Recovery and relocation of a distributed name service in a cluster filesystem

• US 7,593,968 B2
• Filed: 01/16/2003
• Issued: 09/22/2009
• Est. Priority Date: 06/05/2001
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Silicon Graphics Incorporated (Graphics Properties Holdings, Inc.)
• Inventors: Beck, Ken
• Primary Examiner(s): Ehichioya; Fred I
• Application Number:
  US10/345,357
• Publication Number:
  US 20040143607A1
• Time in Patent Office: 2,441 Days
• Field of Search: 707/1, 707/8, 707/104.1, 707/200, 707/201, 707/205, 707/10, 707/103.R, 709/219, 709/226, 709/203, 709/224
• US Class Current: 1/1
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users

Abstract

A cluster of computer system nodes share direct read/write access to storage devices via a storage area network using a cluster filesystem. During relocation of a server for a distributed name service and recovery of a cluster, entries related to the distributed name service for filesystems is updated. During relocation, a new server for a filesystem informs all nodes in the cluster of the new server'"'"'s location. During recovery, a process executing on each node deletes entries related to the distributed name service for any filesystem that does not have a server in the recovering cluster.

Claims

1. A method of administering a distributed name service for a cluster filesystem accessed by computer system nodes sharing direct read/write access to storage devices via a storage area network, comprising:

automatically updating entries related to a distributed name service stored on the computer system nodes upon one of relocation of a server for the cluster filesystem and recovery of a cluster if no node in the cluster provides the server for the cluster filesystem.

2. A method as recited in claim 1, wherein said updating includes deleting the entries related to the distributed name service stored on the computer system nodes during recovery of the cluster if no node in the cluster provides the server for the cluster filesystem.

3. A method as recited in claim 2, wherein said deleting is performed by a process executing independently on each of the computer system nodes.

4. A method as recited in claim 1, wherein said updating includes sending, during relocation of the cluster filesystem, location information of a new server for the cluster filesystem from the new server to the computer system nodes in the cluster.

5. A method of automatically recovering a cluster filesystem accessed by computer system nodes sharing direct read/write access to storage devices via a storage area network, comprising:

executing a process on each of the computer system nodes in a recovering cluster to delete entries related to a distributed name service for any filesystem for which no server is provided by any of the computer system nodes in the recovering cluster.

6. A method of relocating a server of a filesystem in a cluster filesystem accessed by computer system nodes sharing direct read/write access to storage devices via a storage area network, comprising:

sending, during relocation of the cluster filesystem, location information of a new server for the filesystem in a message to the computer system nodes from one of a previous server for the filesystem and the new server for the filesystem.

7. A method as recited in claim 6,

wherein the new server is provided by one of the computer system nodes, and

wherein said sending sends the location information to at least all of the computer system nodes which do not provide the new server.