Granted Patent B2

US 7,461,302 · App. 10/918,201 · Granted Dec 2, 2008

Abstract

Claims (12)

1. In a distributed object-based storage system that includes a plurality of object storage devices for storing object components, a manager coupled to each of the object storage devices, wherein the object storage devices coordinate with the manager, and one or more clients that access and store distributed, object-based files on the object storage devices, a method for recovering from an I/O error, comprising:

attempting, by a client, to perform a data storage operation wherein the client attempts to store data that is the subject of the operation on one or more of the object storage devices, the operation selected from the group consisting of: a data write operation to an object storage device, a set attribute operation to an object storage device, and an object create operation to an object storage device; and

upon failure of the operation, sending a single message from the client to the manager that includes (a) information representing a description of the failure and (b) the data that the client attempted to store that was the subject of the operation.

2. The method of claim 1 , wherein the data that is the subject of the operation and that the client attempted to store is user-data or parity data.

3. The method of claim 1 , further comprising using the data in the message, a parity equation associated with the data in the message and other data on one or more of the object storage devices, to correct data associated with the failure.

4. The method of claim 1 , wherein the distributed object-based system is a RAID system.

5. In a distributed object-based storage system that includes a plurality of object storage devices for storing object components, a manager coupled to each of the object storage devices, wherein the object storage devices coordinate with the manager, and one or more clients that access and store distributed, object-based files on the object storage devices, a system for recovering from an 110 error, comprising:

a client that attempts to perform a data storage operation where the client attempts to store data that is the subject of the operation on one or more of the object storage devices, the operation selected from the group consisting of: a data write operation to an object storage device, a set attribute operation to an object storage device, and a create object operation from to an object storage device; and

wherein, upon failure of the operation, the client sends a single message from the client to the manager that includes (a) information representing a description of the failure and (b) the data that the client attempted to store that was the subject of the operation.

6. The system of claim 5 , wherein the data that is the subject of the operation and that the client attempted to store is user-data or parity data.

7. The system of claim 5 , wherein said system uses the data in the message, a parity equation associated with the data in the message and other data on one or more of the object storage devices, to correct data associated with the failure.

8. The method of claim 5 , wherein the distributed object-based system is a RAID system.

Assignments (8)