Failover device monitoring?

Tedman Eng teng at dataway.com
Thu Oct 2 05:20:28 CEST 2003


Hi folks,

I've had Nagios deployed for quite some time now and am fairly familiar with
most configuration scenarios.  However, I'm stumped trying to figure a
solution to one particular configuration dilemma:

Our company monitors a cluster of High Availability services.  At any one
time, only one service out of the cluster is "OK", the others are down but
waiting to take over if there's a problem.

Cluster is OK if      Service A - OK,      Service B - down,      Service
C - down
Cluster is OK if      Service A - down,      Service B - OK,      Service
C - down
Cluster is Critical if      Service A - down,      Service B - down,
Service C - down

The check_cluster plugin works well (see
http://nagios.sourceforge.net/docs/1_0/clusters.html ), since it checks to
make sure at least one of the services is up.  We don't actually care to
about the state of the individual services, so long as the Cluster checks
out OK, so we turn off notifications for the individual services .  In
addition to not being notified, is there a way to "hide" those services
somehow so that they don't show as always down in the summary screens?

One possible solution I considered was moving the individual and cluster
checks to a separate Nagios machine, and then send only the cluster results
back to the main Nagios.  This effectively hides everything but the cluster
from the main Nagios, but also adds another machine and another Nagios to
have to administer.  Anyone else monitoring "HA" or failover devices?  How
do you do it?







-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null





More information about the Users mailing list