 |
» |
|
|
|
The next step in the process is to configure the system. The cluster_config utility assigns roles to all nodes in your system based on the size of the system (and other considerations). A node role is defined by the services provided by the node. A role is an abstraction that groups one or more services. You have the option to modify this default configuration. When you are satisfied with the node role assignments, the configuration procedure builds the system configuration database, installs LSF (either LSF-HPC with SLURM or standard LSF), installs and configures SLURM (if LSF-HPC with SLURM is chosen), enables Nagios, starts all services on the head node and other service nodes, and creates the golden image from which all nodes are replicated. Follow this procedure to configure the system: Enter the following command to back up the system configuration database to a file in the current directory: # managedb backup ./filename |
You can save the backup file to a directory of your choice, and you may want to add a date and time stamp to the file name to easily determine when the backup file was created. Enter the following command to capture command output into a file; this step is optional: Change to the configuration directory: # cd /opt/hptc/config/sbin |
Begin the cluster configuration process: When the following command-line menu options are displayed, enter the letter h to display a description of each option; entries are not case sensitive: Cluster Configuration - XC Cluster version HP XC V3.0 <kit> <date>
[L]ist Nodes, [M]odify Nodes, [A]nalyze, [H]elp, [P]roceed, [Q]uit: h
This utility enables you to modify node role assignments
and define external network connections on nodes other than the head node.
Valid commands are:
L - Lists discovered nodes with current location, assigned roles,
and external network connections
M - Modifies nodes with the Modify Node submenu
A - Analyze current configuration
H - Displays this help message
P - Proceeds to the next cluster configuration step
Q - Quits this utility leaving the cluster unmodified
[L]ist Nodes, [M]odify Nodes, [C]heck, [H]elp, [P]roceed, [Q]uit: |
HP recommends that you enter the letter l to list the default configuration; then enter the letter b or the letter a to view an abbreviated or full version of the output: [L]ist Nodes, [M]odify Nodes, [A]nalyze, [H]elp, [P]roceed, [Q]uit: l
Output Type - a[B]breviated, [A]ll, [C]ancel: |
|
The a[B]breviated output looks similar to the following. Nodes with the same role assignments are condensed into one line item. node: n16 location: Level 1 Switch 172.20.65.1, Port 42
CURRENT HEAD NODE
roles assigned: management_server management_hub external disk_io compute
External Ethernet Name: penguin.southpole.com, ipaddr: 192.0.2.0, netmask:
255.255.252.0, gateway: 192.0.2.51
node: n15 location: Level 1 Switch 172.20.65.1, Port 41
roles assigned: resource_management compute
The following nodes are configured with the common role "compute":
n[11-15] |
[A]ll output looks similar to the following and displays specific information about the head node and each client node. Be aware that output can be quite lengthy on a system with a large number of nodes. For example, the output for a 500 node system with 495 compute-only nodes results in approximately 1500 lines of output. |
node: n16 location: Level 1 Switch 172.20.65.1, Port 22
CURRENT HEAD NODE
roles assigned: compute disk_io external management_hub management_server
resource_management
External Ethernet Name: penguin.southpole.com, ipaddr: 192.0.2.0,
netmask: 255.255.252.0, gateway: 192.0.2.51
node: n15 location: Level 1 Switch 172.20.65.1, Port 21
roles assigned: compute
node: n14 location: Level 1 Switch 172.20.65.1, Port 20
roles assigned: compute
node: n13 location: Level 1 Switch 172.20.65.1, Port 19
roles assigned: compute
node: n12 location: Level 1 Switch 172.20.65.1, Port 18
roles assigned: compute
node: n11 location: Level 1 Switch 172.20.65.1, Port 17
roles assigned: compute
Continue? [y/n] y |
Do the following: See Chapter 8 for a description of the default role assignments based on system size. Also, read about the special considerations regarding the default role assignments to determine if they are suitable for your environment. If you do not modify the role assignments, your system will be configured with the default assignments. After reading this information, determine if you are satisfied with this configuration or if you want to modify role assignments. If necessary, refer to Appendix F, and use the Modify Nodes and Analyze options to adjust the role assignments or create an external Ethernet connection on any node. When you are satisfied with the role assignments and Ethernet connections, continue to step 8.
When you are satisfied with the current system configuration, enter the letter p to apply it to the system: [L]ist Nodes, [M]odify Nodes, [A]nalyze, [H]elp, [P]roceed, [Q]uit: p
Do you want to apply your changes to the cluster configuration? [y/n] y |
Do one of the following when this prompt is displayed: [S]ervices Config, [P]roceed, [Q]uit: |
Enter S to perform customized services configuration. This option is intended for experienced XC administrators who want to customize role services. Intervention like this is typically not required for HP XC systems. HP recommends that you enter p to continue with the system configuration process and proceed to step 10. Enter P to continue with the system configuration process. When the system configuration process begins, the database is not consistent and complete again until after cluster_config processing is complete. Therefore, do not press Ctrl+c during this time. Enter Q to exit from the system configuration process.
Proceed to “Task 7: Respond to Configuration Questions” to respond to the remainder of the system configuration questions. |
Printable version
