Storage Array ACL I/O Fencing for Failed Node Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In computer clusters, a node that has failed can still access shared resources, leading to data corruption, and existing solutions like SCSI3 PR, FC switch-based disabling, and power fencing have limitations such as requiring additional hardware or being protocol-dependent.
Innovation Solution
Implementing an access control list (ACL) on a logical unit number (LUN) in a storage array to identify and isolate a failed node by removing its access rights, using a REST framework to manage host access and prevent further access to shared resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If SCSI3 PR or FC switch-based disabling is used for I/O fencing, then the failed node can be isolated from shared resources, but additional hardware or protocol-specific infrastructure is required
Solution Approach 1:
The patent extracts the fencing functionality from hardware-dependent mechanisms (SCSI3 PR, FC switch disabling) and relocates it to the storage array's software layer through access control lists. This removes the need for additional hardware while maintaining the isolation of failed nodes from shared resources
Solution Approach 2:
The access control list mechanism provides universal fencing capability across different protocols (FC, iSCSI, NVMe) and storage arrays without requiring protocol-specific hardware. The same ACL infrastructure handles multiple protocols and fencing scenarios, eliminating protocol dependency
2Reliability
If SCSI3 PR or FC switch-based disabling is used for I/O fencing, then the failed node can be isolated from shared resources, but the solution becomes protocol-dependent
Solution Approach 1:
The access control list mechanism provides universal fencing capability across different protocols (FC, iSCSI, NVMe) and storage arrays without requiring protocol-specific hardware. The same ACL infrastructure handles multiple protocols and fencing scenarios, eliminating protocol dependency
3Reliability
If traditional fencing methods are used, then failed nodes can be isolated, but the implementation becomes complex and hardware-specific
Solution Approach 1:
The patent extracts the fencing functionality from hardware-dependent mechanisms and relocates it to the storage array's software layer through access control lists, simplifying implementation while maintaining reliability
Solution Approach 2:
The storage array automatically manages ACL updates based on cluster node status. The system self-services by monitoring node health and dynamically adjusting access permissions without manual intervention or complex external hardware control
Data Source
AI summary
A method, computer program product, and computer system for creating, by a computing device, a logical unit number (LUN) on a storage array node of a storage system. An identifier of the LUN of the storage array node may be provided to a computing system, wherein the computing system includes one of a host, a server, and the storage array node. An access control list (ACL) of the computing system may be created. The ACL of the computing system may be applied to the LUN based upon, at least in part, the identifier. The LUN may be discovered and mapped at the computing system. It may be determined that the computing system has failed. The ACL of the computing system that has failed may be removed from the LUN to prevent the computing system that has failed from accessing the LUN.