Storage Controller Rebuild Priority Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face performance degradation during failure recovery due to improper rebuild ordering, which is not adaptive to changes in I/O patterns, leading to prolonged performance decreases.
Innovation Solution
A storage system that dynamically adjusts rebuild priority based on access frequencies for both a first period and a shorter second period, allowing for responsive changes in rebuild ordering to match evolving I/O patterns, ensuring efficient data repair and recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If priority ordering is performed based on access frequency determined in advance by monitoring for a given period, then rebuilding can be performed according to a stable priority order, but the system cannot adapt to unexpected changes in I/O patterns, leading to prolonged performance degradation
Solution Approach 1:
The patent implements dynamic rebuild priority adjustment by introducing a short-period monitoring mechanism that operates alongside long-period monitoring. When access frequency patterns change unexpectedly during rebuilding, the system detects these changes through the short-period monitoring and adjusts the rebuild priority order in real-time, transforming the static rebuild process into a dynamic one that adapts to changing I/O patterns.
Solution Approach 2:
The system employs feedback mechanisms by continuously monitoring access frequencies during the rebuild process and using this information to adjust rebuild priorities. The short-period monitoring provides feedback about current access patterns, which feeds back into the rebuild control logic to modify the rebuild order, ensuring the system responds to actual usage patterns rather than relying solely on historical data.
2Device complexity
If rebuilding is performed based on long-period access frequency monitoring only, then the rebuild process is simple to manage, but storage system performance decreases for a longer period due to inability to respond to changed I/O patterns
Solution Approach 1:
The patent implements periodic monitoring at two different time scales: long-period monitoring for stable baseline access frequency determination, and short-period monitoring for detecting recent changes in access patterns. This multi-periodic approach allows the system to balance monitoring complexity with responsiveness, using the appropriate time scale for different phases of the rebuild process.
Solution Approach 2:
The system performs preliminary monitoring during the rebuild process to detect changes in access patterns before they significantly impact performance. By proactively detecting I/O pattern changes through short-period monitoring and adjusting priorities in advance, the system prevents prolonged performance degradation rather than reacting after damage has occurred.
3Quantity of substance
If the distributed parity method is used for data protection, then space efficiency is improved, but storage system performance is reduced due to necessary data repair processing when failures occur
Solution Approach 1:
The patent changes the parameter of rebuild priority dynamically based on access frequency patterns. Instead of using a fixed rebuild order, the system adjusts which regions are rebuilt first based on their current and historical access frequencies, optimizing the balance between completing necessary repair work and maintaining system performance during the rebuild process.
Data Source
AI summary
This invention provides a storage system enabling it to properly rebuild a storage device involved in failure. In the storage system, a controller repairs data for which an access request has been issued, returns a reply to the source of the access request, and stores the repaired data. As regards data for which access is not requested, the controller executes rebuilding of storage regions corresponding to rebuild management units in priority-based order and changes priority for executing the rebuilding, based on access frequencies for a first period and access frequencies for a second period that is shorter than the first period.