Cluster member transfer for raid system expansion
AssigneeDell Products L.P.
InventorsKuolin Hua, Kunxiu Gao
Life of this patent
Estimating the term…
Abstract
In a storage system that implements RAID (D+P) with an existing cluster of drives in which the drives have (D+P) partitions that are protection group members, cluster member transfer code creates a new drive cluster when fewer than D+P new drives are added to the storage system. The cluster member transfer code moves one or more drives from the existing cluster into a new cluster so that the number of new drives plus the number of moved drives equals D+P. One or more protection groups may be moved to the new cluster.
Claims20
An apparatus, comprising:
- at least one computing node comprising a processor and non-transitory memory;
- an existing cluster of C drives managed by the at least one computing node, wherein redundant array of independent drives (RAID) protection groups each comprising no more or less than (D+P) members are implemented on the existing cluster, wherein D and P are constants and each D member contains data respectively being used to rebuild at least one parity member of a protection group to which the D member belongs and each P member of that protection group contains parity information respectively being used with remaining data to rebuild a data member of the protection group;
- and
- the processor, in response to introduction of N new drives, to:
- partition each new drive into (D+P) partitions;
- and
- calculate that N is less than (D+P) and C is greater than (D+P) and select a donor drive from the existing cluster and form a new cluster using the donor drive and the new drives.
A method implemented in a storage system comprising:
- at least one computing node having a processor and non-transitory memory, and an existing cluster of C drives managed by the at least one computing node, wherein redundant array of independent drives (RAID) protection groups each comprising no more or less than (D+P) members are implemented on the existing cluster, wherein D and P are constants and each D member contains data respectively being used to rebuild at least one parity member of a protection group to which the D member belongs and each P member contains parity information respectively being used with remaining data to rebuild a data member of the protection group, wherein:
- responsive to introduction of N new drives, partitioning each new drive into (D+P) partitions;
- calculating that N is less than (D+P) and C is greater than (D+P);
- selecting a donor drive from the existing cluster;
- and
- forming a new cluster using the donor drive and the new drives.
A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for using a computer system to increase storage capacity, the method comprising:
- responsive to introduction of N new drives to a system comprising an existing cluster of C drives managed by at least one computing node, wherein redundant array of independent drive (RAID) protection groups each comprising no more or less than (D+P) members that are implemented on the existing cluster, and wherein D and P are constants and each D member contains data respectively being used to rebuild at least one parity member of a protection group to which the D member belongs and each P member contains parity information capable of respectively being used with remaining data to rebuild a data member of the protection group, partitioning each new drive into (D+P) partitions;
- responsive to calculating that N is less than (D+P) and C is greater than (D+P), selecting a donor drive from the existing cluster;
- and
- forming a new cluster using the donor drive and the new drives.
Dependencies are inferred from the claim text.