This chapter contains the following topics:
Your SGI ICE X system was tested and configured at the factory. At the factory, SGI configured the following:
A factory-specified root password. One of the first steps in the configuration procedure is to change this root password on the system admin controller (SAC) node.
Two slots. SGI configured the operating system that you ordered on slot 1. The operating system can be either Red Hat Enterprise Linux (RHEL) or SUSE Linux Enterprise Server (SLES). Slot 2 is blank.
The SGI ICE X system supports a maximum of five slots. If you need more than two slots, you need to reconfigure the system. During the reconfiguration, you reinstall the operating system and perform many other tasks. For the reconfiguration procedure, see Chapter 3, “Installing and Configuring an SGI ICE X System”.
A console configured with a serial port.
If you want to retain the factory-installed configuration, complete the following procedures:
Your configuration session can proceed more quickly if you gather some information before you start. You need to update the factory-installed, system-wide root password and the time zone. In addition, you need to provide information about your site network for the SAC's eth0 network interface card (NIC).
The following procedure explains the information that you need to gather.
Procedure 2-1. To obtain information for the customization
Complete the following table:
Information Needed Specifics for this SGI ICE X System
Factory-installed password ____________________________________
Password for this system at your site ____________________________________
Time zone ____________________________________
IP address ____________________________________
Netmask ____________________________________
Hostname ____________________________________
Default route/Gateway ____________________________________
Fully qualified domain name (FQDN) ____________________________________
House NTP server ____________________________________
First house (site) DNS resolver IP address ____________________________________
(Optional) Second house DNS resolver IP address ____________________________________
(Optional) Third house DNS resolver IP address ____________________________________
House (site) domain ____________________________________
SGI ICE X system subdomain name ____________________________________
Proceed to the following:
The following procedure explains how to change the password on the SAC and how to update the operating system configuration files with your site's networking information.
Procedure 2-2. To customize a factory-installed RHEL operating system
Use the console attached to the SAC, and log into the SAC as the root user.
Type the following command to change the root password on the SAC and on all other nodes:
# cpasswd --h
If you do not have the current password, you can obtain the factory-installed password from your SGI representative.
For example:
SAC:~ # cpasswd Enter new password: Enter new password (again): admin: updating /etc/shadow r1lead: updating /etc/shadow service0: updating /etc/shadow SAC:~ #
Change the system time zone.
Type the following command:
# system-config-date
The system-config-date command starts a graphical user interface (GUI) tool. Within the GUI tool, change only the system time zone. The tool enables you to change other aspects of the configuration, but for this step, change only the system time zone.
Note: Do not use this tool to change the NTP server, the time, or other configuration data. Use a text editor, such as vi or vim, to open file /etc/sysconfig/network-scripts/ifcfg-eth0 .
Edit the /etc/sysconfig/network-scripts/ifcfg-eth0 file.
Add lines for the IPADDR and NETMASK values appropriate for your public (house) network. Also set ONBOOT=yes.
For example:
IPADDR=128.162.244.88 NETMASK=255.255.255.0 ONBOOT=yes
Save and close file /etc/sysconfig/network-scripts/ifcfg-eth0 .
Use a text editor to create file /etc/sysconfig/network .
Add the following three lines to file /etc/sysconfig/network :
NETWORKING=yes HOSTNAME=SAC_hostname GATEWAY=gateway_IP_address
For SAC_hostname, type the hostname you want to assign to the SAC.
For gateway_IP_address, type the IP address of the gateway for your house network.
For example:
NETWORKING=yes HOSTNAME=my-system-admin GATEWAY=128.162.244.1
Save and close file /etc/sysconfig/network.
Use a text editor to open file /etc/hosts.
Add a line in the following format to file /etc/hosts :
SAC_IP SAC_FQDN SAC_hostname
The variables in the preceding line are as follows:
For SAC_IP, type the IP address of the SAC.
For SAC_FQDN, type the fully qualified domain name (FQDN) of the SAC.
For SAC_hostname, type the hostname of the SAC.
For example, add the following line:
128.162.244.88 acme-admin.acme.usa.com acme-admin
Save and close file /etc/hosts.
Type the following command to set the SAC hostname:
# hostname SAC_hostname
For SAC_hostname, type the hostname of the SAC.
For example:
# hostname acme-admin
Proceed to the following:
The following topic explains how to use the cluster configuration tool to add information about your site's network to the cluster database.
Procedure 2-3. To customize the cluster database
Type the following command to start the cluster configuration tool:
# /opt/sgi/sbin/configure-cluster
For general information about the cluster configuration tool, see “Running the Cluster Configuration Tool” in Chapter 3.
On the cluster configuration tool's main menu select Configure the Time Client/Server (NTP) and select OK.
The system guides you through the process to specify your house NTP server in file /etc/ntp.conf. This process differs, depending on your platform, as follows:
On RHEL platforms, follow the the instructions that the cluster configuration tool presents to you.
On SLES platforms, the cluster configuration tool opens a YaST2 menu. Follow the prompts in the YaST2 menu to set your NTP server.
On the cluster configuration tool's main menu select Configure House DNS Resolvers and select OK.
You can specify up to three house DNS resolvers.
Select Quit and select OK to log out from the cluster configuration tool.
Use the cadmin command, in the following format, to set the house (site) domain:
# cadmin --set-admin-domain site_domain
For site_domain, specify the full name of your house domain. For example, usa.acme.com.
Use the cadmin command, in the following format, to change the subdomain name for the SGI ICE X system:
# cadmin --set-subdomain cluster_name
For cluster_name, specify the name of the system. For example, sleet.usa.acme.com.
For more information about the cadmin command, type cadmin -h at the system prompt.
Type the following command to retrieve the path to the system images:
# cinstallman --show-images
An SGI ICE X system includes the following major nodes:
One system admin controller. This node is also called the SAC or the admin node.
One or more rack leader controllers. These nodes are also called the RLCs. There is one RLC per rack.
Several compute nodes. These nodes are also called blades. The compute nodes are housed in racks, and each rack has one RLC.
One or more service nodes. Users log into the system through the service nodes to run jobs.
The SAC hosts the master images for each of the preceding node types. If you need to change some aspect of a node's configuration, SGI recommends that you change the configuration in the master image and push out the changed master image to the affected nodes. This practice maintains consistency between the master images on the SAC and the production images on the nodes. For more information about the system images and how they install, see “About Performing a New Installation and Configuring the Software on an SGI ICE X System” in Chapter 3.
The following example shows how to retrieve the paths to the compute node images on the SAC:
# cinstallman --show-images Image Name BT Path compute-rhel6.5 1 /var/lib/systemimager/images/compute-rhel6.5 service-rhel6.5 0 /var/lib/systemimager/images/service-rhel6.5 lead-rhel6.5 0 /var/lib/systemimager/images/lead-rhel6.5
The following steps explain how to update the master node images with your site's time zone information.
Use the cp(1) command, in the following format, to set the time zone in the system images.
The format of this command is as follows:
cp /etc/localtime /var/lib/systemimager/images/image_name/etc
For image_name, type the name of one of the system images you retrieved with the preceding cinstallman command. Type one of these commands for each master node system image.
For example, type the command three times, one for each master image, as in the following example:
# cp /etc/localtime /var/lib/systemimager/images/compute-rhel6.5/etc # cp /etc/localtime /var/lib/systemimager/images/service-rhel6.5/etc # cp /etc/localtime /var/lib/systemimager/images/lead-rhel6.5/etc
Type the following commands to update the RLC image and service node images with your site's time zone information:
# pdcp -g leader /etc/localtime /etc/localtime # pdcp -g service /etc/localtime /etc/localtime
Proceed to the following:
SGI ICE X systems can host multiple forms of system images for the rack leader controller (RLC) nodes, the service nodes, and the compute nodes. For example, you can have some production images and some test images, and you can push the images to the nodes as needed. The preceding procedure explained how to update the master images for the RLC nodes, the service nodes, and the compute nodes. In addition, the preceding procedure explained how to push the updated the RLC and service nodes images to the ICE X system components.
In this topic's procedure, you push the updated compute node images to the compute nodes. The push action enables the compute nodes to run with the updated networking information that you configured in “Completing the Customization”.
Procedures that describe system operations, later on in this manual, refer back to this procedure because you need to update system images and push out new images as part of several system administration tasks.
The following procedure explains how to push compute node system images to compute nodes.
Procedure 2-4. To push software system images
Type the following command to stop the compute nodes:
# cpower --halt r*i*n*
The preceding command stops all the compute nodes. Use the preceding command when you need to push an updated image out to all the nodes.
(Optional) Provide information about the number of racks on your system.
Perform this step if you have a small system with fewer than eight IRUs per RLC.
The procedure pushes the updated compute image to all the compute nodes. This process can run for a long time on large systems. If you have a large number of IRUs, you need the system to perform expansions that enable you to change many compute nodes at a time. If you have fewer than eight IRUs per RLC, however, the expansions are not needed.
The following substeps explain how to prepare the system to work on a smaller number of compute nodes:
Type the following command to retrieve the identifiers for the RLCs on your system:
# cnodes --leader
Type the following command one or more times to suppress unnecessary processing:
cadmin --set-max-irus --node rlc_id number_of_racks
For rlc_id, specify the identifier for one of the RLCs in your system.
For number_of_racks, specify the number of IRUs associated with this RLC.
For example, the following command specifies that there is only one IRU associated with the RLC identified as r1lead:
# cadmin --set-max-irus --node r1lead 1
Use the cimage command, in the following format, to push the changes:
cimage --push-rack compute-image_name rack
For image_name, specify the name of the compute node image that you updated.
For rack, specify the nodes. To specify all compute nodes, specify r\* or r*i*n*. To specify only selected nodes, specify rxi xnx, and substitute specific integer numbers for the x characters.
For example, the following command pushes the time zone changes (from “Completing the Customization”) to all the compute nodes:
# cimage --push-rack compute-rhel6.5 r\*
Type the following command to power-up the compute nodes:
# cpower --boot r*i*n*
Proceed to the following:
The following procedure explains where you can obtain information about how to configure additional features.
Procedure 2-5. To configure additional features
Configure additional features.
SGI recommends that you configure certain additional features, such as the hardware event tracker (HET). Other features, such as CPU frequency scaling might benefit your installation. For information about these additional features, see the following:
(Optional) Configure optional features.
The SGI ICE X system supports several optional features, for example, networking features such as network address translation (NAT). For information about how to configure optional features, see the SGI ICE X Administration Guide.