
• The needed space for current data is extremely small.
• You don’t need a whole DBMS for storing the data.
• Ndo2fs consumes much less cpu time and memory than ndo2db and the used DBMS together.

At the moment ndo2fs is trimmed and optimized to present the current configuration and state of your Nagios installation. There is no focus on saving historical data like state history or trends. This is optimal for using with NagVis.

The ndo2fs script can be run attached and as a daemon. At the moment there is no init script for ndo2fs but that is no drawback compared to ndo2db cause the init script delivered with ndoutils is not very generic (only valid for SuSE distribuions). I’m sure we will get an init-script for ndo2fs soon.

Ndo2fs is developed in a GIT repository on Nagiosprojects.org. You can browse it at this URL and check it out with this command:

git clone git://git.nagiosprojects.org/ndo2fs.git

There is a nearly complete backend included in the current release of NagVis 1.4 (rc1). Only the correct handling of downtimes is missing there. This will be included in NagVis 1.4rc2.
I had a lot of fun computing this backend – handling the data is much easier than SQL-Queries agains the database of ndo2db. First performance tests on small development systems showed of the performance of NagVis using the ndo2fs backend and data is faster than using the standard ndo2db way with MySQL. I will test this on bigger systems with real server hardware the next days.