IBM NeXtScale System, Planning And Implementation Manual

Page: 173 / 266

Chapter 6. Factory integration and testing
155
6.2 Testing
The following tasks are typical of the testing IBM does at the factory. Other
testing might be done based on unique hardware configurations or client
requirements:
򐂰
All servers are fully tested as individual units before rack integration.
򐂰 After they are installed in the rack, there is a post assembly inspection to
assure all components are installed and positioned correctly.
򐂰
IBM configures all switches; terminal servers; keyboard, video, and mouse
(KVM) over IP units, and other devices with IP addresses and host names to
allow for communication and control.
These can be set by using client-supplied information or to a default scheme.
For more information about the default scheme, see this website:
http://download.boulder.ibm.com/ibmdl/pub/systems/support/system_x_p
df/intelligent_cluster_factory_settings_102411.pdf
򐂰 Power redundancy testing:
If there is a client-provided redundant power domain scheme, IBM tests with
one domain powered down, then the opposite domain powered down, to
ensure that all devices with redundant power feeds stay powered on.
Otherwise, remove and restore power from each PDU to assure all devices
with redundant power feeds stay powered on.
򐂰
Flash all servers, adapters, and devices to current IBM development-provided
best recipe.
򐂰 From a cluster management node, discover servers and program their
integrated management module (IMM). This allows for remote control of the
computational infrastructure.
򐂰
Serial (console) over LAN (SoL) setup: Unless the cluster has terminal
servers, SoL is configured and tested.
򐂰 Set up RAID arrays on local disks as defined by the client or client architect,
or per Intelligent Cluster best practices.
򐂰
Set up shared storage devices, configure arrays with all disks present, and
create and initialize logical disks to verify functionality.