The overlayfs file system is an implementation of a union file system that makes several file systems appear as a single file system when mounted. An overlayfs file system consists of a lower file system and an upper file system which share a single file system namespace. After a file is opened in an overlayfs file system, all operations go directly to the underlying lower or upper file systems, which simplifies the implementation and allows native performance compared to other union file system implementations. A typical use case is to use a read-only OS image as the lower file system and a writeable RAM-backed file system as the upper file system. Modified data is written to the upper file system only and not to the OS image.

Both the upper and lower file systems can be directory trees within the same file system and neither needs to be the root of a file system. The lower file system can be any supported file system, including an overlayfs file system, and does not need to be writable. If the upper file system is writable, as is usually the case, it must support the creation of trusted.* extended attributes and it must provide valid d_type file type in the direct structure returned by readdir(). For example, an NFS file system cannot be used for the upper file system.

The overlayfs file system is not available with UEK R3.

For more information, see https://www.kernel.org/doc/Documentation/filesystems/overlayfs.txt.