System and method for failover

• US 20060190760A1
• Filed: 04/18/2006
• Published: 08/24/2006
• Est. Priority Date: 03/12/2004
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Saika, Nobuyuki
• Granted Patent: US 7,313,722 B2
• Time in Patent Office: Days
• US Class Current: 714/4.1
• CPC Class Codes: G06F 11/2025 using centralised failover control functionality
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 11/2046 where the redundant components share persistent storage G06F11/2043 takes precedence
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029

Abstract

A node 1 and a node 2 are in a mutual failover relationship and share information used in failover through a shared LU. Of filesystems FS1A, FS1B that are mounted at the node 1, the actions of level 1 are allocated to FS1A and the actions of level 2 are allocated to FS1B. The level 1 filesystem FS1A is taken over to the node 2 simultaneously with commencement of failover. The level 2 filesystem FS1B is taken over to the node 2 when an access request for FS1B is generated after commencement of failover. In this way, business services with high availability can be restarted at an early stage.

Claims

1. A failover cluster system in which a plurality of computers are connected and, in a prescribed case, failover object resources of a failover source computer are taken over by a failover target computer, comprising a control section that is capable of taking over said failover object resources in stepwise fashion.

2. The failover cluster system according to claim 1 wherein said control section is capable of taking over said failover object resources in stepwise fashion in accordance with a priority ranking set for said failover object resource.

3. The failover cluster system according to claim 2 wherein said control section sets up said priority ranking beforehand for the failover object resources, based on the state of use of said failover object resources.

4. The failover cluster system according to claim 3 wherein each of said computers employs a shared memory device to share takeover information relating to takeover of said failover object resources and said control section can thus take over said failover object resources in stepwise fashion in accordance with said priority ranking, by referring to the takeover information of said shared memory device.

5. The failover cluster system according to claim 4 wherein said takeover information is constituted by associating information for specifying said failover object resources with takeover processing actions set for said failover object resources in accordance with said priority ranking.

6. The failover cluster system according to claim 5 wherein said priority ranking includes a first ranking whereby takeover processing is immediately executed and a second ranking whereby takeover processing is executed when an access request for said failover object resources is generated.

7. The failover cluster system according to claim 6 wherein said priority ranking further includes a third ranking whereby takeover processing of said failover object resources is executed if said failover target computer is in a prescribed low-load condition.

8. The failover cluster system according to claim 6 wherein said priority ranking further includes a fourth ranking whereby takeover processing is not executed.

9. The failover cluster system according to claim 6 wherein said failover object resources are filesystems and said priority rankings are respectively set beforehand for each of these filesystems.

10. A method of failover of a failover cluster system constituted by connecting a plurality of computers between which a mutual failover relationship has been established, comprising the steps of:

monitoring the state of use of a failover object resource;

setting a priority ranking of said failover object resource in accordance with said state of use;

storing on a shared disk shared by each said computer takeover information constituted by associating information for specifying said failover object resource with a takeover processing action set for said failover object resource in accordance with said priority ranking;

determining whether or not a failover execution condition has been established; and

taking over said failover object resource of a failover source computer in stepwise fashion onto a failover target computer in accordance with said priority ranking, by referring to said takeover information stored on said shared disk, if it is determined that said failover execution condition has been established.

11. A failover cluster system comprising a failover source computer, a failover target computer connected with this failover source computer and a shared disk shared by said failover source computer and said failover target computer, wherein, in said failover source computer, there is provided a priority ranking determination processing section that classifies filesystems constituting the failover objects into one of a first category, second category or third category in accordance with the state of use of these respective filesystems and that stores in said shared disk the correspondence relationship of these respective filesystems and said respective categories and, in said failover target computer, there are provided a failover processing section that executes immediate mounting for the filesystems belonging to said first category and an access request acceptance processing section that, if an access request is generated in respect of a filesystem belonging to said second category, executes mounting for the filesystem belonging to said second category but does not execute mounting in respect of a filesystem belonging to said third category irrespective of whether or not there is a request for access.