IBM Power Systems 775 Manual page 88

For aix and linux hpc solution
Table of Contents

Advertisement

The number of DHCP servers does not need to equal the number of TFTP or HTTP servers.
TFTP servers NFS mount read-only the /tftpboot and image directories from the management
node to provide a consistent set of kernel, initrd, and file system images.
xCAT version 2 provides the following enhancements that address the requirements of a
Power 775 cluster:
Improved ACLs and non-root operator support:
– Certificate-authenticated client/server XML protocol for all xCAT commands
Choice of databases:
– Use a database (DB) like SQLite, or an enterprise DB like DB2 or Oracle
– Stores all of the cluster config data, status information, and events
– Information is stored in DB by other applications and customer scripts
– Data change notification is used to drive automatic administrative operations
Improved monitoring:
– Hardware event and simple Network Management Protocol (SNMP) alert monitoring
– More HPC stack (GPFS, LL, Torque, and so on) setup and monitoring
Improved RMC conditions:
– Condition triggers when it is true for a specified duration
– Batch multiple events into a single invocation of the response
– Micro-sensors: ability to extend RMC monitoring efficiently
– Performance monitoring and aggregation that is based on TEAL and RMC
Automating the deployment process:
– Automate creation of LPARs in every CEC
– Automate set up of infrastructure nodes (service nodes and I/O nodes)
– Automate configuration of network adaptors, assign node names/IDs, IP addresses,
and so on
– Automate choosing and pushing the corresponding operating system and other HPC
software images to nodes
– Automate configuration of the operating system and HPC software so that the system
is ready to use
– Automate verification of the nodes to ensure their availability
Boot nodes with a single shared image among all nodes of a similar configuration
(diskless support)
Allow for deploying the cluster in phases (for example, a set of new nodes at-a-time by
using the existing cluster)
Scan the connected networks to discover the various hardware components and firmware
information of interest:
– Uses the standard SLP protocol
– Finds: FSPs, BPAs, hardware control points
Automatically defines the discovered components to the administration software,
assigning IP addresses, and hostnames
Hardware control (for example, powering components on and off) is automatically
configured
ISR and HFI components are initialized and configured
All components are scanned to ensure that firmware levels are consistent and at the
wanted version
74
IBM Power Systems 775 for AIX and Linux HPC Solution

Advertisement

Table of Contents
loading

Table of Contents