I am working with Nagios 4.0.7 on two servers to monitor my network. Recently I developed the third one in order to have more redundancy so I start reading information about failover redundancy mostly with NSCA. Has anyone a little bit of experience in this domain that could tell me which is the most efficient and easy way to do it. I read this slides that already helped me a lot but of course the issue is still the time synchronization between the master and the slaves:


Is NSCA the best option?Is failover, redundancy or the use of both the best choice?

  • I would also like to see this.. hope someone have a good answer to this.. if it exist. – Sverre Jul 31 '14 at 9:35
  • Also I have never worked with Nagios XI, is it worth it for this issue? – Javier Suárez Jul 31 '14 at 9:52

There is some great documentation available on Nagions Master/Slave setups at http://nagios.sourceforge.net/download/contrib/documentation/misc/HighAvailability/NagiosHA_EN.pdf

In my use case, I used a slightly different approach that I found easier to setup and recover in the event of a failure. In DC 1, there is a VM running Nagios that checks everything in DC 1 & 2. It is setup using DRBD to replicate to a hot-standby VM and also 3rd slave in DC 2. The reason I used DRBD is because it replicates the entire block device (even across data centers via VPN) so recovering is a snap. It also made the configuration really simple and didn't require a complicated configuration in Nagios. There are a number of ways you can automate fail over of a downed node, like keepalived, corosync, heartbeat if you want it totally automated.

In DC 2, we have the exact same setup, only the reverse.

My environment is not that big (450 servers), so this setup has worked well. In much larger environments, it may require distributing the monitoring across Nagios nodes to meet acceptable performance goals.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.