HP-UX SNAplus2 R7 Administration Guide

Appendix DUsing SNAplus2 in a High Availability Environment
Identifies which SNAplus2 LS is being monitored.
interval
Specifies the number of seconds that snapmon waits between attempts to obtain the status of the LS. If this
parameter is not specified, snapmon will pause 5 seconds between queries. Any number between 1 and
3600 (inclusive) can be specified.
retry_count
Specifies how many times snapmon will allow the LS to be reported in a state other than active when
snapmon is starting. This option is useful if the LS is configured to be initially active, and the SNAplus2
control daemon, node, port, LS, and snapmon are all started by ServiceGuard. It allows the LS enough time
to establish communications with the remote system and become active. If this parameter is not specified, 10
retries will be allowed. Any number between 0 and 600 (inclusive) can be specified.
The return code of snapmon is 0 if the LS was active at some time while snapmon was running. Otherwise, a
non-zero value is returned.
One snapmon Service Command will be listed in the Package Control Script for each SNAplus2 LS that should
be monitored. For example, to monitor an SNAplus2 LS called HALS, you might put the following line in the
Package Control Script:
SERVICE_CMD[0]="/opt/sna/bin/snapmon -i1 -r60 HALS"
When ServiceGuard starts the SNAplus2 package, it will also start the snapmon Service to monitor the package.
The snapmon utility will obtain the status of the HALS connection from SNAplus2 every second. snapmon will
wait 60 seconds for the LS to become active. If HALS fails to become active within 60 seconds, or if it activates
and then deactivates, the snapmon process will terminate. The termination of the process will signal ServiceGuard
that the SNAplus2 package is no longer providing SNA network connectivity, and the package should be migrated
to another server.
Once you have created the basic package, read this section to determine how to customize your SNAplus2 package.
You will need to modify the SNAplus2 Package Control Script to complete your SNAplus2 package definition.
D.3.5 Specifying a Package IP Address
An IP Address must be associated with the SNAplus2 package if you are using an IP based network for SNA
network connectivity. This is a ServiceGuard requirement. This address is called a "floating" IP address, because
whenever ServiceGuard migrates a package (or performs a local failover), the floating IP address moves with the
package. Since ServiceGuard moves the floating IP address with the package, associating an IP address with a
package provides a high degree of availability when access to a particular LAN adapter has been cut off.
ServiceGuard prevents TCP/IP connections from being disconnected when a local LAN card failover occurs.
Following are four different ways in which applications can gain SNA network connectivity through an SNAplus2
server using a LAN.
SNAplus2 Client/Server LAN Connections
As described above, SNAplus2 can use either a standalone or a client/server configuration. In a client/server
configuration, you run applications on client systems (HP 9000s or PCs) that access an SNAplus2 server through
the slim process running on the server system.
214