Shared Volume Transformation via Storage Controller Quiesce
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In shared-resource computing environments, transforming operations on a shared volume can render the resource unavailable to other host systems, leading to data corruption or undetected data loss, especially when one host system performs operations without ensuring other systems are not actively using the volume.
Innovation Solution
A method where an initiating host system sends a notification command to a storage controller, which broadcasts an interrupt to other host systems to vote on whether the transforming operation can proceed, ensuring all connected systems quiesce the shared volume before the operation, and notifies the initiating host when all systems have successfully completed the action, allowing the operation to proceed only if all systems agree.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a transforming operation is performed on a shared volume by one host system, then the operation can be completed, but the shared resource becomes unavailable to other connected host systems during the operation, potentially causing data corruption or undetected data loss
Solution Approach 1:
The storage controller sends a quiesce command to all host systems before performing the transforming operation, requiring them to complete current I/O operations and flush buffers. This preliminary action ensures data is safely written before the transformation begins, preventing data corruption while allowing the operation to proceed.
Solution Approach 2:
The storage controller acts as an intermediary between the initiating host system and other connected host systems. It coordinates the quiescing process, collects responses from all hosts, and manages the transforming operation, ensuring that no host system accesses the volume during the critical transformation period, thus preventing data corruption.
2Ease of operation
If the shared resource remains available to host systems during the transforming operation, then access continuity is maintained, but data corruption or undetected data loss can occur
Solution Approach 1:
The storage controller mediates between access continuity and data integrity by coordinating a controlled quiescing process. It notifies all host systems of the upcoming transformation, allows them to complete current operations, and then temporarily suspends access during the transformation. After completion, it restores access, thus maintaining overall access continuity while preventing data corruption during the critical transformation window.
Solution Approach 2:
The storage controller performs preliminary quiescing actions by sending commands to all host systems before the transformation begins. This allows host systems to safely complete current I/O operations and flush buffers, ensuring data integrity is maintained while transitioning to the transformed state, and then restores access afterward.
3Speed
If a host system performs a transforming operation without notifying other systems, then the operation can proceed quickly, but other systems may attempt to write to moved data sets, causing file directory corruption
Solution Approach 1:
The storage controller performs preliminary notification by sending quiesce commands to all host systems before the transforming operation begins. This ensures all systems are aware of the upcoming transformation and can take necessary precautions, such as flushing buffers and completing current operations, preventing them from writing to moved data sets during the transformation.
Solution Approach 2:
The storage controller implements feedback by collecting responses from all host systems regarding their quiesce status before proceeding with the transformation. This feedback mechanism ensures that all systems are properly quiesced and aware of the transformation, preventing them from attempting to write to moved data sets and causing file directory corruption.
Data Source
AI summary
In one embodiment, a computer-implemented method includes making a determination, by a computer, to perform a transforming operation on a shared volume. A notify command is sent to a storage controller of the shared volume, the notify command including an action to be taken by other hosts connected to the shared volume. A message is received from the storage controller, the message indicating that responses have been received from the other host systems connected to the shared volume. A determination is made as to whether to perform the transforming operation based on the responses. The transforming operation is performed on the shared volume in response to the responses from the other host systems indicating that the other host systems successfully performed the action specified in the notify command.