The present invention is related to data storage systems. More particularly, the present invention is directed to a method and system of reclaiming storage space in data storage systems.
Data storage systems utilize various mechanisms to identify available storage space. One approach is to maintain a “high water mark” to indicate a boundary between the portion of a data container that is unavailable, i.e., has been allocated for storage, and the portion of the data container that is available, i.e., has not been allocated for storage. The “high water mark” may be a pointer to the most recently allocated space in the data container. This mechanism optimizes the efficiency of allocating space by making it easy to locate available storage space.
The “high water mark” mechanism, however, may result in less efficient use of storage space in data storage systems because data objects that are initially very large frequently shrink in size. For example, a table that starts out with 1 million rows may end up with only a few hundred rows after several transactions. As a result, much of the storage space allocated for the table will be left unused. Since the “high water mark” only moves in one direction, the newly freed storage space below the “high water mark” will not be available to store other data objects as the system assumes that available storage space can only be found above the “high water mark.”
Having data containers with unused storage space scattered throughout can impact the performance of scans and DML (Data Manipulation Language) operations. Scans may be affected because the amount of storage space read may not be proportional to the data retrieved. In addition, the length of time it takes for a scan to complete may also impact various operations that are scan-based.
In OLTP (Online Transaction Processing) systems, large tables called staging tables are often used for staging data. For example, data may be inserted into the staging tables for pickling. Pickling is a process of transforming data from a source representation to a uniform target representation. A large amount of the data in the staging tables is frequently deleted after pickling. The space that has been allocated for the staging tables, however, may remain unavailable to other objects for a considerable period of time after the data has been deleted.
One method of making available, i.e., reclaiming, unused storage space in an existing data container that is below the “high water mark” is to create a new data container, allocate storage space for objects in the existing data container in the new data container, move those objects to the new data container, and delete the existing data container. This solution, however, requires extra storage space. Hence, addition equipment, e.g., data storage devices, disk drives, etc., may need to be purchased. In addition, the objects may be offline, i.e., inaccessible, during the reclamation process, which may not be acceptable to end-users. Furthermore, dependent objects may have to be recreated as a result of the reclamation process.
Thus, it is desirable to provide a method and system where unused storage space in data containers can be reclaimed in place, i.e., without requiring extra storage space, where concurrency is preserved (i.e., objects in the data containers remain accessible during the reclamation process), and where data dependencies are maintained.
The present invention provides a method and system of reclaiming storage space in data storage systems. In one embodiment, data in a data container is compacted. A high water mark of the data container is then adjusted and unused space in the data container is reclaimed.
Further details of aspects, objects, and advantages of the invention are described below in the detailed description, drawings, and claims. Both the foregoing general description and the following detailed description are exemplary and explanatory, and are not intended to be limiting as to the scope of the invention.
The accompanying drawings are included to provide a further understanding of the invention and, together with the Detailed Description, serve to explain the principles of the invention.
Reclamation of storage space in data storage systems is disclosed. Rather than move data in an existing data container into a new data container in order to reclaim unused storage space below a high water mark of the existing data container, unused storage space in the existing data container is reclaimed by compacting its content, which isolates the unused storage space in the existing data container, and adjusting the high water mark of the existing data container to exclude the emptied storage space. This allows for reclamation of freed storage space without needing to create a new data container, which requires extra storage space. In addition, concurrency of access and data dependencies can be maintained as well.
Data containers into which data is to be inserted may exist at any level of granularity. For example, a data container may be a table space, a file, a segment, a data block, or a row. A data storage system, at the finest level of granularity, may store data in rows. The rows, in turn, may be stored in one or more data blocks. Each data block may correspond to a specific number of bytes of physical space on a data storage device, e.g., a disk drive, etc. The next level of space is an extent. An extent may be a specific number of contiguous data blocks. Segments are the next level of granularity. Each segment may include one or more extents. Although in the following description, for purposes of explanation, specific types or granularity of data containers are set forth in order to provide a thorough understanding of the invention, the present invention is not limited to any particular type or granularity of data container.
Illustrated in
As the reclamation process proceeds, data in in-use space 208d, which is closest to high water mark 202a is moved to unused space 206a, which is farthest from high water mark 202a. Filling in unused space 206a frees up new space at high water mark's 202a end of data container 200. In the embodiment, unused space 206d expands as a result of the data movement.
Data in data container 200 remains online, i.e., accessible to users, when relocation of data is carried out transactionally. For example, a piece of data in space 208d can be inserted into space 206a before the same piece of data in space 208d is deleted. In one embodiment, a piece of data in space 208d may be deleted first before the piece of data is inserted into space 206a. For those embodiments, if movement of data is done incrementally, e.g., a row or a block at a time, other rows or blocks of data will remain accessible even when one row or block of data is locked. Thus, optimal concurrency can still be maintained.
In an embodiment, the granularity in which data is relocated depends upon the granularity at which the data storage system maintains data integrity. For example, if a data storage system maintains integrity at the row level, then data is moved on a row by row basis. Additionally, since data can be relocated using typical functions of the data storage system, e.g., insert operation, delete operation, etc., inherent maintenance functions in the system can automatically preserve the integrity of the data, e.g., data dependencies, indexes, constraints, triggers, etc., being moved. Moreover, because the relocation process uses typical data storage system functions, the process is recoverable, can be rolled back, etc.
Once unused space 206a is filled, shown in
As shown in
Scans and DML (Data Manipulation Language) operations may run in parallel with the data movement phase. Searches for space by DML operations may be synchronized with searches for unused space during the data movement phase. DML operation searches for space may attempt to optimize the search path by caching addresses of spaces where data has been successfully inserted. If those spaces are close to high water mark 202a, then the progress made by the data movement phase may be countered. Hence, spaces being emptied that are near high water mark 202a may be marked with a state that would discourage future use of the emptied spaces, e.g., prevent inserts from happening.
After all of the data in data container 200 are coalesced at one end of data container 200, high water mark 202a is adjusted to make unused space 206d reclaimable, i.e., high water mark 202a is moved to exclude space(s) that have been successfully emptied. As shown in
Depicted in
At 304, an in-use storage space in the data container is identified. The identified in-use storage space may be an in-use storage space closest to a high water mark of the data container. An unused storage space in the data container is selected at 306. The selected unused storage space may be an unused storage space farthest from the high water mark of the data container. Data in the identified in-use storage space are then moved, i.e., copied, relocated, etc., to the selected unused storage space (308).
Movement of data may be implemented in two stages. In one embodiment, the first stage will attempt to move as much data as possible. For instance, if a piece of data is locked, e.g., being accessed by another transaction, the process will skip that piece of data and proceed to move other pieces of data that are unlocked. In the embodiment, the second stage will wait for data to become unlocked in order to move data that was skipped in the first stage. In another embodiment, whenever the process realizes that it is competing against another transaction, e.g., a user transaction, for access to the same piece of data, the process will allow the other transaction to access the data first.
If the identified in-use storage space is emptied as a result of the movement of data to the selected unused storage space, the now empty identified storage space may be marked to prevent the storage space from being used again. In addition, if a bitmap is being maintained for the data container, the bitmap may be updated to indicate that the identified storage space is now empty. Tracking usage as the reclamation process progresses allows the reclamation process to restart where it left off on occasions where the process ends abruptly due to an error.
A determination is made at 310 as to whether any unused storage space in the data container is located in between two in-use storage spaces. If one or more such unused storage spaces remain, the process returns to 304. When none of the unused storage spaces in the data container is located in between two in-use storage spaces, the high water mark of the data container is adjusted to exclude all of the unused storage spaces in the data container (312). The unused storage spaces in the data container are then reclaimed at 314.
Since unused storage space 406b is now large enough to accommodate all of the data in in-use storage space 408c, all of the data in in-use storage space 408c are moved to unused storage space 406b. A new unused storage space 406c is then created as seen in data container 400c. High water mark 402a can now be adjusted as none of the unused storage spaces in data container 400c is in between two in-use storage spaces. After adjustment of high-water mark 402a to the beginning of unused storage space 406c, reclaimable space 404a enlarges to become reclaimable space 404b. Reclaiming reclaimable space 404b results in data container 400e.
System Architecture Overview
According to one embodiment of the invention, computer system 500 performs specific operations by processor 504 executing one or more sequences of one or more instructions contained in system memory 506. Such instructions may be read into system memory 506 from another computer readable medium, such as static storage device 508 or disk drive 510. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions to implement the invention.
The term “computer readable medium” as used herein refers to any medium that participates in providing instructions to processor 504 for execution. Such a medium may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Non-volatile media includes, for example, optical or magnetic disks, such as disk drive 510. Volatile media includes dynamic memory, such as system memory 506. Transmission media includes coaxial cables, copper wire, and fiber optics, including wires that comprise bus 502. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.
Common forms of computer readable media includes, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, carrier wave, or any other medium from which a computer can read.
In an embodiment of the invention, execution of the sequences of instructions to practice the invention is performed by a single computer system 500. According to other embodiments of the invention, two or more computer systems 500 coupled by communication link 520 (e.g., LAN, PTSN, or wireless network) may perform the sequence of instructions required to practice the invention in coordination with one another.
Computer system 500 may transmit and receive messages, data, and instructions, including a program, i.e., application code, through communication link 520 and communication interface 512. Received program code may be executed by processor 504 as it is received, and/or stored in disk drive 510, or other non-volatile storage for later execution.
In the foregoing specification, the invention has been described with reference to specific embodiments thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention. For example, the above-described process flows are described with reference to a particular ordering of process actions. However, the ordering of many of the described process actions may be changed without affecting the scope or operation of the invention. The specification and drawings are, accordingly, to be regarded in an illustrative rather than restrictive sense.
The present application claims the benefit of U.S. Provisional Application Ser. No. 60/500,467, filed Sep. 5, 2003, which is hereby incorporated by reference in its entirety for all purposes as if fully set forth herein.
Number | Date | Country | |
---|---|---|---|
60500467 | Sep 2003 | US |