After you update the golden image successfully, and all nodes are optionally set to network boot, you are ready to deploy the updated image to all client nodes. This section describes the methods for distributing files to all the client nodes.
Using the Full Imaging Installation |
 |
A recommended procedure to propagate the golden image is to install all client nodes automatically. This ensures that they receive the updated image and any updated configuration information automatically. When all nodes are set to network boot, a reboot of each client node starts the automatic installation. After each node completes its installation, it automatically reboots and is available for service.
Use the stopsys and startsys commands to reboot all the client nodes:
# stopsys
# startsys --image_and_boot |
Using the si_updateclient Utility |
|
Another recommended procedure is to use the si_updateclient utility to propagate the golden image to all client nodes. The si_updateclient utility eliminates the need to reinstall all nodes.
The si_updateclient command's --quiet option suppresses the display of the transferred file names when you run this command.
Alternatively, the --dry-run option enables you to verify which files would be updated and deleted without performing the operation.
The following is a general procedure for distributing the golden image to all client nodes using the si_updateclient utility:
Log in as superuser on the head node.
Ensure that the HP XC system is relatively inactive (quiesced).
Use the appropriate drain utility to drain jobs from the node. For example, “Draining Nodes” describes how to drain jobs from a node using SLURM.
Run the updateimage utility on the head node to update the golden image. Be sure to use the --no-netboot option so that the client nodes are not set to network boot the next time they are rebooted:
# updateimage --gc `nodename` --no-netboot |
For more information on the updateimage utility, see “The updateimage Command” and updateimage(8).
Run the si_updateclient utility in conjunction with the cexec command to distribute the golden image to all client nodes:
# cexec -x `nodename` -a "si_updateclient \
--quiet --server nh --image base_image --no-bootloader\
--override base_image" |
If you have additional site-specific overrides that you need to deliver, use the following cexec command instead:
# cexec -x `nodename` -a "si_updateclient \
--quiet --server nh --image base_image --no-bootloader\
--override base_image --override site-override-dir" |
The site-override-dir parameter in this command refers to a site-specific directory under /var/lib/systemimager/overrides.
For more information on file overrides of the golden image, see “Using File Overrides to the Golden Image”. For information on the --override option, see si_updateclient(8).
Run the per-node configuration service on the head node to update it:
# service nconfig nconfigure
# service nconfig restart |
Run the following commands to run the per-node configuration service on all the client nodes:
# cexec -x `nodename` -a "service nconfig nconfigure"
# cexec -x `nodename` -a "service nconfig restart" |
Using the cexec Command |
|
You can use the cexec command to copy designated files to client nodes.
Its advantage is that it is performed immediately, and potentially with little effect on the HP XC system configuration files, presuming the file list does not contain any configuration files that require the restarting of services.
Its disadvantages are that you must copy the files or directories and that you must update the golden image to sustain the change when the node is reimaged.
Printable version
