The following steps must be performed on each additional OTP host in your clustered OTP system.
The OTP high availability framework must be set up and configured on the first OTP host as described in To Set Up the OTP High Availability Framework on the First OTP Host
The additional OTP host must be added to the clustered OTP system as described in To Add Additional OTP Hosts to the Clustered OTP System
Log in as root (su - root) to the OTP host.
Create the additional OTP host high availability state file.
Edit the /var/tmp/setupAvailabilityServiceOnOtherHosts.dat as follows.
mediaDirectory=OTP_Installation_directory: The fully qualified path name of the Open Telecommunications Platform installation directory, for example: /net/otpinstall/otp1.0.
sponsorNode=first OTP host: The name of the first OTP host.
privateInterface1=Ethernet interface 2 where Ethernet interface 2 is bge2 or ce2 depending on platform type.
privateInterface2=Ethernet interface 3 where Ethernet interface 3 is bge3 or ce3 depending on platform type.
transportTypeInterface1=dlpi
transportTypeInterface2=dlpi
quorumAutoConfiguration=yes: Required by the OTP high availability framework
autoConfigureIPMP:
autoConfigureIPMP=no if you do not want to set up IPMP.
To set up IPMP, set autoConfigureIPMP=yes, and add the following three lines:
secondaryInterface=Ethernet interface 2
secondaryIP=111.112.113.114 where 111.112.113.114 is the IP address of Ethernet interface 2.
testIPAddress=111.112.113.222 where 111.112.113.222 is the IP address used for IPMP configuration.
allPatches=yes to install all patches.
allPatches=no to install mandatory patches only.
Save and close the file.
The setup additional OTP hosts high availability state file should be similar to the following:
mediaDirectory=/net/otpinstall/otp1.0 sponsorNode=otpclient1 privateInterface1=bge2 privateInterface2=bge3 transportTypeInterface1=dlpi transportTypeInterface2=dlpi quorumAutoConfiguration=yes autoConfigureIPMP=yes secondaryInterface=bge1 secondaryIP=10.11.55.174 testIPAddress=10.11.55.175 allPatches=yes
Set up the OTP high availability framework.
Type /opt/SUNWotp10/CLI/setupAvailabilityServiceOnOtherHosts /var/tmp/setupAvailabilityServiceOnOtherHosts.dat
The setupAvailabilityServiceOnOtherHosts script performs the following tasks:
Installs required Solaris OS patches
Installs and configures the OTP high availability framework
Reboots the OTP host
The setupAvailabilityServiceOnOtherHosts installation process logs to the file /var/tmp/OTPInstaller.log. You can use the tail -f command during installation to view the log file.
When the first OTP host has rebooted, log in as root again.
Run the setupAvailabilityServiceOnOtherHosts script again.
Type /opt/SUNWotp10/CLI/setupAvailabilityServiceOnOtherHosts /var/tmp/setupAvailabilityServiceOnOtherHosts.dat
The setupAvailabilityServiceOnOtherHosts script verifies the OTP high availability framework installation and configuration.
If you set quorumAutoConfiguration=no on a two-host cluster, you must manually select and configure the quorum disk after cluster configuration is complete as follows.
The following sub-steps apply only to a two-host cluster. If you are setting up the OTP high availability framework on a three-host or more clustered OTP system, this step is optional.
Open a separate terminal window and log in as root to the first OTP host.
Type /usr/cluster/bin/scdidadm -L to display the cluster disk information. For example:
# /usr/cluster/bin/scdidadm -L 1 otpclient1:/dev/rdsk/c0t8d0 /dev/did/rdsk/d1 1 otpclient2:/dev/rdsk/c0t8d0 /dev/did/rdsk/d1 2 otpclient1:/dev/rdsk/c0t9d0 /dev/did/rdsk/d2 2 otpclient2:/dev/rdsk/c0t9d0 /dev/did/rdsk/d2 3 otpclient1:/dev/rdsk/c1t0d0 /dev/did/rdsk/d3 4 otpclient1:/dev/rdsk/c1t1d0 /dev/did/rdsk/d4 5 otpclient2:/dev/rdsk/c1t0d0 /dev/did/rdsk/d5 6 otpclient2:/dev/rdsk/c1t1d0 /dev/did/rdsk/d6 |
In the above example, disks d1 and d2 are shared by both hosts of the two-host cluster. The quorum disk must be a shared disk.
Configure a quorum disk.
Type /usr/cluster/bin/scconf -a -q globaldev=shared disk ID where shared disk ID is a shared disk ID. For example:
# /usr/cluster/bin/scconf -a -q globaldev=d1 |
Type /usr/cluster/bin/scconf -c -q reset to reset the two-host cluster to normal mode.
Create the clustered OTP system shared storage as described in the next procedure.