[Nagiosplug-help] Help me with FAIL OVER REDUNDANCY

Ricardo David Martins ricardo.martins at gmail.com
Fri Aug 27 09:55:08 CEST 2004

Where on Ethan Galstad documentation it refers "fail over redundancy",
it is mentioned making a ocds command to invoke the changes to the slave host. 

Well, if this is done, wont the slave lose synchronization if it goes
down? How to make this to work without ever losing synchronization? I
thought of mysql replication with the master and slave machines of
nagios working both like master/slave and slave/master to always
having both reporting changes to each other. No need to use ocds
commands. Altough Nagios slave shouldnt be able to write do the mysql
data base. The only way i see this hapenning is not to have nagios
slave started and have a cron job (mentioned on the fail over
"additional issues") that checks if nagios process is working on the
master and launches nagios slave if it isnt.

This will work fine i think but the problem is synchronization again.
Does Nagios proactively read the mysql data base and proceed to
actions based on what has changed?????????????????????   If it doesnt,
we must guarantee that the master (not the slave, since its down while
master is up, so no problem there) is only launched after the mysql
data base has synchronized with the one that is up (the slave). After
synch it launches the master. The slave has an "ok master nagios
process" and goes down again.

Ricardo David Martins

More information about the Help mailing list