![]() |
|||
![]() |
![]() ![]() |
![]() |
![]() ![]() |
![]() |
![]() ![]() |
![]() |
| ||||||||||||||||||
Chapter 8Deploying Your Software Load to a ClusterThis chapter describes how to create the Solaris JumpStart environment for target cluster. You are then ready to deploy the software load by using Solaris JumpStart. This chapter contains the following sections: Software Load Deployment ProcedureThe following figure shows the sequence of events that deploy a software load to a cluster: Figure 8-1 Software Load Deployment Procedure ![]() Verifying that a Flash Archive is Configured CorrectlyThis section describes how to use the flar command to verify that the flash archive is correctly configured.
|
# flar -i flash-archive |
The flar command displays information that is stored in the header of a flash archive. This information includes the name and version of the software load and the target node group. The following table gives an example of information in the header of a flash archive:
Table 8-1 Example of Information in the Header of a Flash Archive
Header Information | Description |
---|---|
X-NHAS-SWL-ID=1 | The software load version |
X-NHAS-SWL-NAME=12N | Software load name |
X-NHAS-NODE-GROUP-NAME=master_el | Target node group |
X-NHAS-SOE-RELEASE=Solaris-9-s9_58shwpl3-SPARC | Solaris operating system associated with the flash archive. |
For more information, see the flar(1M) man page.
You can configure a direct link between the master-eligible nodes to prevent a split brain situation, where there are two master nodes in the cluster because the cluster network fails.
Connect the serial ports of the master-eligible nodes.
For an illustration of the connection, see the Netra High Availability Suite Foundation Services 2.1 6/03 Hardware Guide.
Configure the direct link in the machine.conf configuration file, for example:
ELEMENT board T1105@peerNode1 type T1_105 arch SPARC USE nic nic0@peerNode1 nic nic1@peerNode1 nic cgtp@peerNode1 link serial-b |
Before the software load can be deployed, each diskless node must be configured according to the boot policy that was defined in configuration stage 1.
Get the ok prompt on the diskless node.
For example, if you are connected to the console with a terminal server using telnet, run:
telnet> send brk ok> |
Configure the OBP parameters:
ok> setenv boot-device net:dhcp,,,,,5 net2:dhcp,,,,,5 |
In the machine.conf file, add the client ID in hexadecimal for the board associated to the diskless node.
You can calculate the client ID as follows:
Set the board ID at the OBP level
Convert it to hexadecimal
Add two zeros (00) as a prefix
For example, the ID configured at the OBP level is 11, therefore, the client ID in hexadecimal in the machine.conf file is 003131:
ELEMENT board CP2160@peerNode3 type CP2160 arch SPARC clientId 003131 CONTAIN nic nic0@peerNode3 nic nic1@peerNode3 nic cgtp@peerNode3 |
Get the ok prompt on the diskless node.
For example, if you are connected to the console with a terminal server using telnet:
telnet> send brk ok> |
Configure the OBP parameter dhcp-clientid:
ok> setenv dhcp-clientid client-id-value |
Before you boot the cluster, verify that the installation server directories are mounted on the prototype machine and the cluster. For more information, see To Mount the Installation Server Directories Onto the Prototype Machine and the Cluster.
This section describes how to deploy the deployable flash archives to a cluster by using the Solaris JumpStart environment and a boot net installation.
The boot net installation installs and starts the deployable flash archives on the cluster.
This procedure boots all of the master-eligible node and dataless nodes at the same time.
Get the ok prompt on each master-eligible node and dataless node.
Perform a boot net install:
ok> boot net - install |
The master-eligible nodes and dataless nodes perform a Solaris JumpStart from the installation server. These nodes then install and run the software load.
![]() ![]() |