The evolution of storage solutions on Linux
Summary
Lenz Grimmer traces the path from local file systems through LVM and cluster file systems to distributed, parallel and fault-tolerant Linux storage. Every additional storage layer trades simplicity for flexibility or resilience. Distribution removes single hardware failures but creates new consistency and operational problems.
Ideas
- Local file systems optimise consistency and performance within a single machine.
- LVM separates logical storage allocation from physical disks.
- Cluster file systems coordinate parallel access by several machines.
- Distributed systems scale capacity and redundancy across many nodes.
Recommendations
- Choose a storage architecture by its failure model rather than by maximum features.
- Regularly test data loss, network partitions and complete recovery.
References
Links to the original source and the Web Archive open in a new tab.