Method for switching node and an information processing system

Abstract

In an information processing system including host computers and disk devices, each of an execution-node host and a standby-node host includes an I / O request unit and an access-right change command unit. The I / O request unit issues an I / O request. The access-right change command unit transmits an access-right change command. The access-right change command results from causing I / O-enable / disable information and the host identification information to correspond to each other. The disk device includes an access control table, an access control unit, and an access-right change unit. The access control table stores information of the access-right change commands from the hosts. The access control unit judges the execution enablement / disablement for the I / O requests from the host identification information and the access control table. The access-right change unit, in accordance with the access-right change commands from the hosts, changes in batch the I / O-enable / disable information on each host basis within the access control table.

Description

BACKGROUND OF THE INVENTION [0001] The present invention relates to a node switching method and an information processing system. More particularly, it relates to an I / O node-switching method and an information processing system in the following case: At the time of a failure occurrence, a processing, which is being executed by a host where the failure has occurred, is taken over to another host so as to allow another host to continue this processing. [0002] Generally speaking, starting with a financial system and a securities system, systems which form and support the social infrastructure are requested to exhibit a high reliability. Namely, these systems are not permitted to fall into the service interruption, i.e., the system down. On account of this requirement, these systems are configured as follows: Namely, devices which configure these systems, such as hosts and paths for connecting the hosts with a disk device, are formed into a dual-redundant structure. For example, even i...

AI Technical Summary

Benefits of technology

The present invention provides a method for performing I / O cut-off operations at a disk device in a storage area network (SAN) environment, where multiple hosts and logical disks are connected via switches. The method allows for a faster and more efficient switching of the disk device to another host in case of failure. The method includes steps of transmitting access-right change commands to the disk device in advance, changing the I / O-enable / disable information based on the host identification information, and simultaneously storing and holding the access-right change commands. The method can be applied to a large-scale system with many hosts, logical disks, and paths connected to the disk device. The invention also provides an information processing system that controls the execution enablement / disablement for I / O requests from multiple host computers.

Problems solved by technology

This results in a problem that the above-described method is inapplicable in the SAN environment.

As a consequence, it is impossible to apply the first method to situations where accesses from plural paths are wished to be permitted, such as a multi-path environment where there are provided plural paths from one host to a logical disk, and a case where plural operation-node hosts exist.

This results in a problem that a limitation is imposed on the configuration to which the method is applicable.

This gives rise to a problem that this total amount of time needed leads to an increase in the service interruption time-period at the time of the failure occurrence.

Images