Methods for operating a log device

Claim 1

A method of storing and retrieving data by a computer system executing an operating system and supporting first and second persistent storage devices, said operating system including a filesystem module coupled through a first device driver to said first persistent storage device to transfer filesystem data blocks, said filesystem module providing support for synchronous write transactions, said method comprising the steps of: a) providing a second device driver selectively coupled between said filesystem module and said first device driver; b) collecting a predetermined set of data blocks provided from said filesystem module as part of a synchronous write transaction into a data segment, storing said data segment on said second persistent storage device, and signaling completion of said synchronous write transaction to said filesystem module, wherein said step of collecting further provides for 1) constructing a map relating said predetermined set of data blocks to said data segment, whereby said second device driver can identify said data segment by reference to any of said predetermined set of data blocks by said filesystem module; and 2) storing said map, including progressively updated versions of said map, on said second persistent storage device, whereby said map can be reconstructed from said second persistent storage device; c) migrating said predetermined set of data blocks from said second persistent storage device to said first persistent storage device by said second device driver through use of said first device driver independent of said filesystem module, wherein said data segment includes a predetermined set of address references for said predetermined set of data blocks and wherein said step of migrating said predetermined set of data blocks to said first persistent storage device is performed by said second device driver dependant on said predetermined set of address references; d) retrieving any of said predetermined set of data blocks from said first persistent storage device; e) selectively bypass writing said predetermined set of data blocks, as provided from said filesystem module, through said first device driver to said first persistent storage device; and f) selectively bypass reading said data segment from said second persistent storage device to transfer any of said predetermined set of data blocks to said filesystem module.

Claim 2

A method of storing and retrieving data by a computer system executing an operating system and supporting first and second persistent storage devices, said operating system including a filesystem module coupled through a first device driver to said first persistent storage device to transfer filesystem data blocks, said filesystem module providing support for synchronous write transactions, said method comprising the steps of: a) providing a second device driver selectively coupled between said filesystem module and said first device driver; b) collecting a predetermined set of data blocks provided from said filesystem module as part of a synchronous write transaction into a data segment, storing said data segment on said second persistent storage device, and signaling completion of said synchronous write transaction to said filesystem module, wherein said step of collecting further provides for 1) constructing a map relating said predetermined set of data blocks to said data segment, whereby said second device driver can identify said data segment by reference to any of said predetermined set of data blocks by said filesystem module; and 2) storing said map, including progressively updated versions of said map, on said second persistent storage device, whereby said map can be reconstructed from said second persistent storage device; c) migrating said predetermined set of data blocks from said second persistent storage device to said first persistent storage device by said second device driver through use of said first device driver independent of said filesystem module, wherein said data segment includes a predetermined set of address references for said predetermined set of data blocks and wherein said step of migrating said predetermined set of data blocks to said first persistent storage device is performed by said second device driver dependant on said predetermined set of address references; d) retrieving any of said predetermined set of data blocks from said first persistent storage device; e) selectively bypass writing said predetermined set of data blocks, as provided from said filesystem module, through said first device driver to said first persistent storage device; and f) selectively bypass reading said data segment from said second persistent storage device to transfer any of said predetermined set of data blocks to said filesystem module.

Claim 5

A method of storing and retrieving data within a computer system that is coupleable to a plurality of disk drives, where said computer system executes an operating system, including a filesystem and a main filesystem device driver and where said filesystem defines a data layout for the storage of data blocks within the addressable storage space of a main disk drive, said method comprising the steps of: a) including a log device driver transparently coupleable between said filesystem and said main filesystem device driver and coupleable to a log disk drive; b) selectively transferring data blocks between said filesystem, said main disk drive and said log disk drive, wherein data blocks, identified by filesystem address and provided by said filesystem, are preferentially written to said log disk drive and wherein data blocks identified by filesystem address are preferentially read from said main disk drive for transfer to said filesystem; c) determining, by said log device driver, whether to transfer a predetermined data block, having a predetermined filesystem address, to said main disk drive or said log disk drive based on a data block transfer load comparison between said log device disk drive and said main disk drive, whereby said log device driver seeks to balance the data block transfer load of said log device and main disk drives.

Claim 6

A method of storing and retrieving data within a computer system that is coupleable to a plurality of disk drives, where said computer system executes an operating system, including a filesystem and a main filesystem device driver and where said filesystem defines a data layout for the storage of data blocks within the addressable storage space of a main disk drive, said method comprising the steps of: a) including a log device driver transparently coupleable between said filesystem and said main filesystem device driver and coupleable to a log disk drive; b) selectively transferring data blocks between said filesystem, said main disk drive and said log disk drive, wherein data blocks, identified by filesystem address and provided by said filesystem, are preferentially written to said log disk drive and wherein data blocks identified by filesystem address are preferentially read from said main disk drive for transfer to said filesystem; c) determining, by said log device driver, whether to transfer a predetermined data block, having a predetermined filesystem address, to said main disk drive or said log disk drive based on a data block transfer load comparison between said log device disk drive and said main disk drive, whereby said log device driver seeks to balance the data block transfer load of said log device and main disk drives.

Claim 19

A method of storing and retrieving data within a computer system that is coupleable to a plurality of disk drives, where said computer system executes an operating system, including a filesystem and a main filesystem device driver to provide a first addressable storage space, having a first data storage organization defined by said filesystem, within a main disk drive to allow for the storage of filesystem data blocks, said method comprising the steps of: a) providing a log device driver coupled between and effectively transparent to said filesystem and said main filesystem device driver, said log device driver providing a second addressable storage space within a log device disk drive and having a second data storage organization defined by said log device driver, said second addressable storage space providing for the storage of a predetermined data log segment that includes a predetermined filesystem data block; b) enabling the selective transfer of said predetermined filesystem data block from said filesystem to said log device disk drive, and from said main disk drive and said log device disk drive to said filesystem; and c) selecting to transfer said predetermined filesystem data block between said filesystem, said log device disk drive and said main disk drive to preferentially isolating, from the apparent perspective of said filesystem, read transfers of said predetermined data block to said main disk drive and write transfers of said predetermined data block to said log device disk drive, thereby enabling said operating system to transparently access and selectively transfer filesystem data blocks with respect to said second addressable storage space in place of said first addressable storage space.

Claim 20

A method of storing and retrieving data within a computer system that is coupleable to a plurality of disk drives, where said computer system executes an operating system, including a filesystem and a main filesystem device driver to provide a first addressable storage space, having a first data storage organization defined by said filesystem, within a main disk drive to allow for the storage of filesystem data blocks, said method comprising the steps of: a) providing a log device driver coupled between and effectively transparent to said filesystem and said main filesystem device driver, said log device driver providing a second addressable storage space within a log device disk drive and having a second data storage organization defined by said log device driver, said second addressable storage space providing for the storage of a predetermined data log segment that includes a predetermined filesystem data block; b) enabling the selective transfer of said predetermined filesystem data block from said filesystem to said log device disk drive, and from said main disk drive and said log device disk drive to said filesystem; and c) selecting to transfer said predetermined filesystem data block between said filesystem, said log device disk drive and said main disk drive to preferentially isolating, from the apparent perspective of said filesystem, read transfers of said predetermined data block to said main disk drive and write transfers of said predetermined data block to said log device disk drive, thereby enabling said operating system to transparently access and selectively transfer filesystem data blocks with respect to said second addressable storage space in place of said first addressable storage space.