RAID Expansion Failure Avoiding Extent Shuffling

Solution Overview

Problem

The expansion of a RAID storage pool can fail, leading to a rollback process that takes a long time, significantly impacting I/O performance as disk extents are shuffled back, causing delays and performance issues.

Innovation Solution

A method and device for managing RAID that prevent the triggering of a shuffling operation upon failure of expansion, allowing for immediate notification of failure without initiating the shuffling process, thus avoiding the lengthy rollback and maintaining I/O performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a shuffling operation is triggered during RAID expansion failure rollback, then data integrity is maintained through proper redistribution of disk extents, but I/O performance is significantly degraded and the process takes a long time (many days)

Engineering Contradiction:

Improvedata integrityVSAvoidI/O performance

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent extracts the harmful shuffling operation from the rollback process. When expansion fails, the system now skips the shuffling step that redistributes disk extents, directly shrinking the RAID volume instead. This removes the source of I/O performance degradation while maintaining data integrity through alternative rollback mechanisms.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent inverts the traditional rollback sequence. Instead of first shuffling disk extents and then shrinking, the system directly shrinks the RAID volume upon expansion failure. This reversal eliminates the performance-degrading shuffling operation while achieving the same rollback goal.

Inventive Principle:

Principle #13The other way round (Inversion)

2Stability of the object's composition

If a shuffling operation is performed during rollback, then disk extents are properly redistributed across storage devices, but the time consumed is excessive (many days)

Engineering Contradiction:

Improvedisk extent distributionVSAvoidrollback time

Core Design Contradiction:

Stability of the object's compositionVSLoss of time

Solution Approach 1:

The patent removes the shuffling operation from the rollback process entirely. When expansion fails, the system directly shrinks the RAID volume without performing the time-consuming shuffling of disk extents, reducing rollback time from many days to a much shorter duration.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs the shrink operation immediately upon detecting expansion failure, before any shuffling could occur. This preliminary action of direct shrinking eliminates the need for the subsequent time-consuming shuffling process while maintaining system stability.

Inventive Principle:

Principle #10Preliminary action

3Reliability

If the expansion rollback process is executed completely with shuffling, then the RAID structure is properly restored, but I/O performance is greatly affected during the process

Engineering Contradiction:

ImproveRAID structure restorationVSAvoidI/O performance

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent extracts the harmful shuffling operation from the rollback process. When expansion fails, the system now skips the shuffling step that redistributes disk extents, directly shrinking the RAID volume instead. This removes the source of I/O performance degradation while maintaining data integrity through alternative rollback mechanisms.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent inverts the traditional rollback sequence. Instead of first shuffling disk extents and then shrinking, the system directly shrinks the RAID volume upon expansion failure. This reversal eliminates the performance-degrading shuffling operation while achieving the same rollback goal.

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS11119705B2Method, device, and computer readable storage medium for managing redundant array of independent disks

Publication Date: 2021.09.14 EMC IP HLDG CO LLC

AI summary

Techniques manage a redundant array of independent disks. In such a technique, an instruction to expand the RAID is received, the RAID being formed by a first group of storage devices and including a first group of RAID extents. A second group of RAID extents created on a second group of storage devices and logical units are created on the second group of RAID extents. In response to receiving an indication for a failure of the creation of the second group of RAID extents or the logical units, a notification is generated for indicating a failure of the expansion of the RAID without triggering a shuffling operation, the shuffling operation causing storage extents in the first and second groups of RAID extents to be distributed in the first and second groups of storage devices.