Automating Failover Event Handling on Two HP Systems Insight Manager Servers White Paper 532451-002, February 2010

2
Introduction
In very large environments, one HP Sy
stems Insight Manager (HP SIM)
might
not provide all the
coverage needed for all the devices in disaster recovery.
You might want
two HP
SIM servers
covering each other. In this scenario,
you can
create a set of E
vent
H
andlers that
allow two HP SIM
servers
to watch each other and take over coverage for the other’s Event Notification in the event of
failure of one
central management server (
CMS
)
.
The
white paper
describes how to create an Event Handler using the CMS
c
ommand
-
l
ine tool
mxtask
, one of many CLI
tools that come with HP
SIM
and are documented in the Information Library
at:
http://h18013.www1.hp.com/products/servers/management/hpsim/infolibrary.html
.
The goa
l is to have two HP SIM servers that manage Event Notification for local devices, say, City A
and City B. In normal operation, CMS A manages only City A servers and CMS B manages only City
B servers. In some situation, the HP SIM server in City A fails,
yo
u want
City B
to
automatically detect
this and run a task that would enable it to handle alerts on City A server
and
its own population. The
basic configuration is that each CMS has a full population of all devices being monitored, both from
City A and B.
This
white paper
describes a method to achieve that failover.
The
basic
steps are
as follows
:
1.
Create two Event Collections using the customize lists option
:
one for City A events
and one
for
City B events.
2.
Create an Event Handlers on Server A and B that wa
tch each other and run a custom task that
enables a new Event Handler for the other cities devices when the CMS
fails
.
3.
Create an Event Handlers on Server A and B that watch each other and run a custom task that
removes the event handler for the other citie
s devices when the CMS comes back online.
Server
A processes events on servers in A’s territory,
Server
B processes events from servers in B’s
territory. A Watch Dog Event Handlers at each
server
watches for the other
server
to become
reachable or unreacha
ble and creates or deletes event handler tasks.