File management program, file management process, and file management apparatus

A computer-readable medium storing a file management program which makes a computer perform processing for management of files in a storage device, the program comprising:

• a request reception unit which receives a processing request and determines whether a file to be processed is a directory file or a non-directory file;
• a hash-value generation unit, if the file to be processed in response to said processing request is the directory file, applies a predetermined hash function to the name information, and generates a hash value of the file to be processed;
• a physical-block-number acquisition unit which, if the file to be processed in response to said processing request is the non-directory file, acquires a logical-block number of data to be accessed with reference to a dinode, and looks up an extent table indicating correspondence between logical-block numbers and physical-block numbers to acquire a physical-block number of the data to be accessed, and if the file to be processed in response to said processing request is the directory file, looks up, in place of the extent table, a hash-value-range table containing at least one hash-value-range information set each of which is constituted by a hash-value range defining a range of hash values and a physical-block number uniquely indicating a data storage region in said storage device, extracts one of the at least one hash-value-range information set in which a hash-value range containing the hash value of the name information is included, and acquires a physical-block number from the one of the at least one hash-value-range information set;
• and
• a directory processing unit which, if the file to be processed in response to said processing request is the directory file, performs processing of a physical block in said storage device corresponding to said physical-block number acquired by said physical-block-number acquisition unit, where the processing relates to a record containing said name information and is performed in accordance with said processing request,
• wherein when said processing request instructs deletion of said name information, said directory processing unit deletes said record containing said name information from data acquired from said physical block, and puts the data from which the record is deleted, back into said physical block;
• wherein after said record is deleted from said data, said directory processing unit determines whether or not combining of the at least one hash-value-range information set is necessary, and combines said one of the at least one hash-value-range information set with another of the at least one hash-value-range information set when the combining of said at least one hash-value-range information set is determined to be necessary;
• and
• wherein when an amount of data remaining in said physical block reaches a predetermined lower limit by deletion of said record, said directory processing unit determines that combining of said at least one hash-value-range information set is necessary.

A file management process for managing files in a storage device by using a computer, comprising the steps of:

• (a) receiving a processing request and determining whether a file to be processed is a directory file or a non-directory file, by a request reception unit;
• (b) if the file to be processed in response to said processing request is the directory file, applying a predetermined hash function to name information contained in said processing request, and generating a hash value of the name information, by a hash-value generation unit;
• (c) if the file to be processed in response to said processing request is the non-directory file, acquiring a logical-block number of data to be accessed with reference to a dinode, and looking up an extent table indicating correspondence between logical-block numbers and physical-block numbers to acquire a physical-block number of the data to be accessed, and if the file to be processed in response to said processing request is the directory file, looking up, in place of the extent table, a hash-value-range table containing at least one hash-value-range information set each of which is constituted by a hash-value range defining a range of hash values and a physical-block number uniquely indicating a data storage region in said storage device, extracting one of the at least one hash-value-range information set in which a hash-value range containing the hash value of the file to be processed is included, and acquiring a physical-block number from the one of the at least one hash-value-range information set, by a physical-block-number acquisition unit;
• and
• (d) if the file to be processed in response to said processing request is the directory file, performing, by a directory processing unit, processing of a physical block in said storage device corresponding to said physical-block number acquired by said physical-block-number acquisition unit, where the processing relates to a record containing said name information and is performed in accordance with said processing request,
• wherein said directory processing unit determines whether or not said record can be inserted into said data, and inserts the record into said data when the record can be inserted into the data;
• and
• when the record cannot be inserted into the data, said directory processing unit divides said hash-value range containing the hash value of the file to be processed, into a plurality of hash-value ranges, updating said hash-value-range table by registering in the hash-value-range table a plurality of hash-value-range information sets corresponding to the plurality of hash-value ranges, and performs processing for insertion of said record by looking up the updated hash-value-range table.

A file management apparatus for managing files in a storage device, comprising:

• a request reception unit stored in a metadata server which receives a processing request and determines whether a file to be processed is a directory file or a non-directory file;
• a hash-value generation unit which, if the file to be processed in response to said processing request is the directory file, applies a predetermined hash function to the name information, and generates a hash value of the file to be processed;
• a physical-block-number acquisition unit which, if the file to be processed in response to said processing request is the non-directory file, acquires a logical-block number of data to be accessed with reference to a dinode, and looks up an extent table indicating correspondence between logical-block numbers and physical-block numbers to acquire a physical-block number of the data to be accessed, and if the file to be processed in response to said processing request is the directory file, looks up, in place of the extent table, a hash-value-range table containing at least one hash-value-range information set each of which is constituted by a hash-value range defining a range of hash values and a physical-block number uniquely indicating a data storage region in said storage device, extracts one of the at least one hash-value-range information set in which a hash-value range containing the hash value of the name information is included, and acquires a physical-block number from the one of the at least one hash-value-range information set;
• and
• a directory processing unit which, if the file to be processed in response to said processing request is the directory file, performs processing of a physical block in said storage device corresponding to said physical-block number acquired by said physical-block-number acquisition unit, where the processing relates to a record containing said name information and is performed in accordance with said processing request,
• wherein said directory processing unit determines whether or not said record can be inserted into said data, and inserts the record into said data when the record can be inserted into the data;
• and
• when the record cannot be inserted into the data, said directory processing unit divides said hash-value range containing the hash value of the file to be processed, into a plurality of hash-value ranges, updating said hash-value-range table by registering in the hash-value-range table a plurality of hash-value-range information sets corresponding to the plurality of hash-value ranges, and performs processing for insertion of said record by looking up the updated hash-value-range table.

A file management process for management of files in a storage device, the process comprising:

• receiving a processing request and determining whether a file to be processed is a directory file or a non-directory file;
• generating a hash value of name information contained in said processing request by applying a predetermined hash function to the name information if the file to be processed in response to said processing request is the directory file;
• acquiring a physical-block-number by acquiring a logical-block number of data to be accessed with reference to a dinode and looking up an extent table indicating correspondence between logical-block numbers and physical-block numbers to acquire a physical-block number of the data to be accessed if the file to be processed in response to said processing request is a non-directory file, and if the file to be processed in response to said processing request is the directory file, looking up a hash-value-range table containing at least one hash-value-range information set, each of which is constituted by a hash-value range defining a range of hash values and a physical-block number uniquely indicating a data storage region in said data storage device, extracting one of the at least one hash-value-range information sets in which a hash-value range containing the hash value of the name information is included, and acquiring a physical-block number from the one of the at least one hash-value-range information sets;
• and
• processing a physical block in said storage device corresponding to said physical-block number acquired by physical-block-number acquisition unit if the file to be processed in response to said processing request is the directory file, where the processing relates to a record containing said name information and is performed in accordance with said processing request,
• wherein when said processing request instructs deletion of said name information, said record containing said name information is deleted from data acquired from said physical block, and the data from which the record was deleted is put back into said physical block,
• wherein after said record is deleted from said data, it is determined whether or not combining of the at least one set hash-value-range information sets is necessary, and said one the at least one hash-value-range-information sets is combined with another of the at least one hash-value-range information sets when the combining of said at least one hash-value-range information sets is deemed to be necessary, and
• wherein when an amount of data remaining in said physical block reaches a predetermined lower limit by deletion of said record, it is determined that combining said at least one hash-value-range information sets is necessary.

A file management apparatus for management of files in a storage device, comprising:

• a request reception unit stored in a metadata server which receives a processing request and determines whether a file to be processed is a directory file or a non-directory fi1e;
• a hash-value generation unit that generates a hash value of name information contained in said processing request if the file to be processed in response to said processing request is the directory file by applying a predetermined hash function to the name information;
• a physical-block-number acquisition unit that, if the file to be processed in response to said processing request is the non-directory file, acquires a logical-block number of data to be accessed with reference to a dinode, and looks up an extent table indicating correspondence between logical-block numbers and physical-block numbers to acquire a physical-block number of the data to be accessed, and if the file to be processed is the directory file, acquires a physical-block-number by looking up a hash-value-range table containing at least one hash-value-range information set, each of which is constituted by a hash-value range defining a range of hash values and a physical-block number uniquely indicating a data storage region in said data storage device, extracting one of the at least one hash-value-range information sets in which a hash-value range containing the hash value of the name information is included, and acquiring a physical-block number from the one of the at least one hash-value-range information sets;
• and
• a directory processing unit that, if the file to be processed in response to said processing request is the directory file, processes a physical block in said storage device corresponding to said physical-block number acquired by physical-block-number acquisition unit, where the processing relates to a record containing aid name information and is performed in accordance with said processing request,
• wherein when said processing request instructs deletion of said name information, said directory processing unit deletes said record containing said name information from data acquired from said physical block, and puts the data from which the record was deleted back into said physical block,
• wherein after said record is deleted from said data, said directory processing unit determines whether or not combining the at least one hash-value-range information sets is necessary, and combines said one of the at least one hash-value-range-information sets with another of the at least one hash-value-range information sets when the combining of said at least one hash-value-range information sets is determined to be necessary, and
• wherein when an amount of data remaining in said physical block reaches a predetermined lower limit by deletion of said record, said directory processing unit determines that combining said at least one hash-value-range information sets is necessary.