System and method for removing a storage server in a distributed column chunk data store
AssigneeYahoo! Inc.
InventorRadha Krishna Uppala
Life of this patent
Estimating the term…
Abstract
Assuring recovery from failure of a storage server in a distributed column chunk data store of operably coupled storage servers, includes: partitioning a data table into chunks; implementing a distribution scheme with a specified level of redundancy for recovery of one or more failed servers among multiple storage servers; distributing the column chunks according to the distribution scheme; calculating column chunk parity; storing the calculated column chunk parity; managing metadata for the column chunk data store; and updating the metadata for distributing the column chunks among remaining storage servers upon receiving an indication to remove a storage serve.
Claims20
A computer-implemented method for assuring recovery from failure of a storage server in a distributed column chunk data store of operably coupled storage servers, said method comprising:
- partitioning a data table into column chunks, wherein a column chunk is a column of said data table partitioned using at least one column as a key;
- implementing a distribution scheme with a specified level of redundancy for recovery of one or more failed servers among multiple storage servers;
- distributing the column chunks according to the distribution scheme;
- calculating column chunk parity;
- storing the calculated column chunk parity;
- managing metadata for the column chunk data store;
- and
- updating the metadata for distributing the column chunks among remaining storage servers upon receiving an indication to remove a storage server.
An information processing system for assuring recovery from failure of a storage server in a distributed column chunk data store of operably coupled storage servers, comprising:
- a memory comprising computer executable instructions that when executed by a processor device, cause a computer to perform:
- partitioning a data table into column chunks, wherein a column chunk is a column of said data table partitioned using at least one column as a key;
- implementing a distribution scheme with a specified level of redundancy for recovery of one or more failed servers among multiple storage servers;
- distributing the column chunks according to the distribution scheme;
- calculating column chunk parity;
- storing the calculated column chunk parity;
- managing metadata for the column chunk data store;
- and
- updating the metadata for distributing the column chunks among remaining storage servers upon receiving an indication to remove a storage server;
- and the processor device.
Dependencies are inferred from the claim text.