High Availability

[deepavaidya]

Hi,

We are looking for a high availability (failover) solution for Nagios XI. We have 2 sites to be monitored by Nagios XI and the filtered trap messages need to be sent to 3rd party monitoring tool. Can you please suggest what will be the best high availability solution to monitor site A and B using Nagios XI.

[tmcdonald]

Have you taken a look at our High Availability Options documentation?

[deepavaidya]

We have installed Nagios XI in CentOS 6. I want to have a master and slave setup where the salve just monitors the master and takes over the master if master is down. Should this setup have a shared database ? How do we syncronize their data ?

[sreinhardt]

Presently a shared db is not possible. Most likely the best option would be to forward checks from your main installation to the secondary, this will cause all results and perfdata to be shared between them. Additionally you can do either freshness checking or find a way to enable active checks if the main system goes down, on the secondary to enforce failover. By forwarding checks, you would also ensure that any services that need to be added to the secondary system would be in unconfigured objects, making adding them especially easy.

[deepavaidya]

Hi

I just went through this PPT http://www.slideshare.net/nagiosinc/mike-weber-failover

I have few doubts regarding this.

1) Will all the configurations (creating user, creating template) also be replicated from master to slave ? (service and Host check transferred using NSCA)

2) When the Master comes up again, will the configurations and history be synced from slave to master ?

3) If I have a global manager nagios console to display the alerts from separate nagios monitoring individual site, will there be separate license for the global manager though it does not monitor any.

[deepavaidya]

Hi
For the solution discussed in my previous post, where the master shares the service & host check history with the slave to avoid double polling, how does the SNMP configuration be for the monitored hosts, should they send traps to both the nagios (master & slave) everytime ? Can you also please give me some in-depth technical details of how the DB replication would work in this case ?

[deepavaidya]

Please can you also let me know how the end user can be prevented from knowing about the failover, I mean how will the url be dynamically changed. Please can you suggest the best possible solution for this also.

[abrist]

1) Will all the configurations (creating user, creating template) also be replicated from master to slave ? (service and Host check transferred using NSCA)

Yes, at the point you make teh slave. Additional configs should be updated using weber's method, the postgres may not be.

2) When the Master comes up again, will the configurations and history be synced from slave to master ?

Puppet or an offloaded mysql database would handle config replication. Hostory should be saved with rsync or again with an offloaded db.

3) If I have a global manager nagios console to display the alerts from separate nagios monitoring individual site, will there be separate license for the global manager though it does not monitor any.

Yes. You should look at nagios fusion.

should they send traps to both the nagios (master & slave) everytime ?

Yes. Any passive checks need to report to both server ips.

Can you also please give me some in-depth technical details of how the DB replication would work in this case ?

I am not quite sure what you are looking for here. We provide a document for offloading the db, so you could just move the mysql databases to another server cluster. With postgres or a non-offloaded mysql solution, you need to backup the databases periodically to update them on the slave/master.
http://assets.nagios.com/downloads/nagi ... Server.pdf

Please can you also let me know how the end user can be prevented from knowing about the failover, I mean how will the url be dynamically changed.

This is essentially dns/routing/apache question. It all depends on the environment. You can control it with routing, though dynamic dns may be a better option. Do you have internal dsn servers in your environment?

[deepavaidya]

Hi,

Thank you so much for your answers. Please can let me know if there is any document available to follow this approach apart from the PPT ? Is the NSCA which is mentioned is already installed with Nagios XI only needs to be configured ?

[slansing]

Beyond that slide-show we do not have further documentation on this, that was from a conference attendee at our Wold Conference. NSCA is installed on all XI systems by default and it need only be configured to forward checks:

http://assets.nagios.com/downloads/nagi ... ith_XI.pdf
http://assets.nagios.com/downloads/nagi ... ith_XI.pdf