SAS Driver Firmware Upgrade Notification Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The process of upgrading firmware for disk array enclosures (DAEs) leads to a flood of notifications due to multiple restarts, overwhelming other services and causing performance degradation and recognition issues, especially in large-scale storage systems.
Innovation Solution
Implementing a framework that includes precheck logic, data path failover, SAS port state management, and DAE firmware upgrade start/monitor logic, with the introduction of an 'expander_only' state in the SAS driver to filter out disk online/offline events during the upgrade process, reducing the number of notifications propagated to upper layers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If firmware upgrade is performed for DAEs, then firmware version is updated to fix bugs and improve performance, but a flood of notifications is generated causing system performance degradation
Solution Approach 1:
The patent extracts and filters out topology change notifications during firmware upgrade operations. The notification filtering mechanism selectively removes upgrade-related notifications from the notification stream, preventing them from overwhelming the system while still allowing other critical notifications to pass through. This resolves the contradiction by extracting the harmful element (excessive notifications) while preserving the beneficial element (firmware updates).
Solution Approach 2:
The patent introduces an intermediary notification filtering mechanism between the firmware upgrade process and the notification processing services. This intermediary layer intercepts notifications, identifies those related to firmware upgrades, and filters them before they reach the notification processing services. This mediator prevents the flood of notifications from degrading system performance while allowing the firmware upgrade to proceed normally.
2Reliability
If DAEs are restarted multiple times during firmware upgrade, then firmware changes are properly applied, but services monitoring DAEs are overwhelmed by restart notifications
Solution Approach 1:
The patent extracts notifications specifically related to DAE restarts during firmware upgrades and filters them out from the notification stream. By identifying and removing these specific notification types, the system avoids overwhelming monitoring services while still ensuring that firmware changes are properly applied through the necessary restarts.
Solution Approach 2:
The patent implements preliminary action by pre-configuring the notification filtering mechanism to recognize and filter firmware upgrade-related notifications before they reach processing services. The system proactively identifies the upcoming firmware upgrade operation and prepares the filtering logic in advance, so that when restart notifications are generated, they are automatically filtered without requiring complex real-time analysis.
3Loss of information
If all topology change events are propagated to upper layers, then complete device status information is provided, but device creation and deletion processes are overwhelmed
Solution Approach 1:
The patent extracts and filters topology change notifications that occur during firmware upgrades, removing them from the stream of notifications propagated to upper layers. This selective extraction maintains complete device status information for non-upgrade operations while eliminating the excessive notifications during upgrades that would overwhelm device creation and deletion processes.
Solution Approach 2:
The patent applies local quality by making the notification propagation behavior context-dependent. During firmware upgrade operations, the system changes the notification propagation characteristics to filter out upgrade-related topology changes. For normal operations, full topology change information is propagated. This local modification of notification behavior resolves the contradiction by adapting the information flow to the specific operational context.
Data Source
AI summary
A request is received to upgrade firmware for disk array enclosures (DAEs). Each DAE is reachable by a host server using first and second paths. The first path extends from a first port of a first HBA of the host server to each of a set of first input/output modules (IOMs) in the DAEs. The first IOMs have first SAS expanders to which storage disks are connected. The second path extends from a second port of a second HBA of the host server to each of a set of second IOMs in the DAEs. The second IOMs have second SAS expanders to which the disks are connected. The first port is changed from a first state during which a SAS driver exposes all topology change events to an OS of the host server to a second state during which the SAS driver does not expose all of the topology change events.