Virtual Data Mover Failover Network Interface Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional VDM-style data storage systems face issues with data unavailability and potential loss due to duplicate IP addresses when a VDM fails, as they simply replicate changes from a production to a replica side without adequate control over network interfaces.
Innovation Solution
The system performs a network interface control operation that enables or disables network interfaces based on replication usage, as determined by a configuration file, to prevent duplicate IP addresses and ensure data availability, by creating a replacement VDM and controlling its network interfaces after a failover event.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a new VDM is created on the second array with the same IP addresses after a VDM failure, then data service continuity is maintained and switchover is transparent to host computers, but duplicate IP addresses are created that can cause data unavailability or data loss
Solution Approach 1:
The system performs preliminary actions by checking the replication status and configuring the replacement VDM's network interfaces before they are activated. The configuration file is examined in advance to determine whether to enable or disable network interfaces, preventing duplicate IP address conflicts before they occur.
Solution Approach 2:
The system uses feedback from the configuration file to dynamically control the network interface state. The replacement VDM monitors replication session status and adjusts its network interface availability accordingly, enabling interfaces only when safe to do so and disabling them when replication conflicts might occur.
2Productivity
If the replacement VDM enables all network interfaces after creation, then data access is restored quickly, but duplicate IP addresses may be created if the original VDM was part of a replication session
Solution Approach 1:
Before enabling network interfaces, the system performs preliminary checks of the configuration file to determine replication status. This preliminary action ensures that network interfaces are only enabled when it is safe to do so, preventing duplicate IP address conflicts while still enabling quick restoration when appropriate.
Solution Approach 2:
The network interface state is made dynamic rather than static. The replacement VDM's network interfaces are configured to be enabled or disabled based on real-time replication status information from the configuration file, allowing the system to adapt to different operational scenarios optimally.
3Reliability
If the system checks the configuration file to determine replication status before enabling network interfaces, then duplicate IP addresses are prevented, but additional processing time is required during VDM recovery
Solution Approach 1:
The system performs a targeted check of only the necessary configuration file information required to determine replication status, rather than a complete system-wide verification. This partial action approach prevents duplicate IP addresses while minimizing the time added to the recovery process.
Data Source
AI summary
A technique handles a VDM failover situation. The technique involves adjusting a configuration file on a first platform to indicate whether data managed by an initial VDM on that platform is being replicated to a second platform. The technique further involves, following a VDM failover event, creating a replacement VDM on the first platform to replace the initial VDM. The technique further involves, after the replacement VDM is created, performing an operation that controls interfaces provided by the replacement VDM. The operation enables the interfaces when the operation determines that the data managed by the initial VDM on the first platform was not being replicated to the second platform at the time of the event, and disables the interfaces when the operation determines that the data managed by the initial VDM on the first platform was being replicated to the second platform at that time of the event.