International Business Machines Corporation

Daniel Lloyd McNabb

Frank Bernhard Schmuck

James Christopher Wyllie

Wayne A Sawdon

Roger Lee Haskin

Marc M Eshel

Martin G Kienzle

Robert J Curran

Boaz Shmueli

Scott Lyon Porter

Raymond Edward Rose

Demetrios K Michalaros

File system backups are carried out by first generating a list of inodes including associated inode numbers in inode number order that have changed since the last backup operation. A table which has file names and inode numbers for all of the files currently in the file system is also generated. This list is sorted by inode number and the table and list are merged to provide a structure for determining which files are to be backed up. This means that relevant inodes and file names are now provided in a single entity. It is also noted that the structure that results from the merge operation is particularly suitable for being read in blocks which thus permits the backup operation to be carried out in parallel. The task of backing up files is also preferably partitioned by file size or other criteria as opposed to being partitioned simply by the number of files assigned to be backed up by any one processor in a distributed or parallel data processing system.

Parallel high speed backup for a storage area network (SAN) file system

A computer system having a shared disk file system running on on multiple computers each having their own instance of an operating system and being coupled for parallel data sharing access to files residing on network attached shared disks. Methods are provided for use as a parallel file system in a shared disk environment by use of a scalable directory service for the system with a stable cursor, a segmented allocation map. Dynamic prefetch and cached balance pools for multiple accesses improve the system. Extended file attributes are used for implementation of Access Control Lists in a parallel file system. Improvements to caching and cache performance developments balance pools for multiple accesses. A metadata node manages file metadata for parallel read and write actions. For our system, tokens are used for metadata node selection and identification, and we have enhanced token modes for controlling file size, as well as smart caching of byte range tokens using file access patterns and a byte range lock algorithm using a byte range token interface. Locking techniques reduce the overhead of a token manager which is also used in the file system recovery if a computer participating in the management of shared disks becomes unavailable or failed.

Parallel file system and method for multiple node file access

A computer system having a shared disk file system running on multiple computers each having their own instance of an operating system and being coupled for parallel data sharing access to files residing on network attached shared disks. Methods are provided for use as a parallel file system in a shared disk environment by use of a scalable directory service for the system with a stable cursor, and a segmented allocation map. Dynamic prefetch and cached balance pools for multiple accesses improve the system. Extended file attributes are used for implementation of Access Control Lists in a parallel file system. Improvements to caching and cache performance developments balance pools for multiple accesses. A metadata node manages file metadata for parallel read and write actions. For managing data for the same file distributed across multiple disk devices, a buffer manager is provided for a file system buffer pool which arbitrates use of memory resources among different system components competing for memory with buffer information for a component that said buffer manager needs for deciding how much memory to allocate among components of the system. The buffer information includes a desired memory size and current activity level. The buffer manager provides to a system component data as to how much memory has been assigned for use by that component of the system, and it schedules account resources to maximize file system throughput.

Parallel file system and buffer management arbitration

An improved method for supporting an empirical measurement of the data bandwidth that one or more computer disks can continuously sustain while reading data. In a computer system having predetermined workload requirements and disk storage for storing an allocation table including entries for meta blocks and data blocks, a method for supporting a dynamic measurement of the read bandwidth of a disk, includes the steps of: creating an alternate allocation table that ignores the existing content of the disk; creating meta blocks for a phantom file by allocating the meta blocks from the allocation table; allocating data blocks for the phantom file from the alternate allocation table; and writing only meta blocks to the disk, but not actual data blocks. Since the actual data blocks are by far the largest component of the files, phantom files require only a very small number of disk blocks, and can be written very quickly. Since the creation of the alternate allocation table ignores the existing content of the disk, disks can be used regardless of their current content and without modifying this content. Further, a disk can be calibrated by reading phantom files--using all system components including hardware and software--in the same way that the application for which the calibration is intended would read ordinary files thereby avoiding any artifacts resulting from the measurement of the read bandwidth.

Phantom files for dynamic read bandwidth measurements of computer disks

A multimedia datastreaming system includes one of a plurality of differing pre-existing file systems each supporting a set of zero copy function interfaces. A method is provided for supporting the zero copy function of a datastream file with the pre-existing file system. A general purpose file system interface is established to the pre-existing file systems. The general purpose file system interface is then invoked to affect zero copy transfer of the datastream from a producer module to a consumer module. An initial prefetch of data associated with the datastream file is initiated in response to an I/O command issued from a multimedia file system daemon, such prefetch being from mass storage to a buffer in shared memory in an interrupt context.

Growing community of inventors

Los Gatos, CA, United States of America

Daniel Lloyd McNabb