4 preparing for scalable deployment – HP Insight Control Software for Linux User Manual
Page 110
TIP:
For information about the importance of choosing unique names for items in the
repository, see
The name you supply is appended with a unique identifier and the date and time when the
task occurred. The image is stored as a gzipped tar file in the /opt/repository/image/
{prefix_date_and_time}
directory on the CMS.
7.
Select a Precapture script, a Postcapture Script, or both.
A Precapture script is run on the managed system before the image is captured. A Postcapture
script is run on the managed system after the image is captured. The default behavior is to
not run either type of script.
8.
Do one of the following:
•
Select Run Now to launch the image capture operation immediately.
•
Select Schedule to schedule the image capture operation to occur in the future.
9.
Monitor the Task Results window to follow the progress of the deployment operation and the
related task states.
10. Select the following menu item from the Insight Control user interface to view the task results:
Tasks & Logs
→Task Results
NOTE:
After an image is captured from a managed system, a sanity size check is performed on
the captured image to verify that it is valid.
By default, if the captured image size is less than 1,000,000 bytes, the image is considered invalid,
because most operating systems yield an image size greater than that. Typically an image size
less than 1,000,000 bytes indicates that an error occurred that prevented Insight Control for Linux
from capturing the image.
You can change the minimum image size threshold by editing the /opt/mx/icle/
icle.properties
file to change the value of the CAPTURE_IMAGE_MIN_SUCCESS_SIZE
variable; for example:
CAPTURE_IMAGE_MIN_SUCCESS_SIZE=1048575
Setting the CAPTURE_IMAGE_MIN_SUCCESS_SIZE variable to zero overrides the sanity check
and the image size is not checked.
10.4 Preparing for scalable deployment
Scalable deployment is an optional feature that allows you to deploy a Linux image to a large
number of servers (at least 32) in a short amount of time. However, you must inform Insight Control
for Linux which servers are connected to which switch or
beforehand. As illustrated in
, the CMS is connected to other servers through enclosures and switches (for BladeServers
and non-BladeServers respectively). The BladeServers within an enclosure and the servers connected
to a switch form network groups. Each network group has at least one server that acts as the group
leader.
110
Capturing and deploying Linux images