HP XC System Software Release Notes for Version 3.1

During normal operation of an ProLiant DL145 G2 node, the console port may lose connectivity
to the internal Administration Network for periods of tens of seconds to several minutes. This
loss of connection causes a disruption in node management and monitoring.
The connection loss is characterized by "no route to host" messages when using the ping
command to contact the node.
In most cases, the connection loss is related to a power action performed on the node, for instance,
when issuing a command to turn power on or off to a node, a shut down sequence performed
by the operating system, and so on.
The power command and some of the monitoring facilities do not operate properly when this
condition exists. For example,
• If a node's console port loses connectivity to the administration network, you will see the
message “NO_RESPONSE_ERROR” when running the power [--on | --off
|--status] command.
• You may also see the following error in the output of the shownode metrics sensor
command:
# shownode metrics sensors
2005-10-17 06:44:56 |n446 |Error |0 |Unable to retrieve sensor information
• The CMF daemon, cmfd, will also lose its connection to the node's console port and console
logging for that node will temporarily cease. However, cmfd tries to reconnect to the node's
console port every 60 seconds until successful.
5.5 Notes That Apply Before You Invoke the cluster_config Utility
Read the notes in this section before you invoke the cluster_config utility.
5.5.1 Remove the management_hub and console_network Roles From the Head
Node on Systems With More Than 256 Nodes
By default, the head node is assigned with the console_network and management_hub roles.
You have the flexibility to assign the console_network and management_hub roles to multiple
nodes and HP recommends that you consider using one management hub for every 64 to 128
nodes.
However, on systems with more than 256 nodes, HP recommends that you remove the
console_network and management_hub roles from the head node so that the management
hub activity does not interfere with the services that only run on the head node.
Use the cluster_config utility to assign roles or remove roles from a node or nodes.
5.5.2 Follow Role Assignment Guidelines for Improved Availability
When you are configuring services for improved availability, follow the role assignment guidelines
in Table 1-2 in the HP XC System Software Installation Guide. Role assignments for a traditional
HP XC system without improved availability of services is slightly different, thus it is important
to follow the guidelines in Table 1-2.
5.5.3 Update initrd Files With Required Hardware
When client nodes contain additional hardware that is required for booting (such as SCSI
controllers) and the hardware is not present on the head node, you must modify the initrd file
on the head node to include the drivers. Perform this task before the cluster_config utility
is run for the first time. Performing this task after running the cluster_config utility requires
you to remove the golden image.
For example, if a client node contains the Adaptec SCSI controller, the following script updates
all initrd files:
5.5 Notes That Apply Before You Invoke the cluster_config Utility 41