Sharing filesystems (and volume groups) between two or more systems guarentees corruption, usually resulting in a panic (system crash) to prevent further damage. To share filesystems requires a complex interaction between systems to coordinate updates and changes to various files, directories and metadata.
However, NFS handles all this quite well. You can share several filesystems with several clients.
We have a requirement that same mount point will be accessed by 3 hosts simultaneously , if 1 host fails , other two hosts still must be able to work with application , its DRM solution .
What if its not possible with NFS too , then how do we approach this requirement ?
What you are describing is a requirement for Service Guard. Two or more nodes can be clustered together such that the critical application and/or database will always be running. An NFS server, just like any single server (or the network path to the NFS server) is indeed vulnerable. Service Guard can cluster the 3 nodes, even if they are in different cities, and through storage replication, the data will always be available to the nodes. Your requirement, although sounding simple, requires a lot of planning and extra software products to accomplish. Without clustering software and storage replication, the switchover in case of a failure will be quite chaotic.
