This section explains how to start and stop the package registered with the MC/ServiceGuard cluster system.
Starting the package
Start the package using the following procedure.
Start the Systemwalker Operation Manager daemon that has not been registered with MC/ServiceGuard on all the nodes in the cluster system.
# /opt/systemwalker/bin/soperationmgr
Start the Systemwalker Operation Manager package by specifying the node to use as the active node with the cmrunpkg MC/ServiceGuard command on either node. In the following example, "hp01" is specified as the active node and the package name is "omgr ".
# cmrunpkg -v -n hp01 omgr
For N:1 active/standby configurations, start the package "N" times, specifying each active node. For dual node mutual standby configurations, do this twice, once for each node.
Check the operation status of the package and each node by using the cmviewcl MC/ServiceGuard command on either node.
# cmviewcl
Make sure that the "STATUS" and "STATE" of the "PACKAGE" and each "NODE" are "up" and "running".
The following display example is for a 1:1 active/standby configuration.
CLUSTER STATUS cluster1 up NODE STATUS STATE hp01 up running hp02 up running PACKAGE STATUS STATE PKG_SWITCH NODE omgr up running disabled hp01
In this example, the switching status of the package is "disabled", so change the switching status of the Systemwalker Operation Manager package by using the cmmodpkg MC/ServiceGuard command on either node.
In the following example, the package name is "omgr".
# cmmodpkg -v -e omgr
Check the switching status of the Systemwalker Operation Manager package by using the cmviewcl MC/ServiceGuard command with the "-p" option on either node.
In the following example, the package name is "omgr".
# cmviewcl -v -p omgr
Make sure that "PKG_SWITCH" for the "PACKAGE" and "SWITCHING" for the "Primary" node and the "Alternate" node are all "enabled". The cluster will not fail over correctly if these are not "enabled".
The following display example is for a 1:1 active/standby configuration.
PACKAGE STATUS STATE PKG_SWITCH NODE omgr up running enabled hp01 (omitted) Node_Switching_Parameters: NODE_TYPE STATUS SWITCHING NAME Primary up enabled hp01 (current) Alternate up enabled hp02
MC/ServiceGuard packages usually operate in this state.
Information
If the package configuration is changed while the cluster is running, the "STATUS" and "STATE" of the cluster and package displayed with the cmviewcl command may become "down" or "halted" and the switching status may become "disabled". In such cases, check whether each node is running and check the connections to the shared disks for the cluster. Then switch the package status using the cmmodpkg command. Refer to "Switching the package status" below for details.
Stopping the package
Stop the package by executing the cmhaltpkg MC/ServiceGuard command on either node. In the following example, the package name is "omgr".
# cmhaltpkg -v omgr
For N:1 active/standby configurations, stop the package "N" times, specifying the package name. For dual node mutual standby configurations, do this twice, once for each node.
Switching the package status
If the switching status of the package is "disabled", change the status of the package by specifying the package name with the cmmodpkg command. In the following example, the package name is "omgr".
# cmmodpkg -v -e omgr
If the switching status of a node is "disabled", change the status by specifying the "disabled" node and the package name with the cmmodpkg command. In the following example, the name of the "disabled" node is "hp01" and the package name is "omgr".
# cmmodpkg -v -e -n hp01 omgr