Browse Prior Art Database

Method for Providing Enhanced File System Control and Integrity

IP.com Disclosure Number: IPCOM000124033D
Original Publication Date: 2005-Apr-06
Included in the Prior Art Database: 2005-Apr-06
Document File: 2 page(s) / 90K

Publishing Venue

IBM

Abstract

This article describes a method for providing enhanced file system control and integrity.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Method for Providing Enhanced File System Control and Integrity

The Virtual File System (VFS) is an interface specification for the UNIX* file interface. The VFS interface provides abstract objects, vnodes, which represent real objects in a file system. Vnodes can represent directories, files, FIFOs, sockets etc. These vnodes are managed in a single pool of vnodes. Managing all of the vnodes in a single large pool presents several problems:

All file systems compete for resources from this single pool of vnodes. Any required synchronization will cross these file systems. For example, when a vnode is allocated for file system 'A' at the same time that a vnode is being allocated for file system 'B' there will be competition for the vnode pool resources. A certain level of synchronization is required to ensure the integrity of the pool of vnodes. Heavy vnode activity in one file system can negatively impact the performance in another file system.

Vnode corruption can occur when errant file system code incorrectly modifies field(s) within a vnode. For example, file system code may write to fields beyond the end of a vnode being used for file system 'A' and this can easily cause corruption in a vnode being used for file system 'B'. Incorrect manipulation of vnode reference counts is another example of vnode corruption that can be caused by errant file system code. When this situation occurs a system IPL is generally required which can cause costly interruptions to a customer's business operations.

When a dynamically mountable file system is unmounted, there is generally a certain level of vnode cleanup processing that must occur for this file system. For example, the file system must be quiesced and all of the vnodes currently in use for that file system must be located and returned to the vnode pool. Finding all the vnodes that belong to a particular file system can be complicated and time consuming.

     Instead of managing the vnodes for all of the file systems in a single pool of vnodes, each file system is assigned a vnode pool dedicated for use by that particular file system. This approach provides the following benefits:

There is no competition across file systems for resources from the same vnode pool. For example, file system 'A' will be assigned its own dedicated vnode pool and file system...