Applicable servicedependency definition for cluster services

Ralph.Grothe at itdz-berlin.de Ralph.Grothe at itdz-berlin.de
Mon Dec 11 15:35:29 CET 2006


Dear Nagios Experts,

I am currently deliberating over sensible service dependency
definitions for clustered applications
in order to reduce the amount of emitted notifications in case of
a failover or outage.

For a start I want to keep it simple and take one of our less
sophisticated clusters
which is merely a two node active/standby Linux Heartbeat cluster
that hosts a webserver.

Let's assume that NodeA and NodeB be the hostnames, VIP the
designator for the virtual IP address of the webserver,
and Apache the name of this HA service.

I already defined a service that I called Heartbeat which is
monitored by my Nagios server
for all of our Heartbeat clusters (i.e. those belongig to
hostgroup heartbeat_hosts).
The related check_command only verifies that the Heartbeat
service on both NodeA and NodeB is still running.

If for some reason Heartbeat is stopped on NodeA where Apache is
currently servicing clients,
Apache is also torn down on NodeA, and given that Heartbeat is
still running on the standby node NodeB,
it is attempted to being brought up on NodeB, viz. failed over.

Thus a notification of a failed Heartbeat should suffice I
beleive under these circumstances.

I then would start with some dependency definition like this



define servicedependency {
	dependant_host			VIP
	host_name				NodeA, NodeB
	dependant_service_description Apache
	service_description		Heartbeat
	execution_failure_criteria	c,u
	notification_failure_criteria	c,u
}




The pre-flight check (i.e. nagios -v ...) doesn't report any
errors.
Yet, I am not quite convinced if this will work as I intended?

Are there any sample configurations for clustered services and
dependencies,
also for the more sophisticated active/active clusters whith
multiple interdependencies
between clustered packages' or service groups' services,
because I have quite a few HP MC/ServiceGuard and Veritas Cluster
Servers with lots of applications
to monitor?
(n.b. the VCS clusters even allow for traps to be generated that
could be processed by some passive checks
I would assume)

Regards

Ralph

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
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