Browse Prior Art Database

Distributed NVRAM in Multinode Systems, for Setup/Configuration Information

IP.com Disclosure Number: IPCOM000132143D
Original Publication Date: 2005-Dec-02
Included in the Prior Art Database: 2005-Dec-02
Document File: 1 page(s) / 24K

Publishing Venue

IBM

Abstract

In multinode systems, each node starts up as an independent node. They are merged together during the boot process. However, each node needs access to setup/configuration information, before as well as after, the nodes are merged to form a multinode system. The setup/configuration is stored in nonvolatile memory. Since, all nodes need an access to the setup/configuration information, during all stages of the boot process, saving it in the Non-Volatile Random Access Memory (nvram) of the boot node alone is not sufficient. The idea is to distribute the setup/configuration information amongst all nodes of the multinode system. Until the nodes are merged, each node uses the setup/configuration information stored in it's own nvram. After the nodes are merged, each node only uses the setup/configuration information stored in the boot node's nvram. Since, each node can boot as a single node system or as part of the multinode system, the setup/configuration utilities synchronizes nvram of all nodes in the system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Distributed NVRAM in Multinode Systems , for Setup/Configuration Information

Each node of the multinode system contains nvram, because each node is capable of booting by itself, as a single node system. The system firmware stores the configuration/setup information at same locations in nvram of each node, as well as, in the same format. So, the same code can be easily used to access the each node's nvram or the boot node's nvram. Before the nodes are merged, the system firmware always read/writes the local node's nvram for the setup/configuration information, because it only has access to it's own nvram. After the nodes are merged, the system firmware only reads/writes the boot node's nvram for the setup/configuration information, even though each node has access to nvram of all other nodes in the system. The setup/configuration utilities, displays premerge information from each node, but postmerge information only from the boot node.

     Additionally, the setup/configuration utilities provide a way to synchronize the nvram of all nodes in the system, so that any node in the system can become the boot node.

Boot Path

Setup

Utility

Read/Write Local Node's NVRAM

Read Boot Node's NVRAM

Nodes

Merged Into
Partition

Write All Node's NVRAM

Read/Write Boot Node's NVRAM

1