The first step in managing software changes to your HP XC system is to update the golden client node. This can involve adding new software packages, adding a new user, or modifying a configuration file that is replicated across the HP XC system, such as a NIS or NTP configuration file.
If you need to add a software package or service configuration file to the golden client that should not be distributed to all nodes, be sure to prevent these files from becoming part of the golden image by using an exclusion file. See “Exclusion Files” for further details.
If the package or service configuration file is specific to the head node and it is not to be distributed to any other node, then making the change on the golden client and excluding the necessary files from being included in the golden image is sufficient. However, if the software package or configuration file is destined for other (specialized) nodes in the HP XC system, then a procedure is required to distribute these changes to those nodes. You can use either of the following methods to accomplish this:
Installing Additional RPMs from the HP XC System Software Installation DVD |
 |
Use the following procedure to install additional RPMs from the HP XC System Software Installation DVD:
Log in as the superuser on the head node.
Insert the installation DVD into the DVD drive.
Mount the DVD on the default location:
Change to the /mnt/cdrom/LNXHPC/RPMS directory.
Locate the RPM you want to install and issue the rpm command to install it; for example:
# rpm -ivh kde-i18n-French-3.1.3-1.noarch.rpm |
“Updating the Golden Image” discusses how to update the golden image with the new software for subsequent distribution to client nodes.
Using File Overrides to the Golden Image |
|
File overrides overwrite files delivered in the golden image. File overrides are copied to the client nodes after the golden image is transferred, thus overriding files in the golden image itself.
File overrides are organized in a subdirectory hierarchy rooted at /var/lib/systemimager/overrides/name. Each subdirectory of the ./overrides/name directory specifies a complete file override hierarchy that is copied to the root directory of the client's file system.
The necessity to use file overrides is specified in the master autoinstallation scripts, located at /var/lib/systemimager/scripts. Each master autoinstallation script contains an OVERRIDES variable specifying one or more overrides subdirectories to associate with this master autoinstallation script. To specify a select number of nodes to receive a particular override, make a copy of the appropriate master autoinstallation script, and modify the OVERRIDES variable to match your overrides subdirectory name, /name. Finally, in the /var/lib/systemimager/scripts directory, create symbolic links to this master autoinstallation script for the nodes that will receive this override. The symbolic link names must follow the format name.sh, where name is the host name of each node to receive the override. For further information on using overrides in the SystemImager environment, see the FAQ chapter in the SystemImager Manual, located at the following Web site:
http://systemimager.org/doc
The following procedure describes how to set up an override to allow a compiler package to be delivered only to the nodes with the login role:
Create an “overrides” subdirectory, named compiler, to contain the compiler package:
# cd /var/lib/systemimager/overrides
# mkdir ./compiler |
There should be two subdirectories, base_image and compiler, under the ./overrides directory.
# # ls -F
base_image/ compiler/ README |
Install the compiler package into the alternate root location, /var/lib/systemimager/overrides/compiler.
# rpm -ivh --root \
/var/lib/systemimager/overrides/compiler compilername.rpm |
Use the find command to verify the file hierarchy:
# find ./compiler
./compiler
./compiler/usr
./compiler/usr/lib/
.
.
. |
Make a copy of the default HP XC autoinstallation script as a new script to image the login nodes specifically with the compiler overrides. Name the new autoinstallation script compiler.master.0:
# cp base_image.master.0 compiler.master.0 |
Change directory to the SystemImager “scripts” directory:
# cd /var/lib/systemimager/scripts |
Use the text editor of your choice to modify the OVERRIDES variable definition in the compiler.master.0 file as follows:
[ -z $OVERRIDES ] && OVERRIDES="base_image compiler" |
It is necessary to append compiler after the string base_image because the default HP XC installation relies on the base_image override
Save the file and exit the text editor.
Link all the login nodes (that is, those nodes with the lvs service) to the compiler.master.0 autoinstallation script
# for i in $(expandnodes $(shownode servers lvs))
do
ln -sf compiler.master.0 $i.sh
done |
Verify that the links are correct
# ls -l
.
.
.
lrwxrwxrwx 1 root root ... n7.sh -> compiler.master.0
lrwxrwxrwx 1 root root ... n8.sh -> compiler.master.0
lrwxrwxrwx 1 root root ... n9.sh -> compiler.master.0 |
Now the system can be imaged. In this example, when nodes n[7-9] are installed with the full imaging installation, the compiler overrides are automatically delivered after the golden image has been delivered.
Use the --override option of the si_updateclient command to ensure that both overrides are properly delivered:
# si_updateclient --server nh --image base_image \
--no-bootloader --override base_image \
--override compiler |
Using Per-Node Service Configuration |
|
The HP XC system configuration process uses per-node configuration scripts to achieve personalized role configurations as necessary on each node. The per-node configuration process occurs initially during HP XC system configuration, at the time each client node is auto-installed. The HP XC configuration and management database (cmdb) contains the per-node role and service configuration, and is queried on each node's initial boot to identify which roles and services to configure. See “Adding a Service” for more information.
A per-node configuration script is associated with each configurable service, and is executed on the client node if the cmdb identifies this client as hosting this service.
The per-node configuration process actually occurs in two phases:
Global Service Configuration
The global configuration phase is intended to set up the per-node configuration phase by globally configuring the service for use within the HP XC system. The global configuration of a service occurs when a global configuration script (a gconfig script) is executed during the running of the cluster_config utility. It is here that you interact with the cluster_config utility as necessary to configure the service. In order to configure a new service into the HP XC system using gconfig, you must run the cluster_config utility again.
The global configuration script can store information in the following locations:
- cmdb
Use the database as the target for node-specific configuration data, that is, services that are not ubiquitous in the HP XC system, but run on only one or a handful of nodes.
- Golden client file system
By writing to the golden client file system, the script can add or modify files that are propagated to all nodes of the HP XC system by using the golden image. This is useful for configuring services that are ubiquitous in the HP XC system.
- Cluster common storage (/hptc_cluster)
By writing to the cluster common storage, the script is able to add or modify files that are visible to all the nodes of the cluster. Use this method for either per-node or clusterwide services.
As mentioned previously, the database should be the destination for node-specific data. However, some existing services may already place configuration data used by the service in files in known locations. The golden client file system and the cluster common storage are available to support such applications.
Global service configuration scripts are located in the /opt/hptc/etc/gconfig.d directory.
Node-Specific Configuration
The node-specific service configuration step uses the results of the global service configuration step described previously to apply to a specific node its “personality” with respect to the service. User interaction is not permitted because this step runs on a per-node basis.
The configuration of the service is accomplished in a script called by the node-specific service configuration controller (nconfig) script. The nconfig controller script runs at system startup, and it reports if is executing as part of the initial boot of a client node after an installation operation. If so, the nconfig controller script queries the cmdb to determine which node-specific service configuration scripts to execute, thus providing the per-node personality to each node.
Per-node configuration scripts are located in /opt/hptc/etc/nconfig.d directory.
If you used the si_updateclient utility, after successfully synchronizing each client node with the golden image, you must run the per-node configuration scripts on each node with the following commands. This returns the per-node personality following the golden image update:
# cexec -a -x `nodename` -f128 "/sbin/service nconfig nconfigure"
# cexec -a -x `nodename` -f128 "/sbin/service nconfig nrestart" |
The number of simultaneous remote commands are increased by using the cexec command's -f option.
You can monitor the execution of the per-node configuration scripts from a central location by monitoring the /hptc_cluster/adm/logs/imaging.log file. Run the following command from the imaging server node (currently, the head node):
# tail -f /hptc_cluster/adm/logs/imaging.log |
The per-node configuration scripts log their execution locally to the /var/log/nconfig.log file.
Printable version
