Abstract: The present invention provides a recovery method using extendible hashing-based cluster logs in a shared-nothing spatial database cluster, which eliminates the duplication of cluster logs required for cluster recovery in a shared-nothing database cluster, so that recovery time is decreased, thus allowing the shared-nothing spatial database cluster system to continuously provide stable service. In the recovery method, if a failure occurs in a predetermined node, a second node in a group, including the node, records cluster logs in main memory on the basis of extendible hashing. If the node that has failed recovers itself using a local log, the second node in the group transmits cluster logs in packets to a recovery node that is the failed node. If the recovery node reflects the received cluster logs and maintains consistency with other nodes in the group, the recovery node resumes normal service.

Abstract: Disclosed herein is a method of designing a query classificaton component for a multilevel DataBase Management System (DBMS). The method includes determining whether a plan tree, which is used for query processing, is input; if the plan tree is input, designing the query classification component to output locations for respective tables, at which data must be referenced in conjunction with the plan tree, and corresponding predicates; and transferring the locations for respective tables and the corresponding predicates to respective storage managers and causing each of the storage managers to process corresponding data.

Abstract: A journaling method is provided for supporting a recovery when a system is abnormally terminated in a shared disk environment. When a system call operation to take part in a journaling is generated, in order to guarantee a recovery, a transaction is started and new transaction region is assigned. Then, a system is initialized and a transaction type is set up. Lock information on modified data is acquired and added to the transaction so that a transaction manages lock information. A reflection to a disk during a modification of metadata is prevented. Modified metadata added to the transaction and modified information on principal general data are recorded. Then, lock information connected to the transaction is released.

Abstract: The present invention provides a recovery method using extendible hashing-based cluster logs in a shared-nothing spatial database cluster, which eliminates the duplication of cluster logs required for cluster recovery in a shared-nothing database cluster, so that recovery time is decreased, thus allowing the shared-nothing spatial database cluster system to continuously provide stable service. In the recovery method, if a failure occurs in a predetermined node, a second node in a group, including the node, records cluster logs in main memory on the basis of extendible hashing. If the node that has failed recovers itself using a local log, the second node in the group transmits cluster logs in packets to a recovery node that is the failed node. If the recovery node reflects the received cluster logs and maintains consistency with other nodes in the group, the recovery node resumes normal service.

Abstract: The present invention relates to the method for managing directories of a large-scale file system. The present invention the steps of creating root blocks which have a fixed global depth, calculating an index value which shows the location information on how many bits will be referenced for root blocks from the global depth through a hash function when root blocks are completely filled with directories, calculating the block numbers of leaf blocks consisted of extent blocks based on a plurality fixed length platform which are logical and successive objects using the index value and storing the block numbers sequentially to the root blocks, and storing the directory entries of root blocks separately in leaf blocks. As a result, the present invention can improve the search performance by minimizing the cause of frequently change of data storage structure of the file system and flexibly preventing creation of indirect blocks.

Abstract: A system having RAID levels includes a storage medium, a memory and a CPU. The storage medium has at least two disks. The memory stores a striping zone information table of total disks in which a physical address of data recorded in a disk of the storage medium is converted into a logic address. The CPU modifies a striping zone information table stored in the memory in case a new disk is added to the storage medium, stores the modified striping zone information table in the memory, converts a logic address of data to be written/read out from the stored table into a physical address in response to the disk writing/reading instructions, searches positions of a corresponding disk D and a physical block B and controls writing or reading out data.

Abstract: A journaling method is provided for supporting a recovery when a system is abnormally terminated in a shared disk environment. When a system call operation to take part in a journaling is generated, in order to guarantee a recovery, a transaction is started and new transaction region is assigned. Then, a system is initialized and a transaction type is set up. Lock information on modified data is acquired and added to the transaction so that a transaction manages lock information. A reflection to a disk during a modification of metadata is prevented. Modified metadata added to the transaction and modified information on principal general data are recorded. Then, lock information connected to the transaction is released.

Abstract: A method for managing a logical volume for minimizing a size of metadata and supporting dynamic online resizing and software redundant array of independent disks (RAID), and a computer-readable recording medium storing instructions for embodying the method, are disclosed. The method includes the metadata having a disk partition table containing information of a disk partition in which the metadata is stored; a logical volume table for maintaining the information of the logical volume by storing duplicated information of the logical volume onto all disk partitions of the logical volume; an extent allocation table for indicating whether each extent in the disk partitions is used or not; and a mapping table for maintaining a mapping information for a physical address space corresponding to a logical address space which is a continuous address space equal in size of storage space of whole logical volume.

Abstract: The present invention relates to the method for managing directories of a large-scale file system. The present invention the steps of creating root blocks which have a fixed global depth, calculating an index value which shows the location information on how many bits will be referenced for root blocks from the global depth through a hash function when root blocks are completely filled with directories, calculating the block numbers of leaf blocks consisted of extent blocks based on a plurality fixed length platform which are logical and successive objects using the index value and storing the block numbers sequentially to the root blocks, and storing the directory entries of root blocks separately in leaf blocks.

Abstract: A system having RAID levels includes a storage medium, a memory and a CPU. The storage medium has at least two disks. The memory stores a striping zone information table of total disks in which a physical address of data recorded in a disk of the storage medium is converted into a logic address. The CPU modifies a striping zone information table stored in the memory in case a new disk is added to the storage medium, stores the modified striping zone information table in the memory, converts a logic address of data to be written/read out from the stored table into a physical address in response to the disk writing/reading instructions, searches positions of a corresponding disk D and a physical block B and controls writing or reading out data.

Abstract: A method for managing a logical volume for minimizing a size of metadata and supporting dynamic online resizing, computer-readable recoding medium storing a program or data structure for embodying the method are disclosed.