Managing Serviceguard 11th Edition, Version A.11.16, Second Printing June 2004

Planning and Documenting an HA Cluster
Cluster Configuration Planning
Chapter 4154
Specify three things for each policy: USER_NAME,
USER_HOST, and USER_ROLE. For Serviceguard
Manager, USER_HOST must be the name of the
Session node. Policies set in the configuration file of a
cluster and its packages must not be conflicting or
redundant. For more information, see “Editing
Security Files” on page 182.
FAILOVER_OPTIMIZATION
You will only see this parameter if you have installed
Serviceguard Extension for Faster Failover, a
separately purchased product. You enable the product
by setting this parameter to TWO_NODE. Default is
disabled, set to NONE. For more information about the
product and its cluster configuration requirements, go
to http://www.docs.hp.com/hpux/ha and click
Serviceguard Extension for Faster Failover.
NETWORK_FAILURE_DETECTION
When there is a primary and a standby network card,
Serviceguard needs to determine when a card has
failed, so it knows whether to fail traffic over to the
other card. To detect failures, Serviceguard’s Network
Manager monitors both inbound and outbound traffic.
The Manager will mark the card DOWN and begin to
attempt a failover when network traffic is not noticed
for a time. (Serviceguard calculates the time depending
on the type of LAN card.)
The configuration file specifies one of two ways to
decide when the network interface card has failed:
• INOUT - The default method will count inbound and
outbound failures separately, and declare a card
down only when both have reached a critical level.
• INONLY_OR_INOUT - This option combines the
inbound and outbound failure counts, and will
declare a card down when the total failures reach a
critical amount, regardless of their source. With
this method, Serviceguard tries to validate inbound
failure reports by doing additional remote polling.
The default is INOUT.