Browse Prior Art Database

System and method to improve rebuild performance of distributed RAID Disclosure Number: IPCOM000248791D
Publication Date: 2017-Jan-12
Document File: 6 page(s) / 147K

Publishing Venue

The Prior Art Database


This disclosure introduces a method to use the rebuild areas from the storage pool which may contain multiple Distributed RAID(Redundant Arrays of Independent Disks), not a single Distributed RAID. System will monitor the performance data and generate a most suitable target drives list to system. When one drive failure occurs ,the recovered data will be moved to the new target drives which may be from multiple Distributed RAID. With this disclosure most additional write operations on the failed RAID drives are avoided or reduced, and the second drive failure risk is reduced. The data security  and stability are all assured.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.


System and method to improve rebuild performance of distributed RAID

With a limitation life time of a physical drive, or because of an accidental or malicious activity, a disk may come to a failure. To recover data on the failure disk and ensure data security, the system needs to do a fast rebuilding. Most-modern storage system uses Distributed RAID architecture. Rebuild area is distributed on all disks in the Distributed RAID. When rebuild occurs by disk failure, the rebuild task is distributed to all drives in the array, rather than depend on one spare drive. Thus, compared with traditional RAID, writing is distributed to all disks in the raid, the rebuilding time can be reduced. But when disk failure occurs, all left drives in the Distributed RAID will be added burden to rebuild data, it will impact the original I/O on the array, what's more important, it will probably introduce drive failure again. When another drive failed, the data will be lost. To reduce the recovery time and improve the Distributed RAID performance during rebuild period, a system and method is introduced in this disclosure.

This disclosure introduces a method to find the most suitable rebuild areas from the storage pool to store the recovered data. It can reduce the rebuild time and improve the performance when distributed RAID is rebuilding.

This is the current rebuild mode of distributed RAID, it uses the rebuild areas of the left drives in the same array.


Figure 1

This is the new working mode with the new method, it will find target rebuild areas from drives in the storage pool, rather than the array.

To find the most suitable target drives, this disclosure introduces a module to monitor and record the IOPS (Input/Output Operations Per Second) and latency of all drives in the system. This module will generate a drive list which can provide target rebuild a...