The Zebra striped network file system
ACM Transactions on Computer Systems (TOCS)
Serverless network file systems
ACM Transactions on Computer Systems (TOCS) - Special issue on operating system principles
Fibre channel: gigabit communications and I/O for computer networks
Fibre channel: gigabit communications and I/O for computer networks
Petal: distributed virtual disks
Proceedings of the seventh international conference on Architectural support for programming languages and operating systems
Metadata Management of the SANtopia File System
ICPADS '01 Proceedings of the Eighth International Conference on Parallel and Distributed Systems
Volume Management in SAN Environment
ICPADS '01 Proceedings of the Eighth International Conference on Parallel and Distributed Systems
Design and Implementation of an SAN System Based on the Fiber Channel Protocol
IEEE Transactions on Computers
Hi-index | 0.00 |
Logical Volume Managers (LVM) have been a key component of storage systems. They can support volume online resizing and reconfiguration, snapshots, and various software RAID levels. But most logical volume managers are only suitable for single system environments. They are not adequate for Storage Area Network (SAN) environment where cluster hosts share same storage resources. In this paper, we present a cluster LVM (cLVM) for SAN environments, and explain the design and architecture of the cLVM in detail. The cLVM’s key techniques include its management mechanism, device mapping method, metadata synchronization between cluster hosts and fault tolerance. The management mechanism controls all the cLVM operations and all of the storage resources. It provides a uniform view of the cLVM to users and includes a user-friendly interface. The fault tolerance is designed to deal with disk failures and to manage mode or cluster node shutdowns. One agent is set as the management node for synchronizing the metadata in the kernel space between the different nodes in the cLVM. Through the metadata synchronization and the device mapping method, the cLVM provides the file systems with a uniform virtualized view of the storage devices and supports cluster host environments. It achieves online resizing, movement of data between partitions and snapshots. We implemted and tested the cLVM. The result shows that it is also compatible with ordinary common single file systems or GFS. It has comparatively better fault tolerance than the LVM currently in use. The new cLVM system supports the computing needs of large enterprises and provides high scalability and high availability in the SAN environment.