Veritas Storage Foundation Cluster File System
| Trialware Purchasing Data Sheet (PDF) |
New Features
Veritas Storage Foundation Cluster File System includes major enhancements including file lock management, clustered multi-pathing, I/O fencing and multi-volume support.
Multiple Transaction Server
Distributes file locking ownership across all nodes in the cluster, ensuring near linear scalability without requiring unnecessary meta-data communication throughout the cluster. Provides performance and scalability benefits by distributing transaction updates.
- Provide primary like performance from every node.
- Provide scalability. Any one node does not become a bottleneck or limit scalability.
- A single file system scalable up to 32 nodes.
Enabling Parallel I/O with Range Locking
Traditionally, the entire file is locked to perform I/O to a small region. Parallel I/O is supported in CFS with Range Locking that locks ranges in a file that correspond to an I/O request page. Two I/O requests conflict if at least one is a write request, and the I/O range of the request overlaps the I/O range of the other.
- The granularity of the locked range is a page.
- If the file size can be predetermined, the file can be preallocated to avoid block allocations.
Dynamic Storage Tiering
Unimportant or out-of-date files can be moved to less-expensive storage devices without changing the way users or applications access those files. This move is completely transparent to the users and applications that own the files.
- Policies can be created for automating the movement of data.
- File can be moved dynamically without having to be taken offline.
Preventing Data Corruption with I/O Fencing
SFCFS now supports I/O fencing if the required storage configuration is present. I/O fencing prevents data corruption caused by a "split-brain" (network partion) in the cluster. I/O fencing allows write access to members of the active cluster and blocks access to non-members. The physical components of I/O fencing are data disks and coordinator disks.
- Requires arrays that support SCSI 3 Persistant Group Reservations.
- Prevents multiple nodes from overwriting data dute to loss of inter node communication.




