PRIMECLUSTER Global File Services Configuration and Administration Guide 4.2 (Solaris(TM) Operating Environment)

1.3 Features of the GFS Shared File System

This section describes the features of the GFS Shared File System. For details, see Part 3, "Global File Services Shared File System" of this manual.

The GFS Shared File System has API compatibility with the UFS and the GFS Local File Systems and can be accessed simultaneously from multiple Solaris OE. The major functions of the GFS Shared File System are:

• Simultaneous shared access to the file or file system from multiple nodes
• Maintaining of consistency when file data is to be referenced or updated from multiple nodes
• File access is enabled by using the file cache of each node.
• File operation can be continued on another node if a node down occurs while maintaining file system consistency.
• High-speed file system recovery function
• High-speed input and output processing is enabled with contiguous block allocation to the file data area.
• I/O load distribution and mass storage files are enabled with multi partition configuration support.
• The file system data area expansion is enabled without file system reconstruction.
• File system operation with a GUI by using a Web browser

The GFS Shared File System can use only shared volume managed by GDS (Global Disk Services). Other type of disk driver is not applicable.

The GFS Shared File System is shared file system from multiple Solaris OE, so both switchover applications among two or more nodes and scalable applications among several nodes can run on this file system with same contents. When scalable applications are to be executed on multiple nodes, the relevant application can transfer the data on the GFS file system by using the conventional UNIX file API as is.

Also, because file operation can be continued on another node even if a node fails, the GFS Shared File System is particularly suitable for environments in which high availability is required for the file system.

The GFS Shared File System can be used with switchover applications of RMS Cluster Operation.

At this time, the file system can be accessed from standby node.

To use the GFS Shared File System with switchover applications of RMS Cluster Operation, do not specify the GFS Shared File System as mpoint resource. For details, see Chapter 23.9, "Using the GFS Shared File System for RMS Cluster Operation" of this manual.