Degraded Redundant Metadata for Storage Node Upgrades
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During the upgrade of a storage node in a cluster, the existing data protection systems face challenges in maintaining continuous data protection, as the unavailability of one copy of data leaves the system vulnerable to loss, and the data migration process is costly in terms of time and bandwidth.
Innovation Solution
The degraded redundant metadata (DRuM) technique ensures continuous protection by maintaining multiple copies of data across storage nodes, redirecting new incoming data to a standby slice service during the upgrade of a storage node, allowing for failover and rebuilding of complete data copies without loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data replication is maintained during storage node upgrade, then data protection is improved, but system complexity and upgrade time increase due to data migration requirements
Solution Approach 1:
The system performs preliminary actions by pre-positioning a standby slice service on a different storage node before the upgrade begins. This standby service is configured in advance to receive data redirection, ensuring that when the upgrade starts, data can be immediately redirected without waiting for migration during the upgrade process.
Solution Approach 2:
The standby slice service acts as an intermediary that receives data redirected from the primary slice service during node upgrade. This intermediary service allows the primary node to be taken offline for upgrade while maintaining data availability through the standby service, eliminating the need for complex data migration during the upgrade process.
2Reliability
If data is replicated across multiple storage nodes, then data durability is improved, but bandwidth consumption increases during node upgrade due to data migration
Solution Approach 1:
The standby slice service is pre-configured on a different storage node before upgrade begins, positioned to receive redirected data. This preliminary setup eliminates the need for bandwidth-intensive data migration during upgrade, as data is simply redirected to the pre-positioned standby service rather than being transferred across the network.
Solution Approach 2:
Instead of migrating data during upgrade, the system uses a copy approach where the standby slice service creates a copy of the primary slice service's data by receiving redirected writes. This copying mechanism maintains data durability without requiring large bandwidth transfers, as the standby service replicates data through normal write operations rather than bulk migration.
3Ease of repair
If storage node is taken offline for upgrade, then maintenance is improved, but data protection is worsened due to loss of data copy
Solution Approach 1:
The standby slice service serves as an intermediary that maintains data protection during node maintenance. When the primary node is taken offline for upgrade, the standby service continues to receive and store data, ensuring that data protection is maintained without requiring the upgraded node to be online simultaneously.
Solution Approach 2:
The system maintains data protection by having a copy of the slice service running on a different storage node. This copy receives all data writes during the upgrade process, ensuring that data is continuously protected even though the primary node is offline for maintenance.
4Reliability
If data blocks are replicated among storage nodes, then data durability is improved, but device complexity increases during node upgrade
Solution Approach 1:
The system performs preliminary action by pre-configuring the standby slice service on a different storage node before upgrade. This pre-configuration simplifies the upgrade process, as the standby service is already in place and ready to receive redirected data, eliminating complex real-time coordination during upgrade.
Solution Approach 2:
The standby slice service acts as an intermediary that simplifies the upgrade process by absorbing the complexity of data redirection. Instead of the upgrade process needing to coordinate with multiple data blocks and replication mechanisms, the standby service simply receives redirected data writes, significantly reducing upgrade complexity.
Data Source
AI summary
A technique maintains multiple copies of data served by storage nodes of a cluster during upgrade of a storage node to ensure continuous protection of the data served by the nodes. The data is logically organized as one or more volumes on storage devices of the cluster and includes metadata that describe the data of each volume. A data protection system may be configured to maintain two copies of the data in the cluster during upgrade to a storage node that is assigned to host one of the copies of the data but that is taken offline during the upgrade. As a result, a slice service of the node may become unavailable during the upgrade. In response to the unavailability of the slice service, the technique redirects replicated data targeted to the slice service to a standby slice service according to a degraded redundant metadata (DRuM) service of the cluster.