ServerNet Cluster Manual

ServerNet Cluster Description
ServerNet Cluster Manual—520575-003
1-35
SNETMON and the ServerNet Cluster Subsystem
SNETMON Fault Tolerance
For fault tolerance, the SNETMON process pair can be configured to run on a list of
processors. For example, in systems having eight processors, HP recommends that
you configure SNETMON to run in the following CPU list: 2, 5, 6, 3, 7, 4.
The persistence manager ($ZPM) starts the primary SNETMON process on the first
processor in the list (processor 2), and the primary chooses the next processor in the
list (processor 5) to start the backup. A failure in the primary processor (processor 2)
causes a takeover. Processor 5 then becomes the primary, and a new backup
SNETMON is created in the next processor in the list (processor 6). In this way,
SNETMON can survive a variety of outages.
Configuring SNETMON is described in more detail in Section 3, Installing and
Configuring a ServerNet Cluster.
As an alternative, you can use The ZPMCONF macro to automate the process of
adding MSGMON, SANMAN, and SNETMON to the system-configuration database.
For a description of the macro, refer to Appendix E, TACL Macro for Configuring
MSGMON, SANMAN, and SNETMON.
Figure 1-19. ServerNet Cluster Logical Diagram
SNetMon
TSM
(RAL)
SCF
EMS
Collectors
($0, $ZLOG)
NRT
NCP
Service
Processor
MsgMon
Message System
Events
SPI
SPI
Events
Status/
Commands
One Expand-Over-
ServerNet Line-Handler
Process for Each
Remote System
NAM
NAM
Line
Handler
Line
Handler
SANMAN
Application
SPI
File System
VST012.vsd