Storage Unit Fault Handling in Distributed Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed storage systems, repeated IO scheduling and repairing processes waste system resources when a fault occurs in the same backup data on both primary and secondary servers, as existing methods fail to efficiently handle such scenarios.
Innovation Solution
A data processing method where a first server records information about storage units with repairable faults that cannot be repaired, queries whether the storage unit stores the data upon receiving a read request, and either returns a failure message or reads data from a different storage unit, thereby avoiding repeated ineffective repairs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the primary server or secondary server repeatedly schedules IO and restarts the fault repairing process when backup data is faulty on both servers, then the fault repairing may eventually succeed, but system resources are wasted due to repeated ineffective repairs
Solution Approach 1:
The patent applies preliminary action by recording information about storage units with repairable faults before repeated repair attempts are made. When a fault occurs, the system checks the recorded information to determine if this is a previously encountered unrepairable fault, thereby avoiding wasteful repeated repair processes. This proactive recording and checking mechanism prevents resource waste while maintaining reliable fault handling.
2Reliability
If the system performs repeated IO scheduling and fault repairing processes for storage units with persistent faults, then complete fault recovery may be achieved, but the processing time and system efficiency are reduced
Solution Approach 1:
The patent implements feedback by recording the results of fault repair attempts and using this information in subsequent repair decisions. When a storage unit with a recorded unrepairable fault is accessed, the system retrieves the recorded information and immediately determines that further repair attempts are futile, thus avoiding time-wasting repeated processes. This feedback mechanism ensures reliable fault recovery while minimizing time loss.
3Reliability
If the system continuously attempts to repair storage units with persistent faults through repeated IO scheduling, then fault tolerance is maintained, but device complexity increases due to multiple repair attempts and scheduling operations
Solution Approach 1:
The patent reduces device complexity by performing preliminary recording of unrepairable fault information. Instead of implementing complex logic to track and manage multiple repair attempts, the system simply records the fault information once and uses this recorded data to quickly determine whether further repair attempts are necessary. This approach maintains fault tolerance while significantly simplifying the repair process management.
Data Source
AI summary
A data processing method, a device, and a system for a storage unit in order to avoid performing repeated data read operations on the storage unit where the method includes recording, by a first server for at least one storage unit used to store data, information about a storage unit in which a repairable fault occurs but repairing fails, querying, by the first server each time a request for reading data is received, whether the storage unit corresponding to the recorded information about the storage unit stores a part or all of the data, and when the storage unit corresponding to the recorded information stores the data, returning, to a requesting party that sends the request, a message indicating that reading the data fails, or otherwise, reading the data from a storage unit that stores the data, and feeding back the data to the requesting party that sends the request.