[go: up one dir, main page]

CN102968381A - Method for improving snapshot performance by using solid state disk - Google Patents

Method for improving snapshot performance by using solid state disk Download PDF

Info

Publication number
CN102968381A
CN102968381A CN2012104651996A CN201210465199A CN102968381A CN 102968381 A CN102968381 A CN 102968381A CN 2012104651996 A CN2012104651996 A CN 2012104651996A CN 201210465199 A CN201210465199 A CN 201210465199A CN 102968381 A CN102968381 A CN 102968381A
Authority
CN
China
Prior art keywords
snapshot
chunk
metadata
data
volume
Prior art date
Legal status (The legal status 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 status listed.)
Pending
Application number
CN2012104651996A
Other languages
Chinese (zh)
Inventor
王少锋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
IEIT Systems Co Ltd
Original Assignee
Inspur Electronic Information Industry Co Ltd
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 Inspur Electronic Information Industry Co Ltd filed Critical Inspur Electronic Information Industry Co Ltd
Priority to CN2012104651996A priority Critical patent/CN102968381A/en
Publication of CN102968381A publication Critical patent/CN102968381A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明提供一种利用固态硬盘提高快照性能的方法，该方法是兼顾性能的考虑，利用固态硬盘来提高快照性能的方法。提到的兼顾性能的考虑，是因为如果以普通硬盘存储快照数据，快照数据的操作会极大的降低快照的性能。利用固态硬盘来存储快照数据，是利用固态硬盘的高随机读写性能来提高快照数据操作的速率。本发明的一种将固态硬盘作为快照的例外仓库存储设备的方法，利用固态硬盘高随机读写性能，提高快照对元数据的操作性能，而且掉电元数据不丢失，达到了高性能、高可用性快照的技术效果。

Figure 201210465199

The invention provides a method for improving snapshot performance by using a solid-state hard disk. The method is a method for improving snapshot performance by using a solid-state hard disk in consideration of performance. The consideration of taking into account the performance mentioned is because if the snapshot data is stored on a common hard disk, the operation of the snapshot data will greatly reduce the performance of the snapshot. The use of solid-state drives to store snapshot data is to use the high random read and write performance of solid-state drives to increase the speed of snapshot data operations. A method of using a solid-state hard disk as an exception warehouse storage device for snapshots in the present invention utilizes the high random read and write performance of solid-state hard disks to improve the operational performance of snapshots on metadata, and the metadata will not be lost when power is off, achieving high performance and high efficiency. Technical Effects of Availability Snapshots.

Figure 201210465199

Description

A kind of method of utilizing solid state hard disc to improve the snapshot performance
Technical field
The present invention relates to computer system and field of storage, specifically a kind of method of utilizing solid state hard disc to improve the snapshot performance.
Background technology
Solid state hard disc (Solid State Disk; Be called for short: SSD) formed by control module and storage unit.Storage unit non-volatile flash memory (Flash) chips that adopt perhaps adopt Synchronous Dynamic Random Access Memory (SynchronousDynamic Random Access Memory more; Be called for short: SDRAM).SSD does not have mechanical rotation device, has very high random read-write performance.But cost is very high, through being usually used in high-end field of storage.
Snapshot generally refers to a complete usable copy about the specific data set, and this copy comprises that corresponding data at certain time point (for example: the reflection time point of copy beginning).Snapshot can be copy or the duplicate of represented data.The effect of snapshot is can carry out online data to recover, and when memory device generation application and trouble or file corruption, can utilize snapshot to carry out timely data and recover, and data is reverted to the state of snapshot generation time point.The method that snapshot is realized has a variety of, such as: namely write namely and copy, cut apart mirror snapshot etc.
In linux, realize snapshot based on LVM, employing be copy on firest write (cow) mechanism, one of them important concept makes an exception exactly.Be the snapshot mapping relations, show that data were revised on the snapshot source after setting up snapshot, the data before changing have copied or have copied to another chunk on the snapshot to from some chunk of snapshot source device.In order to embody this mapping relations, need metadata to represent this mapping relations.The equipment that metadata is deposited also is exception storehouse volume.
When to snapped volume or source book read-write, at first need to search the corresponding element data, check whether this chunk shone upon, namely create after the snapshot source volume data and whether revised, the words that were it not for also need be set up new mapping relations again, create new metadata.If the deletion snapshot also needs to delete corresponding metadata etc.Each data block must be carried out metadata operation.If metadata is deposited to hard disk, will be become the performance bottleneck of snapshot to the operation of metadata.If metadata is deposited to internal memory, although can increase the performance of snapshot, in case after the power down, metadata will be lost.
Summary of the invention
The purpose of this invention is to provide a kind of method of utilizing solid state hard disc to improve the snapshot performance.
The objective of the invention is to realize in the following manner, the method is to utilize the high random read-write performance of solid state hard disc, with the exception storehouse volume of solid state hard disc as snapshot, metadata and COW data are stored in the solid state hard disc, the organizational form of metadata in the solid state hard disc, with the metadata store of snapshot at solid state hard disc foremost, the COW data are stored in the residue section, to improve snapshot to the operating performance of metadata, system comprises: snapshot module (1), exception storehouse volume module (2), and concrete steps are as follows:
(1) snapshot module
Adopt the mode of incremental snapshot, by way of exception roll up in the storehouse with solid state hard disc, solid state hard disc is divided into meta-data region and cow data field, after source book is set up snapshot, the meta-data region of initialization exception warehouse volume, the form of metadata is old chunk, new chunk, snapID range, wherein, old chunk points to the piece in the source book, new chunk points to piece corresponding to snapped volume, snapID range is the scope of snapshot ID, and each snapshot has an ID, and snapID range is used for showing which snapshot range the piece of metadata appointment is within;
When source book is write, at first search metadata, were it not for and find words, explanation is first revision snapshot source, need to suppose that the position at the data place in snapshot source is old chunk with the data copy in snapshot source to COW equipment, finishing said process needs following three steps:
1) searches position for save data at COW equipment, be designated as new chunk;
2) data with the snapshot data source are saved in new chunk place on the COW equipment;
3) with exceptional information old chunk, new chunk, snapID range are that so-called meta-data preservation is to solid state hard disc;
If find, illustrate that source book upgraded after creating snapshot, the data of source book have been saved in COW equipment and have suffered, and directly source book are carried out write operation;
When source book is carried out read operation, directly source book is carried out read operation and get final product;
When snapped volume is carried out write operation, at first search metadata, if do not find, the data that this chunk place of source book is described are not also revised, new data need to be write on the COW equipment, search position for save data at COW equipment, be designated as new chunk, again with meta-data preservation to solid state hard disc;
If find, directly the new chunk write operation that finds is got final product;
When snapped volume is carried out read operation, at first search metadata, if do not find, read the corresponding chunk of source book, if find, read the corresponding chunk of snapshot storehouse volume;
(2) exception storehouse volume module
First sector of solid state hard disc is used for the root node position of record metadata B+ tree, the degree of depth and bitmap and the root node of B+ tree as superblock;
In the front portion of solid state hard disc as meta-data index storage, metadata is with the form tissue of B+ tree, the leaf directory entry of B+ tree comprises: in the old chunk source book, in the volume of new chunk snapshot storehouse, the snapshot ID scope that this belongs to mapping, the key word Key of B+ tree is by old chunk, and snapshot ID range forms, maximum catalogue item No. in a node has regulation, to such an extent as to a node just fills a piece;
Internal node has identical data structure with leaf node, if internal node, new_chunk then represents the pointer that points to subordinate B+ tree node, if leaf node, new_chunk is exactly the data content that will search;
When searching metadata B+ tree, search when root node is opened, key comes comparison according to key word, keyword sequences in the node is carried out binary chop, just can hit until reach leaf node, internal node is equivalent to the index of leaf node, and leaf node directly points to the COW data field;
Solid state hard disc does not have the physical mechanical arm, the random read-write performance is very high, the COW data in exception storehouse volume, directly store in mode at random with solid state hard disc in, wherein the new_chunk of metadata is used for preserving the sector number that the COW data are deposited, after finding new_chunk in meta-data region, direct read/write exception storehouse volume change the chunk position;
Exception storehouse volume free space is by bitmap management, bitmap is pointed to by a Ji Shu, each internal node comprises 64 pointer, point to slave node, leaf node is bitmask, and each leaf node comprises independent position, 1 means and distributes, 0 expression is idle, and the root node position of bitmap is stored in the superblock, and the degree of depth is by instrument size and the common decision of chunk size;
If the degree of depth is 0, then root bitmap piece only comprises single position, and equipment is very little, to such an extent as to 1 piece is just enough as bitmap, if the degree of depth is 1, then root bitmap piece is one group 64 pointer, and the piece of sensing is as single bitmap, if the degree of depth is 2, then be second rank pointer, by that analogy;
When bitmap is searched, begin traversal according to bitmap piece number from root node, during traversal, obtain the index of every one deck by bitmap piece number and the degree of depth, until find leaf node.
The invention has the beneficial effects as follows: the method is to take into account the consideration of performance, utilizes solid state hard disc to improve the method for snapshot performance.The consideration that takes into account performance of mentioning is because if with common hard disc storage snapshot data, the operation meeting of snapshot data reduces the performance of snapshot greatly.Utilizing solid state hard disc to store snapshot data, is to utilize the high random read-write performance of solid state hard disc to improve the speed of snapshot data operation.Of the present invention a kind of with the method for solid state hard disc as the exception warehouse memory device of snapshot, utilize the high random read-write performance of solid state hard disc, improve snapshot to the operating performance of metadata, and the power down metadata do not lose, reached the technique effect of high-performance, high availability snapshot.
Description of drawings
Fig. 1 is the snapshot module structural drawing;
Fig. 2 is that metadata is organized structural drawing;
Fig. 3 is the bitmap organization chart.
Embodiment
Explain below with reference to Figure of description method of the present invention being done.
Concrete technical scheme is as follows:
1 〉. a kind of mode that realizes snapshot with the exception storehouse volume of solid state hard disc as snapshot, is stored in metadata and COW data in the solid state hard disc;
2 〉. the organizational form of metadata in the solid state hard disc with the metadata store of snapshot at solid state hard disc foremost, is stored in the residue section with the COW data;
(1) snapshot module
This example adopts the mode of incremental snapshot, by way of exception rolls up in the storehouse with solid state hard disc, as shown in Figure 1:
Solid state hard disc is divided into meta-data region and cow data field, after source book is set up snapshot, the meta-data region of initialization exception warehouse volume.The form of metadata is (old chunk, new chunk, snapID range), old chunk points to the piece in the source book, new chunk points to piece corresponding to snapped volume, snapID range is the scope of snapshot ID, and each snapshot has an ID, and snapID range is used for showing which snapshot range the piece of metadata appointment is within;
When source book is write, at first search metadata, were it not for and find words, explanation is first revision snapshot source, need to suppose that the position at the data place in snapshot source is old chunk with the data copy in snapshot source to COW equipment, finishing said process needs three steps;
1) searches position for save data at COW equipment, be designated as new chunk;
2) data with the snapshot data source are saved in new chunk place on the COW equipment;
3) be that so-called meta-data preservation is to solid state hard disc with exceptional information (old chunk, new chunk, snapID range);
If find, illustrate that source book upgraded after creating snapshot, the data of source book have been saved in COW equipment and have suffered.Can directly write source book;
When source book is read, directly source book is read to get final product;
When snapped volume is carried out write operation, at first search metadata, if do not find, illustrate that the data at this chunk place of source book are not also revised, new data need to be write on the COW equipment.Search position for save data at COW equipment, be designated as new chunk.Again with meta-data preservation to solid state hard disc;
If find, the direct new chunk that finds is write gets final product;
When snapped volume is carried out read operation, at first search metadata, if do not find, read the corresponding chunk of source book, if find, read the corresponding chunk of snapshot storehouse volume;
(2) exception storehouse volume module
First sector of solid state hard disc is used for the root node position of record metadata B+ tree, the degree of depth and bitmap and the root node of B+ tree as superblock.
As meta-data index storage, metadata is with the form tissue of B+ tree in the front portion of solid state hard disc.The leaf directory entry of B+ tree comprises: old chunk (in the source book), new chunk (in the volume of snapshot storehouse), the snapshot ID scope that this belongs to mapping.The key word (Key) of B+ tree is by old chunk, and snapshot ID range forms.Maximum catalogue item No. in a node has regulation, to such an extent as to a node just fills a piece.
Internal node has identical data structure with leaf node.If internal node, new_chunk then represents the pointer that points to subordinate B+ tree node.If leaf node, new_chunk is exactly the data content that will search.As shown in Figure 2.
When searching metadata B+ tree, search when root node is opened, key comes comparison according to key word, and the keyword sequences in the node is carried out binary chop, just can hit until reach leaf node, and internal node is equivalent to the index of leaf node.Leaf node directly points to the COW data field.
Solid state hard disc does not have the physical mechanical arm, the random read-write performance is very high, the COW data in exception storehouse volume, directly store in mode at random with solid state hard disc in, wherein the new_chunk of metadata is used for preserving the sector number that the COW data are deposited, after finding new_chunk in meta-data region, direct read/write exception storehouse volume change the chunk position.
Exception storehouse volume free space is by bitmap management.Bitmap is pointed to by a Ji Shu.Each internal node comprises 64 pointer, points to slave node, and leaf node is bitmask.Each leaf node comprises independent position, and 1 means and distributes, and 0 expression is idle.The root node position of bitmap is stored in the superblock, and the degree of depth is by instrument size and the common decision of chunk size.
If the degree of depth is 0, then root bitmap piece only comprises single position (equipment is very little, to such an extent as to 1 piece is just enough as bitmap).If the degree of depth is 1, then root bitmap piece is one group 64 pointer, and the piece of sensing is as single bitmap.If the degree of depth is 2, then be second rank pointer, by that analogy.By shown in Figure 3.
When bitmap is searched, begin traversal according to bitmap piece number from root node, during traversal, obtain the index of every one deck by bitmap piece number and the degree of depth, until find leaf node.
Except the described technical characterictic of instructions, be the known technology of those skilled in the art.

Claims (1)

1.一种利用固态硬盘提高快照性能的方法, 其特征在于，该方法是利用固态硬盘高随机读写性能，以固态硬盘作为快照的例外仓卷，将元数据和COW数据存储在固态硬盘中，固态硬盘中元数据的组织方式，是将快照的元数据存储在固态硬盘的最前端，将COW数据存储在剩余区段，以提高快照对元数据的操作性能，系统包括：快照模块（1）、例外仓卷模块（2），具体步骤如下： 1. A method for improving snapshot performance by using solid-state hard drives, characterized in that the method utilizes the high random read and write performance of solid-state hard drives, uses solid-state hard drives as exception storage volumes for snapshots, and stores metadata and COW data in solid-state hard drives , the organization method of metadata in solid-state disk is to store snapshot metadata at the front end of solid-state disk, and store COW data in the remaining segments, so as to improve the operation performance of snapshot on metadata. The system includes: snapshot module (1 ), the exception volume module (2), the specific steps are as follows: （1）快照模块 (1) Snapshot module 采用增量快照的方式，以固态硬盘作为例外仓卷，将固态硬盘分为元数据区和cow数据区，当对源卷建立快照之后，初始化例外仓库卷的元数据区，元数据的形式为old chunk,new chunk,snapID range,其中，old chunk指向源卷中的块，new chunk指向快照卷对应的块，snapID range为快照ID的范围，每个快照都有一个ID，snapID range用于表明元数据指定的块是在哪个快照范围之内； Incremental snapshots are adopted, and the solid-state disk is used as the exception warehouse volume, and the solid-state disk is divided into metadata area and cow data area. After the snapshot is created for the source volume, the metadata area of the exception warehouse volume is initialized. The metadata is in the form of old chunk, new chunk, snapID range, where old chunk points to the block in the source volume, new chunk points to the block corresponding to the snapshot volume, snapID range is the range of snapshot ID, each snapshot has an ID, and snapID range is used to indicate Which snapshot range does the block specified by the metadata belong to; 当对源卷进行写时，首先查找元数据，假如没有查找到话，说明是第一次修改快照源，需要将快照源的数据拷贝到COW设备上，假定快照源的数据所在的位置为旧chunk，完成上述过程需要以下三个步骤： When writing to the source volume, first search for the metadata. If it is not found, it means that the snapshot source is modified for the first time, and the data of the snapshot source needs to be copied to the COW device. Assume that the location of the snapshot source data is old chunk, the following three steps are required to complete the above process: 1)  在COW设备上查找用于保存数据的位置，记为新chunk; 1) Find the location for saving data on the COW device, and record it as a new chunk; 2)  将快照数据源的数据保存到COW设备上新 chunk处； 2) Save the data of the snapshot data source to the new chunk on the COW device; 3)  将例外信息old chunk，new chunk，snapID range即所谓元数据保存至固态硬盘； 3) Save the exception information old chunk, new chunk, snapID range, the so-called metadata, to the SSD; 如果查找到的话，说明在创建快照之后，源卷已经更新过，源卷的数据已经保存到COW设备中了，直接对源卷进行写操作；  If it is found, it means that after the snapshot is created, the source volume has been updated, and the data of the source volume has been saved to the COW device, and the source volume is directly written; 当对源卷进行读操作时，直接对源卷进行读操作即可； When performing a read operation on the source volume, just directly perform the read operation on the source volume; 当对快照卷进行写操作时，首先查找元数据，如果没有查找到的话，说明源卷该chunk处的数据还没有修改，需要将新数据写到COW设备上，在COW设备上查找用于保存数据的位置，记为新chunk，再将元数据保存至固态硬盘上； When writing to the snapshot volume, first search for the metadata, if not found, it means that the data at the chunk of the source volume has not been modified, and the new data needs to be written to the COW device, and the COW device is used to save the data. The location of the data is recorded as a new chunk, and then the metadata is saved to the SSD; 如果查找到的话，直接对查找到的new chunk写操作即可； If you find it, you can directly write to the found new chunk; 当对快照卷进行读操作时，首先查找元数据，如果没有查找到的话，读源卷的对应chunk，如果查找到的话，读快照仓卷的对应chunk； When performing a read operation on a snapshot volume, first search for metadata, if not found, read the corresponding chunk of the source volume, if found, read the corresponding chunk of the snapshot warehouse volume; （2）例外仓卷模块 (2) Exception warehouse volume module 固态硬盘的第一个扇区作为超级块，用于记录元数据B+树的根节点位置、B+树的深度和位图和根节点；   The first sector of the solid-state disk is used as a super block, which is used to record the root node position of the metadata B+ tree, the depth of the B+ tree, the bitmap and the root node;  在固态硬盘的前一部分作为元数据存储区，元数据以B+树的形式组织，B+树的叶子目录项包含：old chunk源卷中，new chunk快照仓卷中，这对映射属于的snapshot ID范围，B+树的关键字Key由old chunk,snapshot ID range 组成，在一个node中的最大目录项号是有规定的，以至于一个node刚好填充一个块；  The first part of the solid-state disk is used as a metadata storage area, and the metadata is organized in the form of a B+ tree. The leaf directory items of the B+ tree include: the old chunk source volume, the new chunk snapshot warehouse volume, and the snapshot ID range that this pair of mappings belongs to , the keyword Key of the B+ tree is composed of old chunk, snapshot ID range, and the maximum directory entry number in a node is specified, so that a node just fills a block; 内部节点与叶子节点有相同的数据结构，如果是内部节点，new_chunk则代表指向从属B+树node的指针，如果是叶子节点，new_chunk就是要查找的数据内容；  The internal node has the same data structure as the leaf node. If it is an internal node, new_chunk represents the pointer to the subordinate B+ tree node. If it is a leaf node, new_chunk is the data content to be searched; 查找元数据B+树时，从根节点开时查找，根据关键字key来比较，对节点内的关键字序列进行二分查找，直到达到叶子节点才能命中，内部节点相当于叶子节点的索引，叶子节点直接指向COW数据区； When looking for the metadata B+ tree, search from the root node, compare according to the keyword key, and perform a binary search on the keyword sequence in the node until it reaches the leaf node. The internal node is equivalent to the index of the leaf node, and the leaf node Directly point to the COW data area; 固态硬盘没有物理机械臂，随机读写性能很高，COW数据在例外仓卷中直接以随机的方式存储与固态硬盘中，其中元数据的new_chunk用于保存COW数据所存的扇区号，当在元数据区查找到new_chunk后，直接读写例外仓卷的改chunk位置； The solid-state hard disk has no physical mechanical arm, and its random reading and writing performance is very high. The COW data is directly stored in the exception warehouse volume and the solid-state hard disk in a random manner. The new_chunk of the metadata is used to save the sector number where the COW data is stored. After the new_chunk is found in the data area, it directly reads and writes the changed chunk position of the exception warehouse volume; 例外仓卷空闲空间由位图管理，位图由一颗基树指向，每个内部节点包含64位的指针，指向从属节点，叶节点为位掩码，每个叶子节点包含单独的位，1意味着已分配，0表示空闲，位图的根节点位置存于超级块中，深度由设备大小和chunk大小共同决定； The free space of the exception volume is managed by a bitmap. The bitmap is pointed to by a base tree. Each internal node contains a 64-bit pointer pointing to a subordinate node. The leaf nodes are bitmasks. Each leaf node contains a single bit. 1 It means allocated, 0 means free, the root node position of the bitmap is stored in the super block, and the depth is determined by the device size and the chunk size; 如果深度是0，则根位图块只包含单个的位，设备非常小，以至于1个块用作位图就足够，如果深度是1，则根位图块是一组64位的指针，指向的块用作单个位图，如果深度是2，则是二级指针，以此类推； If the depth is 0, the root bitmap block contains only a single bit, and the device is so small that 1 block is sufficient for use as a bitmap, if the depth is 1, the root bitmap block is a set of 64-bit pointers, The block pointed to is used as a single bitmap, if the depth is 2, it is a secondary pointer, and so on; 位图查找时，根据位图块号从根节点开始遍历，遍历时，由位图块号和深度求出每一层的索引，直到查找到叶子节点为止。 When searching for a bitmap, start traversing from the root node according to the bitmap block number. When traversing, calculate the index of each layer from the bitmap block number and depth until the leaf node is found.
CN2012104651996A 2012-11-19 2012-11-19 Method for improving snapshot performance by using solid state disk Pending CN102968381A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2012104651996A CN102968381A (en) 2012-11-19 2012-11-19 Method for improving snapshot performance by using solid state disk

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2012104651996A CN102968381A (en) 2012-11-19 2012-11-19 Method for improving snapshot performance by using solid state disk

Publications (1)

Publication Number Publication Date
CN102968381A true CN102968381A (en) 2013-03-13

Family

ID=47798533

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2012104651996A Pending CN102968381A (en) 2012-11-19 2012-11-19 Method for improving snapshot performance by using solid state disk

Country Status (1)

Country Link
CN (1) CN102968381A (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103631676A (en) * 2013-11-06 2014-03-12 华为技术有限公司 Snapshot data generating method and device for read-only snapshot
CN103984609A (en) * 2014-05-28 2014-08-13 华为技术有限公司 Method and device for recovering check points in file system based on copy-on-write
CN104360914A (en) * 2014-10-22 2015-02-18 浪潮(北京)电子信息产业有限公司 Incremental snapshot method and device
CN106326040A (en) * 2016-08-27 2017-01-11 浪潮(北京)电子信息产业有限公司 Method and device for managing snapshot metadata
CN106919342A (en) * 2015-12-28 2017-07-04 成都华为技术有限公司 Storage resource distribution method and device based on automatic simplify configuration
CN107526691A (en) * 2016-06-21 2017-12-29 深圳市中兴微电子技术有限公司 A kind of buffer memory management method and device
CN107665097A (en) * 2017-08-22 2018-02-06 深圳市硅格半导体股份有限公司 A kind of method, apparatus and computer-readable recording medium of solid state hard disc subregion
CN108984357A (en) * 2018-07-16 2018-12-11 郑州云海信息技术有限公司 A kind of hard disk localization method, device and computer readable storage medium
CN109521962A (en) * 2018-11-14 2019-03-26 郑州云海信息技术有限公司 A kind of metadata query method, apparatus, equipment and computer readable storage medium
CN109871355A (en) * 2019-01-23 2019-06-11 杭州宏杉科技股份有限公司 A kind of snapshot metadata storage method, device and equipment, medium
CN110147332A (en) * 2018-02-13 2019-08-20 纬颖科技服务股份有限公司 The management method and memory device of catalogue data
WO2022016863A1 (en) * 2020-07-21 2022-01-27 苏州浪潮智能科技有限公司 Data snapshot method, apparatus and computer device, and storage medium

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101183383A (en) * 2007-12-17 2008-05-21 中国科学院计算技术研究所 A kind of snapshot system and using method thereof
US20100153620A1 (en) * 2008-12-17 2010-06-17 Mckean Brian Storage system snapshot assisted by SSD technology
CN102567214A (en) * 2011-12-01 2012-07-11 浪潮电子信息产业股份有限公司 Method for quickly initializing bitmap page management by RAID5
CN102591790A (en) * 2011-12-30 2012-07-18 记忆科技(深圳)有限公司 Method for implementing data storage snapshot based on solid state disk, and solid state disk
CN102722449A (en) * 2012-05-24 2012-10-10 中国科学院计算技术研究所 Key-Value local storage method and system based on solid state disk (SSD)

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101183383A (en) * 2007-12-17 2008-05-21 中国科学院计算技术研究所 A kind of snapshot system and using method thereof
US20100153620A1 (en) * 2008-12-17 2010-06-17 Mckean Brian Storage system snapshot assisted by SSD technology
CN102567214A (en) * 2011-12-01 2012-07-11 浪潮电子信息产业股份有限公司 Method for quickly initializing bitmap page management by RAID5
CN102591790A (en) * 2011-12-30 2012-07-18 记忆科技(深圳)有限公司 Method for implementing data storage snapshot based on solid state disk, and solid state disk
CN102722449A (en) * 2012-05-24 2012-10-10 中国科学院计算技术研究所 Key-Value local storage method and system based on solid state disk (SSD)

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张宇 等: "一种在存储子系统中实现依赖快照的技术", 《计算机工程与应用》 *
杨巍: "采用基树的磁盘阵列 Cache 技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103631676A (en) * 2013-11-06 2014-03-12 华为技术有限公司 Snapshot data generating method and device for read-only snapshot
CN103631676B (en) * 2013-11-06 2016-08-24 华为技术有限公司 The snapshot data of a kind of read-only materialized view generates method and device
CN103984609B (en) * 2014-05-28 2017-06-16 华为技术有限公司 A kind of method and apparatus that checkpoint is reclaimed in file system based on copy-on-write
CN103984609A (en) * 2014-05-28 2014-08-13 华为技术有限公司 Method and device for recovering check points in file system based on copy-on-write
WO2015180394A1 (en) * 2014-05-28 2015-12-03 华为技术有限公司 Method and device for recovering checkpoint in copy on write based file system
CN104360914B (en) * 2014-10-22 2017-10-13 浪潮(北京)电子信息产业有限公司 Incremental snapshot method and apparatus
CN104360914A (en) * 2014-10-22 2015-02-18 浪潮(北京)电子信息产业有限公司 Incremental snapshot method and device
CN106919342A (en) * 2015-12-28 2017-07-04 成都华为技术有限公司 Storage resource distribution method and device based on automatic simplify configuration
CN107526691A (en) * 2016-06-21 2017-12-29 深圳市中兴微电子技术有限公司 A kind of buffer memory management method and device
CN106326040A (en) * 2016-08-27 2017-01-11 浪潮(北京)电子信息产业有限公司 Method and device for managing snapshot metadata
CN106326040B (en) * 2016-08-27 2019-12-31 苏州浪潮智能科技有限公司 Snapshot metadata management method and device
CN107665097A (en) * 2017-08-22 2018-02-06 深圳市硅格半导体股份有限公司 A kind of method, apparatus and computer-readable recording medium of solid state hard disc subregion
CN110147332A (en) * 2018-02-13 2019-08-20 纬颖科技服务股份有限公司 The management method and memory device of catalogue data
CN108984357A (en) * 2018-07-16 2018-12-11 郑州云海信息技术有限公司 A kind of hard disk localization method, device and computer readable storage medium
CN109521962A (en) * 2018-11-14 2019-03-26 郑州云海信息技术有限公司 A kind of metadata query method, apparatus, equipment and computer readable storage medium
CN109871355A (en) * 2019-01-23 2019-06-11 杭州宏杉科技股份有限公司 A kind of snapshot metadata storage method, device and equipment, medium
WO2022016863A1 (en) * 2020-07-21 2022-01-27 苏州浪潮智能科技有限公司 Data snapshot method, apparatus and computer device, and storage medium
US11907076B2 (en) 2020-07-21 2024-02-20 Inspur Suzhou Intelligent Technology Co., Ltd. Data snapshot method and apparatus, computer device, and storage medium

Similar Documents

Publication Publication Date Title
CN102968381A (en) Method for improving snapshot performance by using solid state disk
CN110825748B (en) High-performance and easily-expandable key value storage method by utilizing differentiated indexing mechanism
US10956071B2 (en) Container key value store for data storage devices
US9342256B2 (en) Epoch based storage management for a storage device
CN107862064B (en) High-performance and extensible lightweight file system based on NVM (non-volatile memory)
US9489297B2 (en) Pregroomer for storage array
EP2735978B1 (en) Storage system and management method used for metadata of cluster file system
US10055420B1 (en) Method to optimize random IOS of a storage device for multiple versions of backups using incremental metadata
US8819367B1 (en) Accelerated translation power recovery
US8612719B2 (en) Methods for optimizing data movement in solid state devices
CN103150258B (en) Writing, reading and garbage collection method of solid-state memory system
CN103902669B (en) A kind of separate type file system based on different storage mediums
CN106528001B (en) A kind of caching system based on nonvolatile memory and software RAID
CN114281762B (en) A log storage acceleration method, device, device and medium
CN103838853B (en) Mixed file system based on different storage media
CN105659204A (en) Method and apparatus for performing annotated atomic write operations
CN105868396A (en) Multi-version control method of memory file system
CN105453021A (en) Systems and methods for atomic storage operations
CN110347852A (en) It is embedded in the file system and file management method of key assignments storage system extending transversely
CN107391774A (en) The rubbish recovering method of JFS based on data de-duplication
US9996426B1 (en) Sparse segment trees for high metadata churn workloads
WO2022262381A1 (en) Data compression method and apparatus
CN102542057B (en) High dimension data index structure design method based on solid state hard disk
WO2017113211A1 (en) Method and device for processing access request, and computer system
CN103473337A (en) Massive catalogs and files oriented processing method in distributed type storage system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20130313