The following explains how to change the node name after building a PRIMECLUSTER system.
Note
Changing a node name may have a serious impact on the system. Therefore, make this change only when it is absolutely necessary.
Use the "sys-unconfig(1M)" command to change a node name. The "sys-unconfig(1M)" command is used to reset the system. All settings such as the name service, time zone, IP address, IP subnet mask, and root password are all erased and are then reset when the node next boots.
See
For details on the "sys-unconfig(1M)" command, see the manual page describing sys-unconfig(1M).
Operation Procedure:
Stop CF on the node whose node name is to be changed.
For details on how to stop CF, see "4.7 Starting and stopping CF" in "PRIMECLUSTER Cluster Foundation (CF) Configuration and Administration Guide."
Set up the shutdown facility.
When changing the node name assigned to the IP address of the Administrative LAN that is to be used in the shutdown facility, perform the following according to the definition of the /etc/opt/SMAW/SMAWsf/rcsd.cfg file.
When the IP address is defined in "admIP" of the rcsd.cfg file
You do not need to change the rcsd.cfg file.
When the node name registered in the hosts(4) file is defined "admIP" of the rcsd.cfg file
Execute the following command on each cluster node to stop the shutdown facility.
# sdtool -e
Define the node name to be changed to "admIP" of the /etc/opt/SMAW/SMAWsf/rcsd.cfg file.
See
For details on the rcsd.cfg file, see the manual page of rcsd.cfg(4M).
Define "export HV_RCSTART=0" in the hvenv.local file of the node whose node name is to be changed.
This disables automatic startup RMS when the node is booted after changing the node name.
For details on the hvenv.local file and the HV_RCSTART attribute, see the manual page describing hvenv.local.
Change the node name.
For Solaris 10, the "sys-unconfig(1M)" command is used. For Solaris 11, the "sysconfig(1M)" command is used. After the node names are changed, edit the hosts(4) file if necessary. For details, see Oracle Solaris documents.
Note
Do not attempt to change information other than node names.
Restart the system.
# /usr/sbin/shutdown -y -i6 -g0
Start RMS.
For information on how to start up RMS, see "7.1.1 Starting RMS" in "PRIMECLUSTER Reliant Monitor Services (RMS) with Wizard Tools Configuration and Administration Guide."
After confirming that RMS has started up, restore the definition of the HV_RCSTART environment variable corrected in Step 3.