MemFS 2.0 - A Memory-based File System on HP-UX 11iv3

 All the inodes are pre-allocated during the creation of file system, which wastes kernel

memory.

 The maximum file system size and file size of MemFS is restricted to 256GB due to the

design of inodes in HFS and cannot grow up to the size of available system memory.

 It does not scale well for large directories, as the directory entries are stored and

searched sequentially by HFS, which in turn degrades the performance of operations like

create, search on large directories.

 Overheads during allocation of blocks for files and directory

 Pre-allocates virtual memory of file system size (similar to RAM disk) during creation of file

system, which can cause the system to run out of virtual memory.

A new light weight layout is introduced in MemFS on HP-UX 11iv3 to overcome the above

problems, by storing metadata and data in better way to track them faster. The MemFS is entirely

implemented in the kernel, without requiring a separate user process. Metadata are stored in the

non-page able kernel dynamic memory and file data are kept in the Unified File Cache (UFC).

UFC is enhanced to automatically swap the MemFS file data into system swap device under

system memory/UFC pressure. The layout just contains minimum necessary information to provide

basic functionality of a file system.

Design

The memory file-based system which is available as a configurable product in HP-UX 11iv3 is

designed based on the various caches used by the operating system. Figure 1 is an illustration of

how MemFS is implemented on HP-UX 11iv3.

Figure 1. MemFS Design

The metadata needed are carved from the kernel memory (base cell memory). The data in the files

on memory file systems are cached in UFC (Unified file cache) similar to other file systems. But the

data are written on to swap devices transparently by UFC during memory pressure.

VFS

Super

block

Root inode

Hash

table of