WO2017113888A1 - 数据写入装置及方法 - Google Patents

数据写入装置及方法 Download PDF

Info

Publication number
WO2017113888A1
WO2017113888A1 PCT/CN2016/098860 CN2016098860W WO2017113888A1 WO 2017113888 A1 WO2017113888 A1 WO 2017113888A1 CN 2016098860 W CN2016098860 W CN 2016098860W WO 2017113888 A1 WO2017113888 A1 WO 2017113888A1
Authority
WO
WIPO (PCT)
Prior art keywords
hard disk
partition
data
data object
key value
Prior art date
Application number
PCT/CN2016/098860
Other languages
English (en)
French (fr)
Inventor
薛贞文
夏勇
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2017113888A1 publication Critical patent/WO2017113888A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
    • G06F3/0608—Saving storage space on storage systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
    • G06F3/0638—Organizing or formatting or addressing of data
    • G06F3/0644—Management of space entities, e.g. partitions, extents, pools
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
    • G06F3/0601—Interfaces specially adapted for storage systems
    • G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
    • G06F3/0671—In-line storage system
    • G06F3/0673—Single storage device

Definitions

  • the present invention relates to the field of data storage technologies, and in particular, to a data writing apparatus and method.
  • SMR Silicon Magnetic Recording
  • the current SMR hard disks are partitioned according to the magnetic track, and the end of each hard disk partition does not adopt a shingle layout.
  • KV (Key-Value, key-value) storage is a commonly used non-relational storage method.
  • a key-value data object When storing data in key-value storage mode, a key-value data object is usually divided into several sub-data objects and then hashed (hash). ) mode all the child data objects are scattered to the hard disk for storage.
  • hashed (hash). ) mode all the child data objects are scattered to the hard disk for storage.
  • a key-value data object is stored in an SMR hard disk, several sub-data objects of a key-value data object will be scattered among the individual hard disk partitions or in a non-contiguous physical area on the same hard disk partition.
  • the entire key value data object is usually operated, because the sub data object of the same key value data object in the SMR hard disk may be dispersed in multiple hard disk partitions, or the same hard disk partition is discontinuous.
  • the physical area therefore, the deletion of the key value data object will cause multiple holes in the SMR hard disk, which makes the hard disk fragmentation serious and affects the efficiency of the garbage collection of the storage system.
  • the embodiment of the present invention provides a data writing device and The method is used in a key value storage system, and the key value storage system is built in a shingled magnetic recording SMR hard disk.
  • the technical solution is as follows:
  • a data writing method comprising:
  • the SMR hard disk includes a plurality of hard disk partitions; determining a first target partition among the plurality of hard disk partitions; The key value data object is sliced into at least one child data object, and the at least one child data object is sequentially written to adjacent tracks in the first target partition.
  • the data writing method provided by the embodiment of the present invention determines a first target partition in a plurality of hard disk partitions of the SMR hard disk by receiving a write instruction for instructing the key value data object to be written into the SMR hard disk, and determining The first target partition sequentially writes at least one sub-data object included in the key value data object, and when the part or all of the sub-data objects in the key value data object are subsequently deleted, the invalid data formed is relatively concentrated, It will cause too many holes in the SMR hard disk, which will alleviate the fragmentation of the SMR hard disk and improve the garbage collection efficiency of the system.
  • the determining the first target partition in the multiple hard disk partitions includes: determining whether the SMR hard disk has the same as the key value data object The data of the object identifier; if the result of the determination is that the data of the object identifier having the same object identifier as the key value data object already exists in the SMR hard disk, the data having the same object identifier as the key value data object is located. And a hard disk partition whose remaining space is not less than the data amount of the key value data object is determined as the first target partition.
  • the key value data object can be preferentially written into the hard disk partition where the other sub-data objects corresponding to the same object identifier are located, so that the data corresponding to the same object identifier is more concentrated, and further It alleviates the fragmentation of the SMR hard disk when deleting data.
  • the at least one sub data object included in the key value data object respectively corresponds to an object identifier of the key value data object
  • the method further includes: in the SMR When the invalid data in the hard disk is garbage collected, it is detected whether each valid sub-data object corresponding to the object identifier is stored in the same hard disk partition; if the detection result is that the valid sub-data objects are not stored in the same hard disk partition Determining, in the plurality of hard disk partitions, a second target partition; moving the respective valid sub-data objects to the second target partition.
  • garbage collection of invalid data in the SMR hard disk it is detected whether each valid sub data object corresponding to the object identifier is stored in the same hard disk partition, and if the detection result is that the valid sub data objects are not stored in the same a hard disk partition, the second target partition is determined in the plurality of hard disk partitions, and the valid valid data objects are moved to adjacent magnetic tracks in the second target partition, and garbage collection is performed uniformly, and valid data can be obtained.
  • the data with the same object-id continues to be stored in a centralized state, further alleviating the fragmentation of the SMR hard disk and improving the garbage collection efficiency.
  • the determining the second target partition in the multiple hard disk partitions includes: determining to store the respective Whether there is a hard disk partition that meets a preset condition in the hard disk partition of the valid sub data object, the preset condition is: the remaining space of the hard disk partition is not less than other valid sub data stored outside the hard disk partition. And the hard disk partition that satisfies the preset condition is determined as the second target partition.
  • each valid sub-data object corresponding to the same object identifier is preferentially moved to a hard disk partition in which a valid sub-data object corresponding to the object identifier is already stored, so that only other valid sub-data objects need to be moved. It can be in the second target partition, which reduces the amount of data movement.
  • an embodiment of the present invention provides a storage device, including: a processor and a memory; the processor is configured to execute an instruction stored in the memory; and the processor implements the first aspect or the A method of writing data provided by any of the possible implementations in one aspect.
  • an embodiment of the present invention provides a data writing apparatus, where the data writing apparatus includes at least one unit, and the at least one unit is configured to implement any one of the foregoing first aspect or the first aspect.
  • the data writing method provided.
  • the determining the first target partition in the multiple hard disk partitions further includes: if the result of the determination is that the SMR hard disk does not have the same object identifier data as the key value data object, A hard disk partition in which a remaining space of the plurality of hard disk partitions is not less than a data amount of the key value data object is determined as the first target partition.
  • the determining the second target partition in the multiple hard disk partitions further includes: if the result of the determination is that there is no hard disk partition that meets the preset condition, leaving one of the plurality of hard disk partitions A hard disk partition having a space not less than a sum of data amounts of the respective valid sub data objects is determined as the second target partition.
  • FIG. 1 is a schematic structural diagram of a storage device according to an exemplary embodiment of the present invention.
  • FIG. 2A is a flowchart of a method of data writing method according to an exemplary embodiment of the present invention.
  • FIG. 2B is a flowchart of a first target partition determining method involved in the embodiment shown in FIG. 2A;
  • FIG. 2C is a flowchart of a second target partition determining method involved in the embodiment shown in FIG. 2A;
  • FIG. 3 is a block diagram of a data writing apparatus provided by an exemplary embodiment of the present invention.
  • the storage device 100 includes a storage controller 110 and at least one hard disk 120.
  • the storage device 100 runs a key value storage system, and at least one hard disk 120 is an SMR hard disk.
  • the number of the storage controllers 110 of the storage device 100 is one or more.
  • the primary storage controller fails, the primary storage controller is replaced by the secondary storage controller.
  • the storage controller 110 includes a processor 112 and a network interface 114.
  • Processor 112 includes one or more processing cores.
  • the processor 112 executes various functional applications and data processing by running software programs and modules.
  • network interfaces 114 There may be multiple network interfaces 114, some of which are used for storage device 100 to communicate with other hosts or devices, and another portion of network interface 114 for communicating with at least one hard disk 120.
  • the storage controller 110 further includes components such as a memory 116 and a bus 118.
  • the memory 116 is coupled to the processor 112 via a bus 118.
  • Memory 116 can be used to store software programs as well as modules.
  • the memory 116 can store an operating system 1162 and an application module 1164 required for at least one function.
  • the operating system 1162 can be an operating system such as Real Time eXecutive (RTX), LINUX, UNIX, WINDOWS, or OS X.
  • the application module 1164 may be an instruction receiving module 1164a, a partition determining module 1164b, a writing module 1164c, and the like.
  • the instruction receiving module 1164a is configured to receive a write instruction, where the write command is used to instruct to write a key value data object to the SMR hard disk, where the SMR hard disk includes a plurality of hard disk partitions;
  • the partition determining module 1164b is configured to determine a first target partition among the plurality of hard disk partitions
  • the write module 1164c is configured to slice the key value data object into at least one sub data object and sequentially write the at least one sub data object to an adjacent magnetic track in the first target partition.
  • the at least one sub-data object of the above-mentioned key value data object respectively corresponds to the object identifier of the key value data object
  • the application module 1164 may further include: a detecting module 1164d and a moving module 1164e;
  • the detecting module 1164d is configured to detect, when the invalid data in the SMR hard disk is garbage collected, whether each valid sub data object corresponding to the object identifier is stored in the same hard disk partition.
  • the partition determining unit 1164b is further configured to determine the second target partition in the plurality of hard disk partitions if the detection result of the detecting unit 1164d is that the valid valid data objects are not stored in the same hard disk partition.
  • the shifting unit 1164e is configured to move the respective valid sub-data objects to adjacent tracks in the second target partition.
  • memory 116 can be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read only memory
  • EPROM Programmable Read Only Memory
  • PROM Programmable Read Only Memory
  • ROM Read Only Memory
  • Magnetic Memory Magnetic Memory
  • Flash Memory Disk or Optical Disk.
  • the structure of the storage device 100 shown in FIG. 1 does not constitute a limitation on the storage device 100, may include more or less components than those illustrated, or combine some components, or different. Assembly of parts.
  • FIG. 2A is a flowchart of a method of data writing method according to an exemplary embodiment of the present invention, which may be used in a key value storage system, and is executed by the storage device shown in FIG. 1.
  • the data writing method may include:
  • Step 201 Receive a write command for instructing to write a key value data object to the SMR hard disk, where the SMR hard disk includes a plurality of hard disk partitions.
  • the storage device may first buffer a series of key value data objects in the cache, when certain conditions are met, such as the time from the last write reaches a preset time threshold, or When the data amount of the stored data reaches a certain data amount threshold, the storage device generates a write instruction, or the write instruction may also be sent by the user through a specified operation; when the storage device receives the write instruction, the cached device The key-value data object is written to the hard disk in a sequential hard disk write operation, thereby improving the performance of the write.
  • This step 201 can be implemented by the processor of the memory controller in the memory device executing the instruction receiving module 1164a in the memory.
  • Step 202 Determine a first target partition in the plurality of hard disk partitions.
  • the central idea of determining the first target partition is to write the data of the key value data object to the same hard disk partition as much as possible.
  • FIG. 2B shows a flowchart of the first target partition determining method according to the exemplary embodiment.
  • the first target partition determining method may include the following steps 202a to 202.
  • Step 202a it is determined whether data of the same object identifier as the key value data object already exists in the SMR hard disk, and if yes, the process proceeds to step 202b, otherwise, the process proceeds to step 202c.
  • each key value data object corresponds to an object identifier (object-id), and each sub-data object in the key value data object also corresponds to the object identifier, for a key value data object to be stored,
  • the sub-data object corresponding to the identifier of the corresponding object may also be stored in the hard disk.
  • the sub-data object corresponding to the object identifier existing in the hard disk may be Data objects make up a key-value data object.
  • the key value data object when the key value data object is written, the key value data object may be preferentially written into the hard disk partition where the other sub-data objects corresponding to the same object identifier are located, that is, the storage device first determines Whether there is already other data with the same object identifier as the key value data object.
  • Step 202b Determine a hard disk partition in which the remaining space of the data having the same object identifier as the key value data object is not less than the data amount of the key value data object is the first target partition.
  • the exemplary embodiment if other data having the same object identifier as the key value data object already exists in the SMR hard disk, it may be checked that there is another object having the same object identifier as the key value data object.
  • the hard disk partition where the data is located, and the hard disk partition in which the remaining space is not less than the data amount of the key value data object determines the first target partition.
  • a remaining space may be determined from the other hard disk partitions.
  • Step 202c Determine a hard disk partition in which a remaining space of the plurality of hard disk partitions is not less than a data amount of the key value data object as the first target partition.
  • a hard disk partition having a remaining space not less than the data amount of the key value data object may be directly determined as the first target partition.
  • the storage device may not care whether the data of the same object identifier as the key value data object already exists in the SMR hard disk, and only determine a hard disk partition with sufficient remaining space from the plurality of hard disk partitions. It is sufficient as the first target partition.
  • This step 202 can be implemented by a processor in the storage device executing a partition determination module 1164b in the memory of the processor.
  • Step 203 Dividing the key value data object into at least one sub data object, and sequentially writing the at least one sub data object to an adjacent magnetic track in the first target partition.
  • the storage device can write each sub data object included in the key value data object one by one in the first target partition, and ensure that each sub data object of a key value data object is concentrated as much as possible.
  • the storage is on a continuous physical area in the SMR hard disk. If the magnetic track corresponding to the remaining space in the first target partition is physically discontinuous, data movement may be performed inside the first target partition, so that the magnetic track corresponding to the remaining space of the first target partition is physically continuous. Further, if other data having the same object identifier as the key data object is already stored in the first target partition, after the data is moved inside the first target partition, the magnetic space corresponding to the remaining space of the first target partition is made. The tracks are physically contiguous and immediately after the track on which the other data having the same object identification as the keyed data object is located.
  • the key value data object may be sequentially written from the hard disk partition with the largest remaining space. At least one child data object is included until the key value data object is all written to the hard disk.
  • each sub-data object in the key value data object is sequentially written into the same hard disk partition, and when some or all of the sub-data objects in the key value data object are subsequently deleted, the invalid data formed is relatively concentrated, and will not be in the Excessive holes in the SMR hard disk alleviate the fragmentation of the SMR hard disk, thereby improving the garbage collection efficiency of the system.
  • This step 203 can be implemented by the processor of the memory controller in the memory device executing the write module 1164c in memory.
  • Step 204 When garbage collection of the invalid data in the SMR hard disk, detecting whether each valid sub data object corresponding to the object identifier of the key value data object is stored in the same hard disk partition.
  • the at least one sub-data object included in the key value data object respectively corresponds to an object identifier of the key value data object.
  • the data in the SMR hard disk cannot be rewritten directly on the basis of the original data, and can only be rewritten after the invalidation, so when all the sub-data objects corresponding to a key value data object or When the part is deleted, the storage device first sets the deleted sub-data object in the SMR hard disk as invalid data. At a certain point in time, the storage device performs garbage collection uniformly, and moves the valid data together to make the same object. The -id data continues to be in a centralized storage state, further alleviating the fragmentation of the SMR hard disk and improving garbage collection efficiency.
  • the storage device When performing garbage collection, the storage device first queries whether valid sub-data objects corresponding to the respective object identifiers are stored in the same hard disk partition. For an object identifier, if each valid sub-data object corresponding to the object identifier is not stored in the In the same hard disk partition, subsequent processing of each valid sub data object is performed.
  • This step 204 can be implemented by a processor in the storage device executing a detection module 1164d in the memory of the processor.
  • Step 205 If the detection result is that the valid sub-data objects are not stored in the same hard disk partition, determine the second target partition in the plurality of hard disk partitions.
  • FIG. 2C shows a flowchart of a second target partition determining method according to the exemplary embodiment.
  • the second target partition determining method may include the following steps 205a to 205c.
  • Step 205a determining to store a hard disk corresponding to each valid sub-data object of the same object identifier In the partition, is there a hard disk partition that meets the preset conditions? If yes, go to step 205b, otherwise, go to step 205c.
  • the preset condition is that the remaining space of the hard disk partition is not smaller than the sum of the data amounts of other valid sub data objects stored outside the hard disk partition.
  • Step 205b Determine the hard disk partition that meets the preset condition as the second target partition.
  • each valid sub-data object corresponding to the same object identifier may be preferentially moved to a hard disk partition in which a valid sub-data object corresponding to the object identifier is already stored, and thus,
  • the storage device first queries, from the hard disk partition storing the valid sub-data objects corresponding to the object identifier, a hard disk partition having a remaining capacity space sufficient to move other valid sub-data objects into the second. The target partition, so that only other valid sub-data objects need to be moved to the second target partition.
  • Step 205c Determine a hard disk partition in which a remaining space of the plurality of hard disk partitions is not less than a sum of data amounts of the valid data objects of the respective ones as the second target partition.
  • the storage device cannot find out the hard disk partition in which the remaining capacity is sufficient to move other valid sub-data objects from the hard disk partition storing the valid sub-data objects corresponding to the object identifier, it can be determined from other hard disk partitions. A remaining space is sufficient for moving the corresponding valid sub-data object corresponding to the object identifier into the hard disk partition as the second target partition.
  • This step 205 can be implemented by a processor executing a partition determination unit 1164b in the memory of the memory controller of the storage device.
  • Step 206 Move each valid sub-data object to an adjacent magnetic track in the second target partition.
  • This step 206 can be implemented by the processor of the storage controller in the storage device executing the transport unit 1164e in the memory.
  • the data writing method shown in the exemplary embodiment determines the first among the plurality of hard disk partitions of the SMR hard disk by receiving a write command for instructing the key value data object to be written to the SMR hard disk.
  • Target partitioning and sequentially writing at least one sub-data object included in the key value data object in the determined first target partition, and forming invalid when subsequently deleting part or all of the sub-data objects in the key value data object.
  • the data will be relatively concentrated, and will not cause too many holes in the SMR hard disk, which will alleviate the fragmentation of the SMR hard disk, thereby improving the garbage collection efficiency of the system.
  • the data writing method shown in the exemplary embodiment detects whether each valid sub-data object corresponding to the object identifier is stored in the same hard disk partition by performing garbage collection on the invalid data in the SMR hard disk. If the detection result is that the valid sub-data objects are not stored in the same hard disk partition, determining a second target partition in the plurality of hard disk partitions, and moving each valid sub-data object to the second target partition, Unified garbage collection allows you to move valid data together and keep the same object-id data in a centralized storage state, further reducing the fragmentation of SMR hard disks and improving garbage collection efficiency.
  • FIG. 3 shows a block diagram of a data writing device provided by an exemplary embodiment of the present invention.
  • the data writing device can be implemented as all or part of the storage device shown in FIG. 1 by software, hardware or a combination of both.
  • the data writing device may include an instruction receiving unit 301, a partition determining unit 302, a writing unit 303, a detecting unit 304, and a shifting unit 305.
  • the instruction receiving unit 301 has the same or similar function as the instruction receiving module 1164a.
  • the partition determining unit 302 has the same or similar function as the partition determining module 1164b.
  • the writing unit 303 has the same or similar function as the writing module 1164c.
  • the detecting unit 304 has the same or similar function as the detecting module 1164d.
  • the transport unit 305 has the same or similar function as the transport module 1164e.
  • a person skilled in the art may understand that all or part of the steps of implementing the above embodiments may be completed by hardware, or may be instructed by a program to execute related hardware, and the program may be stored in a computer readable storage medium.
  • the storage medium mentioned may be a read only memory, a magnetic disk or an optical disk or the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Signal Processing For Digital Recording And Reproducing (AREA)

Abstract

一种数据写入方法，属于数据存储技术领域。所述方法用于键值存储系统，所述键值存储系统建立于叠瓦式磁记录SMR硬盘中，所述方法包括：接收写入指令，所述写入指令用于指示将键值数据对象写入所述SMR硬盘，所述SMR硬盘包含多个硬盘分区（201）；在所述多个硬盘分区中确定第一目标分区（202）；将所述键值数据对象切分成至少一个子数据对象，并将所述至少一个子数据对象顺序写入所述第一目标分区中的相邻磁轨（203）。在后续删除该键值数据对象中的部分或者全部子数据对象时，形成的无效数据也会相对集中，不会在SMR硬盘上造成过多的空洞，缓解了SMR硬盘的碎片化程度，从而提高系统的垃圾回收效率。

Description

数据写入装置及方法 技术领域
本发明涉及数据存储技术领域，特别涉及一种数据写入装置及方法。
背景技术
SMR(Shingled Magnetic Recording，叠瓦式磁记录)技术是一种新兴的硬盘存储技术。目前的SMR硬盘都是按照磁轨进行分区，每一个硬盘分区的末端不采用叠瓦式布局。
KV(Key-Value，键值)存储是一种常用的非关系型存储方式，采用键值存储方式存储数据时，通常将一个键值数据对象切分成若干子数据对象，然后通过哈希(hash)方式将所有子数据对象打散到硬盘中进行存储。当键值数据对象存储入SMR硬盘时，一个键值数据对象的若干个子数据对象将会分散在各个硬盘分区，或者同一个硬盘分区上不连续的物理区域中。
在实现本发明的过程中，发明人发现现有技术至少存在以下问题：
在键值存储中，业务删除时通常针对整个键值数据对象进行操作，由于SMR硬盘中同一个键值数据对象的子数据对象可能分散在多个硬盘分区，或者同一个硬盘分区上不连续的物理区域，因此，删除键值数据对象会在SMR硬盘上造成多个空洞，使得硬盘碎片化严重，影响存储系统垃圾回收的效率。
发明内容
为了解决现有技术中删除键值数据对象会在SMR硬盘上造成多个空洞，使得硬盘碎片化严重，影响存储系统垃圾回收的效率的问题，本发明实施例提供了一种数据写入装置及方法，用于键值存储系统中，所述键值存储系统建立于叠瓦式磁记录SMR硬盘中，所述技术方案如下：
第一方面，提供了一种数据写入方法，该方法包括：
接收写入指令，所述写入指令用于指示将键值数据对象写入所述SMR硬盘，所述SMR硬盘包含多个硬盘分区；在所述多个硬盘分区中确定第一目标分区；将所述键值数据对象切分成至少一个子数据对象，并将所述至少一个子数据对象顺序写入所述第一目标分区中的相邻磁轨。
本发明实施例提供的数据写入方法，通过接收用于指示将键值数据对象写入该SMR硬盘的写入指令，在SMR硬盘的多个硬盘分区中确定第一目标分区，并在确定的该第一目标分区中顺序写入该键值数据对象包含的至少一个子数据对象，在后续删除该键值数据对象中的部分或者全部子数据对象时，形成的无效数据也会相对集中，不会在SMR硬盘上造成过多的空洞，缓解了SMR硬盘的碎片化程度，从而提高系统的垃圾回收效率。
在第一方面的第一种可能的实现方式中，所述在所述多个硬盘分区中确定第一目标分区，包括：判断所述SMR硬盘中是否已存在与所述键值数据对象具有相同的对象标识的数据；若判断结果为所述SMR硬盘中已存在与所述键值数据对象具有相同的对象标识的数据，则将与所述键值数据对象具有相同的对象标识的数据所在的、剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。对键值数据对象进行写入时，可以优先将该键值数据对象写入到其它对应相同的对象标识的子数据对象所在的硬盘分区中，从而使对应相同的对象标识的数据更加集中，进一步缓解了删除数据时，SMR硬盘的碎片化程度。
在第一方面的第二种可能的实现方式中，所述键值数据对象包含的至少一个子数据对象各自对应所述键值数据对象的对象标识，所述方法还包括：在对所述SMR硬盘中的无效数据进行垃圾回收时，检测对应所述对象标识的各个有效的子数据对象是否存储于同一个硬盘分区；若检测结果为所述各个有效的子数据对象未存储于同一个硬盘分区，则在所述多个硬盘分区中确定第二目标分区；将所述各个有效的子数据对象搬移到所述第二目标分区。通过在对该SMR硬盘中的无效数据进行垃圾回收时，检测对应该对象标识的各个有效的子数据对象是否存储于同一个硬盘分区，若检测结果为该各个有效的子数据对象未存储于同一个硬盘分区，则在该多个硬盘分区中确定第二目标分区，将该各个有效的子数据对象搬移到该第二目标分区中的相邻磁轨，在统一进行垃圾回收，可以将有效数据搬移到一起，使具有相同的object-id的数据继续保持集中存储的状态，进一步缓解SMR硬盘的碎片化程度，提高垃圾回收效率。
结合第一方面的第二种可能实现方式，在第一方面的第三种可能实现方式中，所述在所述多个硬盘分区中确定第二目标分区，包括：判断存储所述各个 有效的子数据对象的硬盘分区中，是否存在满足预设条件的硬盘分区，所述预设条件为：所述硬盘分区的剩余空间不小于存储于所述硬盘分区之外的其它有效的子数据对象的数据量之和；若判断结果为存在满足所述预设条件的硬盘分区，则将所述满足所述预设条件的硬盘分区确定为所述第二目标分区。在进行垃圾回收时，将对应同一对象标识的各个有效的子数据对象优先搬移到已经存储有部分对应该对象标识的有效子数据对象的硬盘分区中，这样只需要将其它有效的子数据对象搬移到第二目标分区中即可，减少了数据搬移量。
第二方面，本发明实施例提供了一种存储设备，该存储设备包括：处理器和存储器；处理器被配置为执行存储器中存储的指令；处理器通过执行指令来实现上述第一方面或第一方面中任意一种可能的实现方式所提供的数据写入方法。
第三方面，本发明实施例提供了一种数据写入装置，该数据写入装置包括至少一个单元，该至少一个单元用于实现上述第一方面或第一方面中任意一种可能的实现方式所提供的数据写入方法。
可选的，所述在所述多个硬盘分区中确定第一目标分区，还包括：若判断结果为所述SMR硬盘中不存在与所述键值数据对象具有相同的对象标识的数据，则将所述多个硬盘分区中一个剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。
可选的，所述在所述多个硬盘分区中确定第二目标分区，还包括：若判断结果为不存在满足所述预设条件的硬盘分区，则将所述多个硬盘分区中一个剩余空间不小于所述各个有效的子数据对象的数据量之和的硬盘分区确定为所述第二目标分区。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明示例性实施例所示出的存储设备的结构示意图；
图2A是本发明一示例性实施例示出的数据写入方法的方法流程图；
图2B是图2A所示实施例涉及的第一目标分区确定方法的流程图；
图2C是图2A所示实施例涉及的第二目标分区确定方法的流程图；
图3是本发明一个示例性实施例提供的数据写入装置的框图。
具体实施方式
为使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明实施方式作进一步地详细描述。
请参考图1其示出了本发明一个示例性实施例提供的存储设备100的结构示意图。该存储设备100包括：存储控制器110以及至少一个硬盘120。其中，该存储设备100中运行有键值存储系统，至少一个硬盘120为SMR硬盘。
可选的，存储设备100的存储控制器110的数量为一个或多个。比如，存储控制器110为两个，其中一个为主存储控制器，另一个为备用存储控制器。当主存储控制器发生故障时，由备用存储控制器代替主存储控制器进行运行。
存储控制器110包括：处理器112和网络接口114。
处理器112包括一个或者一个以上处理核心。处理器112通过运行软件程序以及模块，从而执行各种功能应用以及数据处理。
网络接口114可以为多个，其中一部分网络接口用于存储设备100与其它主机或者设备进行通信，另一部分网络接口114用于与至少一个硬盘120通信。
可选的，存储控制器110还包括存储器116以及总线118等部件。
其中：存储器116通过总线118与处理器112相连。存储器116可用于存储软件程序以及模块。
存储器116可存储操作系统1162以及至少一个功能所需的应用程序模块1164。操作系统1162可以是实时操作系统(Real Time eXecutive，RTX)、LINUX、UNIX、WINDOWS或OS X之类的操作系统。应用程序模块1164可以是指令接收模块1164a、分区确定模块1164b以及写入模块1164c等。
指令接收模块1164a用于接收写入指令，所述写入指令用于指示将键值数据对象写入所述SMR硬盘，所述SMR硬盘包含多个硬盘分区；
分区确定模块1164b用于在所述多个硬盘分区中确定第一目标分区；
写入模块1164c用于将所述键值数据对象切分成至少一个子数据对象，并将所述至少一个子数据对象顺序写入所述第一目标分区中的相邻磁轨。
可选的，上述键值数据对象包含的至少一个子数据对象各自对应该键值数据对象的对象标识，应用程序模块1164还可以包括：检测模块1164d以及搬移模块1164e；
检测模块1164d，用于在对该SMR硬盘中的无效数据进行垃圾回收时，检测对应该对象标识的各个有效的子数据对象是否存储于同一个硬盘分区。
分区确定单元1164b，还用于若该检测单元1164d的检测结果为该各个有效的子数据对象未存储于同一个硬盘分区，则在该多个硬盘分区中确定第二目标分区。
搬移单元1164e，用于将该各个有效的子数据对象搬移到该第二目标分区中的相邻磁轨。
此外，存储器116可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。相应地。
本领域技术人员可以理解，图1中所示出的存储设备100结构并不构成对存储设备100的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
图2A是本发明一示例性实施例示出的数据写入方法的方法流程图，该方法可以用于键值存储系统中，由图1所示的存储设备执行。如图2A所示，该数据写入方法可以包括：
步骤201，接收写入指令，该写入指令用于指示将键值数据对象写入该SMR硬盘，该SMR硬盘包含多个硬盘分区。
在本示例性实施例中，存储设备可以先在缓存中缓冲一系列的键值数据对象，当满足一定条件，比如距离上次写入的时间达到预设时间阈值，或者，缓 存数据的数据量达到一定的数据量阈值时，存储设备生成写入指令，或者，该写入指令也可以由用户通过指定的操作发送；当存储设备接收到该写入指令后，将缓存的键值数据对象在一次顺序的硬盘写操作中全部写入硬盘，从而提升写入的性能。
该步骤201可由存储设备中存储控制器的处理器执行存储器中的指令接收模块1164a来实现。
步骤202，在该多个硬盘分区中确定第一目标分区。
在本示例性实施例中，确定第一目标分区的中心思想是，将该键值数据对象的数据都尽可能的按顺序写入同一个硬盘分区。
具体的，请参考图2B，其示出了本示例性实施例涉及的第一目标分区确定方法的流程图，如图2B所示，该第一目标分区确定方法可以包括下列步骤202a至202。
步骤202a，判断该SMR硬盘中是否已存在与该键值数据对象具有相同的对象标识的数据，若是，进入步骤202b，否则，进入步骤202c。
在键值存储中，每个键值数据对象对应一个对象标识(object-id)，该键值数据对象中每一个子数据对象也都对应该对象标识，对于一个待存储的键值数据对象，在硬盘中可能还存储在其它对应该对象标识的子数据对象，在该待存储的键值数据对象中的各个子数据对象写入硬盘后，可以与硬盘中已有的对应该对象标识的子数据对象组成一个键值数据对象。
在本示例性实施例中，对键值数据对象进行写入时，可以优先将该键值数据对象写入到其它对应相同的对象标识的子数据对象所在的硬盘分区中，即存储设备首先判断是否已存在与该键值数据对象具有相同的对象标识的其它数据。
步骤202b，将与该键值数据对象具有相同的对象标识的数据所在的、剩余空间不小于该键值数据对象的数据量的硬盘分区确定为该第一目标分区。
可选的，在本示例性实施例中，如果SMR硬盘中已存在与该键值数据对象具有相同的对象标识的其它数据，则可以检查存在与该键值数据对象具有相同的对象标识的其它数据所在的硬盘分区，并将其中剩余空间不小于该键值数据对象的数据量的硬盘分区确定该第一目标分区。
如果该SMR硬盘中，与该键值数据对象具有相同的对象标识的其它数据所在的硬盘分区的剩余空间都小于该键值数据对象的数据量，则可以从其它硬盘分区中确定一个剩余空间不小于该键值数据对象的数据量的硬盘分区。
步骤202c，将该多个硬盘分区中一个剩余空间不小于该键值数据对象的数据量的硬盘分区确定为该第一目标分区。
如果SMR硬盘中不存在与该键值数据对象具有相同的对象标识的其它数据，则可以直接将一个剩余空间不小于该键值数据对象的数据量的硬盘分区确定为该第一目标分区。
或者，在实际应用中，存储设备也可以不关心该SMR硬盘中是否已存在与该键值数据对象具有相同的对象标识的数据，只从多个硬盘分区中确定出一个剩余空间足够的硬盘分区作为该第一目标分区即可。
该步骤202可由存储设备中存储控制器的处理器执行存储器中的分区确定模块1164b来实现。
步骤203，将该键值数据对象切分成至少一个子数据对象，并将该至少一个子数据对象顺序写入第一目标分区中的相邻磁轨。
在写入键值数据对象时，存储设备可以在第一目标分区中逐个磁轨写入该键值数据对象包含的各个子数据对象，保证一个键值数据对象的各个子数据对象都尽可能集中的存储在SMR硬盘中连续的物理区域上。若该第一目标分区中的剩余空间对应的磁轨在物理上不连续，则可以在第一目标分区内部进行数据搬移，使得第一目标分区的剩余空间对应的磁轨在物理上连续。进一步的，若第一目标分区内已经存储有与该键值数据对象具有相同的对象标识的其它数据，则在第一目标分区内部进行数据搬移后，使得第一目标分区的剩余空间对应的磁轨在物理上连续，且紧接在与该键值数据对象具有相同的对象标识的其它数据所在的磁轨之后。
可选的，在极端情况下，如果该SMR硬盘中各个硬盘分区的剩余空间都小于该键值数据对象的数据量，则可以从剩余空间最大的硬盘分区，依次顺序写入该键值数据对象包含的至少一个子数据对象，直至该键值数据对象全部写入硬盘。
本示例性实施例上述数据写入方法，在SMR硬盘中存储键值数据对象时， 将键值数据对象中的各个子数据对象顺序写入相同的硬盘分区中，在后续删除该键值数据对象中的部分或者全部子数据对象时，形成的无效数据也会相对集中，不会在SMR硬盘上造成过多的空洞，缓解了SMR硬盘的碎片化程度，从而提高系统的垃圾回收效率。
该步骤203可由存储设备中存储控制器的处理器执行存储器中的写入模块1164c来实现。
步骤204，在对该SMR硬盘中的无效数据进行垃圾回收时，检测对应该键值数据对象的对象标识的各个有效的子数据对象是否存储于同一个硬盘分区。
其中，上述键值数据对象包含的至少一个子数据对象各自对应该键值数据对象的对象标识。
基于SMR硬盘的写入特点，SMR硬盘中的数据无法直接在原数据的基础上进行改写，只能无效之后重新进行顺序写，因此，当一个键值数据对象对应的各个子数据对象中的全部或者部分被删除时，存储设备首先将SMR硬盘中被删除的子数据对象设置为无效数据，后续在某一个时间点，存储设备会统一进行垃圾回收，将有效数据搬移到一起，使具有相同的object-id的数据继续保持集中存储的状态，进一步缓解SMR硬盘的碎片化程度，提高垃圾回收效率。
在进行垃圾回收时，存储设备首先查询各个对象标识对应的有效的子数据对象是否存储在同一个硬盘分区中，对于某一个对象标识，如果该对象标识对应的各个有效的子数据对象未存储在同一个硬盘分区中，则对该各个有效的子数据对象进行后续处理。
该步骤204可由存储设备中存储控制器的处理器执行存储器中的检测模块1164d来实现。
步骤205，若检测结果为该各个有效的子数据对象未存储于同一个硬盘分区，则在该多个硬盘分区中确定第二目标分区。
具体的，请参考图2C，其示出了本示例性实施例涉及的第二目标分区确定方法的流程图，如图2C所示，该第二目标分区确定方法可以包括下列步骤205a至205c。
步骤205a，判断存储对应同一对象标识的各个有效的子数据对象的硬盘 分区中，是否存在满足预设条件的硬盘分区。若是，进入步骤205b，否则，进入步骤205c。
其中，该预设条件为：该硬盘分区的剩余空间不小于存储于该硬盘分区之外的其它有效的子数据对象的数据量之和。
步骤205b，将该满足该预设条件的硬盘分区确定为该第二目标分区。
在本示例性实施例中，为了减少数据搬移量，可以将对应同一对象标识的各个有效的子数据对象优先搬移到已经存储有部分对应该对象标识的有效子数据对象的硬盘分区中，因此，在确定第二目标区域时，存储设备首先从存储对应该对象标识的各个有效的子数据对象的硬盘分区中，查询一个剩余容量空间足够将其它有效的子数据对象搬移进去的硬盘分区作为第二目标分区，这样只需要将其它有效的子数据对象搬移到第二目标分区中即可。
步骤205c，将该多个硬盘分区中一个剩余空间不小于该各个有效的子数据对象的数据量之和的硬盘分区确定为该第二目标分区。
如果存储设备从存储对应该对象标识的各个有效的子数据对象的硬盘分区中，查询不到剩余容量空间足够将其它有效的子数据对象搬移进去的硬盘分区，则可以从其它的硬盘分区中确定一个剩余空间足够将该对象标识对应的各个有效的子数据对象搬移进去的硬盘分区作为第二目标分区。
该步骤205可由存储设备中存储控制器的处理器执行存储器中的分区确定单元1164b来实现。
步骤206，将该各个有效的子数据对象搬移到该第二目标分区中的相邻磁轨。
该步骤206可由存储设备中存储控制器的处理器执行存储器中的搬移单元1164e来实现。
综上所述，本示例性实施例所示的数据写入方法，通过接收用于指示将键值数据对象写入该SMR硬盘的写入指令，在SMR硬盘的多个硬盘分区中确定第一目标分区，并在确定的该第一目标分区中顺序写入该键值数据对象包含的至少一个子数据对象，在后续删除该键值数据对象中的部分或者全部子数据对象时，形成的无效数据也会相对集中，不会在SMR硬盘上造成过多的空洞，缓解了SMR硬盘的碎片化程度，从而提高系统的垃圾回收效率。
此外，本示例性实施例所示的数据写入方法，通过在对该SMR硬盘中的无效数据进行垃圾回收时，检测对应该对象标识的各个有效的子数据对象是否存储于同一个硬盘分区，若检测结果为该各个有效的子数据对象未存储于同一个硬盘分区，则在该多个硬盘分区中确定第二目标分区，将该各个有效的子数据对象搬移到该第二目标分区，在统一进行垃圾回收，可以将有效数据搬移到一起，使具有相同的object-id的数据继续保持集中存储的状态，进一步缓解SMR硬盘的碎片化程度，提高垃圾回收效率。
请参考图3，其示出了本发明一个示例性实施例提供的数据写入装置的框图。该数据写入装置可以通过软件、硬件或者两者的结合实现成为上述图1所示存储设备的全部或者一部分。该数据写入装置可以包括：指令接收单元301、分区确定单元302、写入单元303、检测单元304以及搬移单元305。
指令接收单元301，具有与指令接收模块1164a相同或相似的功能。
分区确定单元302，具有与分区确定模块1164b相同或相似的功能。
写入单元303，具有与写入模块1164c相同或相似的功能。
检测单元304，具有与检测模块1164d相同或相似的功能。
搬移单元305，具有与搬移模块1164e相同或相似的功能。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
以上所述仅为本发明的较佳实施例，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (12)

  1. 一种数据写入装置，其特征在于，用于键值存储系统，所述键值存储系统建立于叠瓦式磁记录SMR硬盘中，所述装置包括：
    指令接收单元，用于接收写入指令，所述写入指令用于指示将键值数据对象写入所述SMR硬盘，所述SMR硬盘包含多个硬盘分区；
    分区确定单元，用于在所述多个硬盘分区中确定第一目标分区；
    写入单元，用于将所述键值数据对象切分成至少一个子数据对象，并将所述至少一个子数据对象顺序写入所述第一目标分区中的相邻磁轨。
  2. 根据权利要求1所述的装置，其特征在于，在所述多个硬盘分区中确定第一目标分区时，所述分区确定单元，用于判断所述SMR硬盘中是否已存在与所述键值数据对象具有相同的对象标识的数据；若判断结果为所述SMR硬盘中已存在与所述键值数据对象具有相同的对象标识的数据，则将与所述键值数据对象具有相同的对象标识的数据所在的、剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。
  3. 根据权利要求2所述的装置，其特征在于，在所述多个硬盘分区中确定第一目标分区时，所述分区确定单元，还用于若判断结果为所述SMR硬盘中不存在与所述键值数据对象具有相同的对象标识的数据，则将所述多个硬盘分区中一个剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。
  4. 根据权利要求1所述的装置，其特征在于，所述键值数据对象包含的至少一个子数据对象各自对应所述键值数据对象的对象标识，所述装置还包括：
    检测单元，用于在对所述SMR硬盘中的无效数据进行垃圾回收时，检测对应所述对象标识的各个有效的子数据对象是否存储于同一个硬盘分区；
    所述分区确定单元，还用于若所述检测单元的检测结果为所述各个有效的子数据对象未存储于同一个硬盘分区，则在所述多个硬盘分区中确定第二目标分区；
    搬移单元，用于将所述各个有效的子数据对象搬移到所述第二目标分区中的相邻磁轨。
  5. 根据权利要求4所述的装置，其特征在于，在所述多个硬盘分区中确定第二目标分区时，所述分区确定单元，用于判断存储所述各个有效的子数据对象的硬盘分区中，是否存在满足预设条件的硬盘分区，所述预设条件为：所述硬盘分区的剩余空间不小于存储于所述硬盘分区之外的其它有效的子数据对象的数据量之和；若判断结果为存在满足所述预设条件的硬盘分区，则将所述满足所述预设条件的硬盘分区确定为所述第二目标分区。
  6. 根据权利要求5所述的装置，其特征在于，在所述多个硬盘分区中确定第二目标分区时，所述分区确定单元，还用于若判断结果为不存在满足所述预设条件的硬盘分区，则将所述多个硬盘分区中一个剩余空间不小于所述各个有效的子数据对象的数据量之和的硬盘分区确定为所述第二目标分区。
  7. 一种数据写入方法，其特征在于，用于键值存储系统，所述键值存储系统建立于叠瓦式磁记录SMR硬盘中，所述方法包括：
    接收写入指令，所述写入指令用于指示将键值数据对象写入所述SMR硬盘，所述SMR硬盘包含多个硬盘分区；
    在所述多个硬盘分区中确定第一目标分区；
    将所述键值数据对象切分成至少一个子数据对象，并将所述至少一个子数据对象顺序写入所述第一目标分区中的相邻磁轨。
  8. 根据权利要求7所述的方法，其特征在于，所述在所述多个硬盘分区中确定第一目标分区，包括：
    判断所述SMR硬盘中是否已存在与所述键值数据对象具有相同的对象标识的数据；
    若判断结果为所述SMR硬盘中已存在与所述键值数据对象具有相同的对象标识的数据，则将与所述键值数据对象具有相同的对象标识的数据所在的、 剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。
  9. 根据权利要求8所述的方法，其特征在于，所述在所述多个硬盘分区中确定第一目标分区，还包括：
    若判断结果为所述SMR硬盘中不存在与所述键值数据对象具有相同的对象标识的数据，则将所述多个硬盘分区中一个剩余空间不小于所述键值数据对象的数据量的硬盘分区确定为所述第一目标分区。
  10. 根据权利要求7所述的方法，其特征在于，所述键值数据对象包含的至少一个子数据对象各自对应所述键值数据对象的对象标识，所述方法还包括：
    在对所述SMR硬盘中的无效数据进行垃圾回收时，检测对应所述对象标识的各个有效的子数据对象是否存储于同一个硬盘分区；
    若检测结果为所述各个有效的子数据对象未存储于同一个硬盘分区，则在所述多个硬盘分区中确定第二目标分区；
    将所述各个有效的子数据对象搬移到所述第二目标分区中的相邻磁轨。
  11. 根据权利要求10所述的方法，其特征在于，所述在所述多个硬盘分区中确定第二目标分区，包括：
    判断存储所述各个有效的子数据对象的硬盘分区中，是否存在满足预设条件的硬盘分区，所述预设条件为：所述硬盘分区的剩余空间不小于存储于所述硬盘分区之外的其它有效的子数据对象的数据量之和；
    若判断结果为存在满足所述预设条件的硬盘分区，则将所述满足所述预设条件的硬盘分区确定为所述第二目标分区。
  12. 根据权利要求11所述的方法，其特征在于，所述在所述多个硬盘分区中确定第二目标分区，还包括：
    若判断结果为不存在满足所述预设条件的硬盘分区，则将所述多个硬盘分区中一个剩余空间不小于所述各个有效的子数据对象的数据量之和的硬盘分区确定为所述第二目标分区。
PCT/CN2016/098860 2015-12-31 2016-09-13 数据写入装置及方法 WO2017113888A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201511031159.0 2015-12-31
CN201511031159.0A CN105677243B (zh) 2015-12-31 2015-12-31 数据写入装置及方法

Publications (1)

Publication Number Publication Date
WO2017113888A1 true WO2017113888A1 (zh) 2017-07-06

Family

ID=56298494

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/098860 WO2017113888A1 (zh) 2015-12-31 2016-09-13 数据写入装置及方法

Country Status (2)

Country Link
CN (1) CN105677243B (zh)
WO (1) WO2017113888A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112954244A (zh) * 2021-01-19 2021-06-11 普联技术有限公司 监控录像的存储实现方法、装置、设备及存储介质
CN113867651A (zh) * 2021-12-03 2021-12-31 武汉深之度科技有限公司 一种硬盘数据清理方法、装置及计算设备
CN114020707A (zh) * 2022-01-06 2022-02-08 阿里云计算有限公司 存储空间回收方法、存储介质及程序产品

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105677243B (zh) * 2015-12-31 2018-12-28 华为技术有限公司 数据写入装置及方法
CN106201353A (zh) * 2016-07-12 2016-12-07 曹蕊 一种用于键值存储系统的数据写入装置及方法
JP2019164854A (ja) * 2018-03-19 2019-09-26 株式会社東芝 磁気ディスク装置およびその記録方法
CN109445681B (zh) * 2018-08-27 2021-05-11 华为技术有限公司 数据的存储方法、装置和存储系统
CN110780812B (zh) * 2019-09-27 2021-08-24 苏州浪潮智能科技有限公司 一种硬盘数据置无效方法与装置
CN111459412B (zh) 2020-03-30 2023-07-04 北京百度网讯科技有限公司 磁盘管理方法、装置以及电子设备
CN111857600B (zh) * 2020-07-30 2024-04-16 江苏金恒信息科技股份有限公司 一种数据读写方法及装置
CN112637327B (zh) * 2020-12-21 2022-07-22 北京奇艺世纪科技有限公司 一种数据处理方法、装置及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103955433A (zh) * 2014-05-09 2014-07-30 华为技术有限公司 盖瓦磁记录硬盘、盖瓦磁记录硬盘写数据的方法及装置
CN104461390A (zh) * 2014-12-05 2015-03-25 华为技术有限公司 将数据写入叠瓦状磁记录smr硬盘的方法及装置
US20150331613A1 (en) * 2013-12-13 2015-11-19 Seagate Technology Llc Media write operation
CN105677243A (zh) * 2015-12-31 2016-06-15 华为技术有限公司 数据写入装置及方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150331613A1 (en) * 2013-12-13 2015-11-19 Seagate Technology Llc Media write operation
CN103955433A (zh) * 2014-05-09 2014-07-30 华为技术有限公司 盖瓦磁记录硬盘、盖瓦磁记录硬盘写数据的方法及装置
CN104461390A (zh) * 2014-12-05 2015-03-25 华为技术有限公司 将数据写入叠瓦状磁记录smr硬盘的方法及装置
CN105677243A (zh) * 2015-12-31 2016-06-15 华为技术有限公司 数据写入装置及方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112954244A (zh) * 2021-01-19 2021-06-11 普联技术有限公司 监控录像的存储实现方法、装置、设备及存储介质
CN113867651A (zh) * 2021-12-03 2021-12-31 武汉深之度科技有限公司 一种硬盘数据清理方法、装置及计算设备
CN113867651B (zh) * 2021-12-03 2022-02-25 武汉深之度科技有限公司 一种硬盘数据清理方法、装置及计算设备
CN114020707A (zh) * 2022-01-06 2022-02-08 阿里云计算有限公司 存储空间回收方法、存储介质及程序产品

Also Published As

Publication number Publication date
CN105677243B (zh) 2018-12-28
CN105677243A (zh) 2016-06-15

Similar Documents

Publication Publication Date Title
WO2017113888A1 (zh) 数据写入装置及方法
US12117911B2 (en) Remote data replication method and system
US11487787B2 (en) System and method for near-synchronous replication for object store
US8001351B2 (en) Data migration method and information processing system
US8924664B2 (en) Logical object deletion
US10572175B2 (en) Method and apparatus of shared storage between multiple cloud environments
US20120246403A1 (en) Write spike performance enhancement in hybrid storage systems
US10379977B2 (en) Data management method, node, and system for database cluster
EP3346387B1 (en) Storage system and system garbage collection method
JP2017079053A (ja) ストレージジャーナリングを改善する方法およびシステム
WO2016107390A1 (zh) 一种管理磁盘文件的方法和装置
EP3120250A1 (en) Solid state drive operations
US10198180B2 (en) Method and apparatus for managing storage device
JP5776474B2 (ja) ストレージ装置、ストレージ制御装置およびコピー先データアクセス方法
WO2018076633A1 (zh) 一种远程数据复制方法、存储设备及存储系统
WO2014157244A1 (ja) ストレージ制御装置、ストレージ制御方法、およびストレージ制御プログラム
US9851907B2 (en) Control device, storage system, and method of controlling a plurality of storage devices
WO2014157243A1 (ja) ストレージ制御装置、ストレージ制御装置の制御方法およびストレージ制御装置の制御プログラム
US10586052B1 (en) Input/output (I/O) inspection methods and systems to detect and defend against cybersecurity threats
US10353641B2 (en) Storage system and method of managing volumes thereof based on received correlation information of the volumes
WO2017054636A1 (zh) 虚拟快照处理方法及装置
KR101686346B1 (ko) 하이브리드 ｓｓｄ 기반 하둡 분산파일 시스템의 콜드 데이터 축출방법
US9489276B2 (en) Implementing enhanced wear leveling in 3D flash memories
WO2016206070A1 (zh) 一种文件更新方法及存储设备
US20170220422A1 (en) Moving data chunks

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16880681

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16880681

Country of ref document: EP

Kind code of ref document: A1