Planning and Configuring HP-UX DCE 1.9

Chapter 6
Configuring HP-UX DCE Cells
Integrating DCE Services with MC/ServiceGuard
69
An alternative configuration is the individual core server failover. In this configuration, the DCE cluster
includes a primary host and several smaller hosts. Each of the smaller hosts can perform one or more of the
functions normally performed by the primary host. In the event of a system failure for the primary host,
ServiceGuard moves only the core servers on the smaller hosts from their individual physical nodes to other
physical nodes. For example, in a DCE cluster with individual server hosts for the Security Service and the
Name Service, ServiceGuard can move the secd or cdsd servers individually to a new host rather than
moving the entire primary host to another node.
Another alternative configuration is the redesignation of master servers through ServiceGuard scripts. This
approach involves changing the role of the existing replicas, making them the master servers, instead of
moving the databases to new locations.
Other alternative configurations are possible, but they are not discussed here. Refer Managing
MC/ServiceGuard (B3936-90003) for more information about alternative configurations.
The remaining sections in this chapter describe the basic configuration only.
The DCE-SGUARD template fileset is available on the CD-ROM that ships with the DCE product. See Table
3-2 in Chapter 3 and “Supported Templates for MC/ServiceGuard Integration with DCE” on page 69" for more
information about filesets.
Supported Templates for MC/ServiceGuard Integration with DCE
As part of the DCE product, HP-UX DCE 1.9 provides a fileset (DCE-SGUARD) that contains a set of
customizable ServiceGuard templates and scripts to integrate MC/ServiceGuard with DCE services. This set
of templates includes the DCE processes dced, cdsadv, secd, and cdsd within a single package. Each process
is monitored separately, and local restart and local network switchover are supported.
NOTE Do not attempt to use the templates in an installation without first customizing them.
The templates and scripts are as follows:
dcepkg.conf—package definition file generated by MC/ServiceGuard. You must supply the ServiceGuard
node names, DCE services, and the network name and address.
rc.dcepkg—package control script generated by MC/ServiceGuard. You must supply the ServiceGuard
node names, DCE services, and the network name and address. You can also specify customer defined
functions.
dce—DCE configuration file, which is copied from /etc/rc.config.d/dce to the DCE/ServiceGuard
package directory. You can set the environment variable RPC_SUPPORTED_NETADDRS to the package
IP address.
dce.start—package startup script, which MC/ServiceGuard calls when the package is activated.
dce.restart—package restart script, which the DCE service monitor uses to start an individual DCE
service.
dce.monitor—DCE Service monitor, which ServiceGuard launches to monitor DCE daemons.
First, the DCE service monitor checks to see if the server is running, and, if it is not, the DCE service
monitor starts it. Then, the DCE service monitor goes into a loop and checks to ensure that the server
process is running. Finally, the DCE service monitor performs a DCE level ping on the server interface.
rc.dcepkg.log—a log file that is created in the package directory during DCE package startup through
MC/ServiceGuard, individual DCE services restart upon failure, and package failover.