US 8,788,667 B2Granted

Computer system and management method

AssigneeHitachi, Ltd.

InventorsNobuyuki Saika, Sadahiro Nakamura

Life of this patent

Estimating the term…

Loading patent…

Abstract

Proposed are a computer system and a management method capable of localizing influence from a failure of a physical node and preventing influence from the physical node failure from spreading. This computer system includes one or more servers that respectively provide a different single namespace to a client, a plurality of physical nodes that are capable of executing a plurality of virtual nodes that respectively belong to one of the single namespaces and provide a file sharing service to the client, and a management apparatus for managing the virtual nodes, and executes control of calculating a degree of contribution to a user for each of the virtual nodes, and selecting the virtual node to be migrated and the physical node of a migration destination so as to average, among the physical nodes, a total value of the degree of contribution of each of the virtual nodes running on the physical nodes based on the degree of contribution to the user for each of the virtual nodes, and migrating the virtual node selected as the migration target to the physical node selected as the migration destination.

Claims10

Dependent claims

A computer system, comprising:

one or more servers configured to respectively provide a plurality of single namespaces, different from one another to a client;
a plurality of physical nodes configured to execute a plurality of virtual nodes that respectively belong to one of the plurality of single namespaces and provide a file sharing service to the client;
and
a management apparatus for managing the virtual nodes;
wherein the management apparatus comprises:
a degree of contribution calculation processing unit configured to calculate a degree of contribution to a user for each of the virtual nodes;
and
a migration control unit configured to:
select the virtual node to be migrated and the physical node of a migration destination so as to average, among two or more of the plurality of physical nodes, a total value of the degree of contribution of each of the virtual nodes running on the physical nodes based on the degree of contribution to the user for each of the virtual nodes;
migrate the virtual node selected as a migration target to the physical node selected as the migration destination;
select the virtual node to be migrated among the virtual nodes running on the physical nodes in which the total value of the degree of contribution of each of the virtual nodes in operation is greater than an average value of the total value in all of the physical nodes and in which dispersion of the degree of contribution for each of the virtual nodes in operation is smallest;
calculate a multiplicity of the single namespace in each of the physical nodes;
select more than one physical node that comprises the multiplicity higher than a threshold value;
determine, for the selected physical nodes, a plurality of average values associated with a degree of contribution of the virtual nodes that belong to the physical node;
and
select, between the selected physical nodes, a physical node that comprises:
a highest average value of degrees of contribution of the virtual nodes that belong to the physical node;
and
a smallest dispersion of degrees of contribution of respective virtual nodes that belong to the physical node;
wherein,
when the virtual node belonging to the same single namespace as the migration target virtual node is running on the migration destination physical node, the migration control unit is configured to merge the migration target virtual node and the virtual node running on the migration destination physical node;
and
when the virtual node belonging to a different single namespace as the migration target virtual node is running on the migration destination physical node, the migration control unit is configured not to merge the migration target virtual node and the virtual node running on the migration destination physical node.

20 more elements

A management method of managing virtual nodes that are running on a plurality of physical nodes capable of executing a plurality of virtual nodes that provide a file sharing service to a client, wherein the virtual nodes respectively belong to a plurality of single namespaces which are different from each other;

wherein the management method comprises:
calculating a degree of contribution to a user for each of the virtual nodes;
and
selecting the virtual node to be migrated and the physical node of a migration destination so as to average, among two or more of the physical nodes, a total value of the degree of contribution of each of the virtual nodes running on the physical nodes based on the degree of contribution to the user for each of the virtual nodes;
migrating the virtual node selected as the migration target to the physical node selected as the migration destination;
selecting the virtual node to be migrated among the virtual nodes running on the physical nodes in which the total value of the degree of contribution of each of the virtual nodes in operation is greater than an average value of the total value in all of the physical nodes and in which a dispersion of the degree of contribution for each of the virtual nodes in operation is smallest;
calculating a multiplicity of each physical node;
selecting more than one physical node that comprises the multiplicity higher than a threshold value;
determining for the selected physical nodes, a plurality of average values associated with a degree of contribution of the virtual nodes that belong to the physical node;
and
selecting, between the selected physical nodes, a physical node that comprises:
a highest average value of the degrees of contribution of the virtual nodes that belong to the physical node;
and
a smallest dispersion of degrees of contribution of respective virtual nodes that belong to the physical node;
wherein,
when the virtual node belonging to the same single namespace as the migration target virtual node is running on the migration destination physical node, merging the migration target virtual node and the virtual node running on the migration destination physical node;
and
when the virtual node belonging to a different single namespace as the migration target virtual node is running on the migration destination physical node, not merging the migration target virtual node and the virtual node running on the migration destination physical node.

15 more elements

Dependencies are inferred from the claim text.

Facts

Application: 12/510,051Filed 27 Jul 2009
Priority: JP 2009-11921915 May 2009
Published: US 2010/0293279 A118 Nov 2010
Granted: 22 Jul 2014
Term ends: Not estimatedThe source records don’t support an estimate.

Classification6 codes

G06F 11/0727
The processing taking place on a specific hardware platform or in a specific software environment: in a storage system, e.g. in a DASD or network…
G06F 11/2094
Where persistent mass storage functionality or persistent mass storage control functionality is redundant: redundant storage or storage space
H04L 67/1001
In which an application is distributed across nodes in the network: for accessing one among a plurality of replicated servers
G06F 11/0793
Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of…

Family

US 2010/0293279 A1Pre-grant publication · published 18 Nov 2010
US 8,788,667 B2This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key