CN1955939A - Backup and recovery method based on virtual flash disk - Google Patents
Backup and recovery method based on virtual flash disk Download PDFInfo
- Publication number
- CN1955939A CN1955939A CN 200610113704 CN200610113704A CN1955939A CN 1955939 A CN1955939 A CN 1955939A CN 200610113704 CN200610113704 CN 200610113704 CN 200610113704 A CN200610113704 A CN 200610113704A CN 1955939 A CN1955939 A CN 1955939A
- Authority
- CN
- China
- Prior art keywords
- data block
- backup
- data
- ssd
- disk
- 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
Links
Images
Abstract
A restoring and backing up method based on virtual internal memory disc includes utilizing asynchronous mirror image policy of snapshot and white-writing copy and using bit map to mark state of each data block in virtual internal memory disc module for realizing restoration and back-up of internal memory under control of asynchronous scanning linear program.
Description
Technical field
Backup and restoration methods based on virtual flash disk belong to field of storage area networks, and the backup and the data that relate in particular to are wherein recovered.
Background technology
Solid magnetic disc (Solid-state disk, SSD) be a kind of high-performance and with the External memory equipment of traditional magnetic disk compatibility, the disc that it uses memory chip (as SDRAM or Flash chip) to replace rotating in the traditional magnetic disk comes store data, and use corresponding techniques to ensure the persistence of data, eliminate the mechanical delay of memory module fully, improved the Performance And Reliability of using greatly.Therefore, SSD equipment is applied to just more and more to data bandwidth and delay requirement during higher key storage is used.
SSD distinguishes according to the storage chip type of its use, mainly is divided into based on the SSD of Flash with based on the SSD two big classes of SDRAM.The former uses non-volatile Flash chip to store data, and data still are kept perfectly after the chip power down.Yet because the Flash chip unit just can lose efficacy after hundreds of thousands writes to millions of times, this has just caused such SSD to be unsuitable for as frequent data updated storage.Using at present is a kind of SSD in back more widely, and the internal memory that its realization is based at a high speed, easily loses is such as SDRAM.This type of SSD has very high data access speed and low-down delay, can be used for quickening the application that can't improve performance because of disk bottleneck.Yet because the easy mistake characteristic of SDRAM, the data after system's power down among the SSD can be lost fully, therefore how to ensure that the integrality of its data is current major issues.
The persistence that a kind of method is to use inner battery and backup disk to ensure data.In case power-fail, battery can support this equipment all to be dumped to backup disk up to data.When system restart, data can copy in the work internal memory via backup disk.The SSD product of favourable in the market in this way realization, its performance is very superior, but owing to be that devices at full hardware realizes, cost is very high, after system's power down, need to spend the persistence of plenty of time disk synchrodata, and the support that recovers for the assurance of data consistency and data is not enough with service data.
In our work formerly, realized SSD system---the SSD of Tsing-Hua University (TH-SSD) of oneself based on the method for virtual flash disk Visual Disk (being called for short VD), the main memory and the backup disk of storage networking control end node are invented unified equipment VD, hide it to upper layer application and realize details, the service of SSD is provided to the memory access of front end.Because the realization carrier of VD is the main memory of volatibility, therefore need take measures to protect data wherein, in case locking system collapse or restart the loss of data that is caused.
Among the present invention, we have proposed the asynchronous mirroring strategy based on snapshot (Snapshot) and Copy on write (Copy-On-Write), with the persistence and the consistance of dynamic maintenance data.The method that this method adopts snapshot and Copy on write to combine makes the magnetic disc storage backup procedure begin partial data among the eve VD.When after system breaks down, restarting, can come data are recovered by the backup disk and the interim daily record of unanimity.Compare with other speedy storage equipment based on DRAM, this method has reduced after system's power down the plenty of time to the disk synchrodata is spent, the support that recovers for the assurance of data consistency and data is good, and it is low to have a cost, portable advantage such as good.
Summary of the invention
Backup and the restoration methods that the object of the present invention is to provide a kind of efficient, portable and be easy to realize based on virtual flash disk.
The invention is characterized in that described method contains following steps successively:
Step (1): on storage networking control end node, be provided with:
Virtual memory disk module SSD_VD, the internal memory of this storage networking control end node and backup disk are invented one to the transparent memory pool of upper layer application, this module is divided into array and the denotation token that a data block is a unit to the virtual flash disk VD in this memory pool, uses the state variation situation of each bit representation respective data blocks in the following bitmap:
Data block X: only refer to the data block of before the scanning T1 zero hour, being visited by write operation, at the piece front end of this data block with position 01 sign;
Data block O: the data block of constantly being visited by write operation between the T2 only at the scanning T1 zero hour and the end of scan; Piece front end in this data block is represented with position 10;
Data block OX: refer to be visited by write operation before T1, T1 represents with position 11 at the piece front end of this data block to the data block of being visited by write operation again between the T2 constantly;
Data block CLEAN:, represent with position 00 at the piece front end of this data block to the data block that T2 is visited by write operation not yet
In this virtual flash disk VD, also be provided with:
Copy on write table COW is used to deposit the data block that backup procedure begins afterwards to be equipped with in the not scan bits of VD the write operation request of access, is data block X but be not scanned the former meaningful of position at this;
Backup Data chained list new_dirty_list is used to deposit and retouches the data block that there is write access request the position at VD after backup procedure begins;
Small computer system interface module TH_SSD_SCSI links to each other with the virtual memory module, is responsible for handling read write command;
Host module links to each other with the virtual memory module, is the interface of this storage networking control end node and management host;
Step (2): described method realizes backup operation by the virtual memory disk module according to following steps:
Step (2.1): start async backup thread ssd_workers constantly at T1, scan all data blocks, the data block that is labeled as X is write to the relevant position of backup disk;
Step (2.2): handle to this thread ssd_workers between the T2 finish time at moment T1, the I/O processing threads receives that write order handles respectively according to following situation on this network storage control end node minicomputer interface module:
(a): if writing position is before the data block pointer curr that thread ssd_workers arrives when pre-treatment, illustrate that pointer curr has pointed to the zone that this thread ssd_workers has finished dealing with, the zone bit of this data block bitmap correspondence directly is set so that when thread ssd_workers startup next time, as a token of handle for the data block of X;
(b): if writing position after pointer curr, is then handled respectively according to following situation:
If to have pointed to was the data block that non-O is X originally, only need the piece of this data block number is recorded in the new_dirty_list table, corresponding marker bit is not set;
Constantly visited at T1 if pointed to originally by write operation, the data block X that between T1 and T2, is visited by write operation again, then the zone bit of this data block is emptied, this piece content and piece number are recorded in the COW chained list, again the content that the writes O class data virtual flash disk VD that writes direct, piece number is recorded in the new_dirty_list table;
Step (2.3): at T2 constantly, all data blocks are traverse scanning, locking virtual flash disk VD, and the corresponding positions of all data blocks in bitmap among the mark new_dirty_list, and empty this chained list, release virtual flash disk VD; At last all data blocks in the COW table are write the backup disk relevant position, the dormancy of ssd_workers thread;
Step (2.4): for the data block that enters VD later on, the corresponding position of a mark;
Step (3): the virtual memory disk module carries out data in the following manner and recovers: from the 0th sector of backup disk, read in the data of data block size in order successively and put into the data block of VD correspondence;
2. according to claim: described backup and restoration methods based on virtual flash disk is characterized in that, asynchronous scanning thread is to be waken up by system clock at interval every regular time, or is being reached threshold value by the number of data block to be written and waken up by the minicomputer interface module.
Description of drawings
The overall software architecture figure of Fig. 1 .TH-SSD.
Fig. 2 .ssd_workers thread work principle.
Fig. 3. based on the mirror back-up mode principle of work of snapshot and Copy on write.
Fig. 4. system construction drawing.
Fig. 5. the readwrite performance of system during backup.
Fig. 6. system flowchart.
Embodiment
Data consistency is to need a major issue considering in the data backup process.Data backup is not simply data to be transferred on the persistent storage medium from volatile storage medium, but will guarantee that the data that shifted have correctly reflected the store status in a certain moment.The method that solves the data consistency problem all depends on persistent storage.Use with storage between contract be exactly the rreturn value that writes the IO operation, when rreturn value is that application thinks that just this partial data is by persistence successfully the time.Because the characteristic of two kinds of storage mediums that are commonly used at present need weighed at a slow speed and between non-volatile disk and the DRAM main memory quick and that easily lose when people store data.
SSD system of Tsing-Hua University (TH-SSD) is our independently developed solid storage device, also is the basis of backing up.It mainly is made up of following three submodules: SCSI (Small Computer System Interface, small computer system interface) module TH-SSD_SCSI, host module TH-SSD_HOST and cache module TH-SSD_Cache.The SCSI module is responsible for handling read write command, and host module is the interface of storage and management main frame, and cache module then utilizes the overall performance of Cache mechanism raising system.The realization of three subsystems all depends on a corn module---virtual memory disk module (being called for short SSD-VD), its major function is that the internal memory of storage networking control end node and backup disk are invented unified equipment, with operating system is that the internal memory that SSD keeps is packaged into a memory pool that can conveniently use, a small amount of interface outwards is provided and hides its complicated inside realization, make that the management of storage space is transparent to upper layer application.SSD-VD has adopted the tracking strategy that writes based on bitmap in the management to virtual flash disk VD, with the write operation of monitoring to VD.In the realization, being divided into VD with data block (being called for short trunk) is the array of unit, with the state variation situation of each record respective data blocks in the bitmap.Front end will be provided with the corresponding positions of data block in bitmap when VD writes data.The backup of SSD-VD and restoration methods are exactly that its operating strategy in this bitmap realizes.
Adopt a backup disk (using hardware RAID1) corresponding one by one based on the asynchronous mirroring backup of Snapshot and Copy-On-Write (COW) with the content among the SSD-VD in order to guarantee reliability, the integrality that can guarantee the async backup copy (obtains the reflection in the unanimity of the system that backs up the beginning eve, promptly be somebody's turn to do snapshot, i.e. a Snapshot of system constantly).Its implementation is: in backup procedure, if there is new write request to arrive, then original location conten among the VD is copied in the middle of interim COW (Copy-On-Write) table space.In realization, the COW table is the chained list of the data block copy of the medium backup disk to be written of VD.
Implementation is described below: asynchronous scanning thread ssd_workers waken up by system clock at set intervals or since the number of data block to be written reach threshold value and waken up by the SCSI module.The whole bitmap region of its sequential scanning writes backup disk with the data in the piece of revising.In this process,, run into different data blocks and use different strategies to handle in order to guarantee the consistance of data.When asynchronous scanning thread ssd_workers moves, if be scanned the zone write operation is not arranged, then need in the meaningful COW of copying to the table in this data block Central Plains, and record, when treating that asynchronous scanning thread ssd_workers scans this data block, again the corresponding contents in the COW table is write back backup disk.In this way, this algorithm can guarantee to back up content for backing up the accurate copy of the zero hour.
In the specific implementation, at first be defined as follows notion:
Definition of T 1: asynchronous scanning thread ssd_workers starts constantly
Definition of T 2: asynchronous scanning thread ssd_workers handles the finish time
For data block, be defined as follows state:
Definition X: the data block of before T1, being visited only by write operation
Definition O: the data block of between T1 and T2, being visited only by write operation
Definition OX: write before T1, T1 is to the data block of being visited by write operation again between the T2
Definition CLEAN: the data block of not write yet to T2
Definition bitmap bitmap: do not have data block of two bit representations, and represent above-mentioned four kinds of states respectively, be i.e. 00 (CLEAN), 01 (X), 10 (O), 11 (OX) by coding
Definition pointer curr (current position): the data block that asynchronous scanning thread ssd_workers arrives when pre-treatment
In addition, two list structures of definition:
Definition COW table: after being used to deposit backup procedure and beginning not scan bits be equipped with the request of writing and original location conten is exactly the data block of X
Definition new_dirty_list table: after being used to deposit backup procedure and beginning not scan bits be equipped with the data block of the request of writing
By definition as can be known: have only O to the T2 at T1.
The concrete scanning strategy of ssd_workers is as follows:
T1 begins constantly, travels through all data blocks, finds the data block that is labeled as X, and it is write to the relevant position of backup disk, and the curr value is moved one backward.When the I/O processing threads in middle layer runs into write order in the meantime:
If writing position is before curr, explanation is to be positioned at the zone that ssd_workers finishes dealing with, and the corresponding positions in the bitmap directly is set, and this data block will be used as X and handle when ssd_workers startup next time;
If writing position is after curr:
Originally not O but X data block: only the piece of this data block number is recorded in the new_dirty_list table, the position in the bitmap is not set.
Originally be exactly the data block of OX: the X content is read, empty the corresponding positions in the bitmap, this piece content and piece number are recorded in the COW chained list, the content of the O VD that writes direct number is recorded in piece in the new-dirty_list table.
Treat that all trunk traversal finishes (T2 is constantly), locking VD, the corresponding positions of all data blocks in bitmap in the mark new_dirty_list table, and empty this chained list, release VD; At last all data blocks in the COW table are write the backup disk relevant position, the dormancy of ssd_workers thread.
For the I/O write request of all arrivals afterwards, all corresponding positions among the set figure.
In T2 finish time, kept the T1 partial data among the VD constantly on the backup disk.The write operation that T1 and T2 take place constantly only acts on the data among the VD, does not note down on backup disk.In next backup procedure, just these changes can be recorded on the backup disk.Mirror image on the backup disk is incomplete at T1 and T2 constantly, and the mirror image on the backup disk is complete before the ssd_workers thread starts next time.This method implements comparatively simple, does not need extra hardware device, has the characteristics of high efficiency, low cost.
When system restarted, data recovery procedure was as follows: virtual flash disk is from the 0th sector of backup disk, reads in the data of trunk size in order successively and puts into the data block of the correspondence of VD.Like this, just realized the reenterability of virtual flash disk.
In conjunction with the accompanying drawings, details are as follows in the present invention:
1. hardware device and network
Hardware device required for the present invention comprises front end main frame, I/O node, high capacity SDRAM and backup disk; Network comprises fiber optic network and Ethernet.Wherein, fiber optic network connects front end main frame, I/O node by optical fiber switch, carries out data transmission; Ethernet connects the I/O node by the ether switch and is managed for configuration.High capacity SDRAM uses to realize SSD for the front end main frame by the read-write behavior of mock disc on I/O node machine.Native system is the maximum SSD that can support 64GB under 32 bit patterns.The (SuSE) Linux OS (2.4.18 above kernel) of I/O node operation through transforming, the front end main frame is any mainstream operation system, comprises Linux, Windows and Solaris.
2. system loads
1) Linux of I/O node machine starts
I/O node machine: utilize the internal memory analog hard disk to realize the target machine of virtual flash disk.
The Linux that moves on I/O node machine is for revising the embedded Linux system of system start-up code.When system start-up, when running to memory headroom detection and internal memory initialization, only a small amount of internal memory is given operating system management to keep the required physical memory of system's operation, keep a large amount of high-end physical memories and transfer to the virtual memory disk management.Its result is kept in the e820 structure.E820 is a kernel data structure of Linux memory management, and it offers all physical memories of system in the mode of segmentation, and the internal memory of each section has identical attribute.In realization, we will keep the memory marker of giving virtual flash disk and be " virtual disk ", so just realize the purpose that keeps.
2) load kernel module th_scsi_target.o, initialization virtual flash disk.
Kernel module th_scsi_target.o: this module is the primary module of virtual flash disk.It has two functions: initialization I/O node machine and processing I./O order.This module is finished following function successively:
1. the information among the e820map that creates when starting according to Linux is mapped as continuous virtual disk address space to the discontinuous memory block in address in the memory pool, again this space is divided according to piece so that manage, utilized bitmap to write down each fast operating position.
2. start kernel thread ssd_read_write and finish various read write commands.This thread meets the iSCSI protocol specification, and the intercepting order is carried out read-write operation to corresponding data block in the virtual flash disk then from the protocol data bag, returns from the iSCSI agreement then.
3. based on the async backup of snapshot and Copy on write, it regularly writes backup disk with the data of having changed of virtual flash disk.Its process is as follows:
1. when loading, virtual flash disk starts thread ssd_works.This thread regularly scans the bitmap of virtual flash disk, and carries out different operations according to each different conditions of bitmap:
2. to the memory block of X state, call Linux bottom read scsi_do_req this piece is write hard disk, and in call back function, change the state of this piece correspondence in the bitmap into CLEAN;
3. to the piece of O state, only need change the state of this piece correspondence into X in bitmap;
4. to the piece of OX state, its deposit data is shown at new_dirty_list, calls Linux bottom read scsi_do_req it is write backup disk and changes the state of the piece of bitmap correspondence into X.
Claims (2)
1. based on the backup and the restoration methods of virtual flash disk, its special this is that described method contains following steps successively:
Step (1): on storage networking control end node, be provided with:
Virtual memory disk module SSD_VD, the internal memory of this storage networking control end node and backup disk are invented one to the transparent memory pool of upper layer application, this module is divided into array and the denotation token that a data block is a unit to the virtual flash disk VD in this memory pool, uses the state variation situation of each bit representation respective data blocks in the following bitmap:
Data block X: only refer to the data block of before the scanning T1 zero hour, being visited by write operation, at the piece front end of this data block with position 01 sign;
Data block O: the data block of constantly being visited by write operation between the T2 only at the scanning T1 zero hour and the end of scan; Piece front end in this data block is represented with position 10;
Data block OX: refer to be visited by write operation before T1, T1 represents with position 11 at the piece front end of this data block to the data block of being visited by write operation again between the T2 constantly;
Data block CLEAN: the data block of being visited by write operation not yet to T2, be illustrated among this virtual flash disk VD with position 00 at the piece front end of this data block, also be provided with:
Copy on write table COW is used to deposit the data block that backup procedure begins afterwards to be equipped with in the not scan bits of VD the write operation request of access, is data block X but be not scanned the former meaningful of position at this;
Backup Data chained list new_dirty_list, be used to deposit backup procedure begin the back VD not scan bits be equipped with the data block of write access request;
Small computer system interface module TH_SSD_SCSI links to each other with the virtual memory module, is responsible for handling read write command;
Host module links to each other with the virtual memory module, is the interface of this storage networking control end node and management host;
Step (2): described method realizes backup operation by the virtual memory disk module according to following steps:
Step (2.1): start async backup thread ssd_workers constantly at T1, scan all data blocks, the data block that is labeled as X is write to the relevant position of backup disk;
Step (2.2): handle to this thread ssd_workers between the T2 finish time at moment T1, the I/O processing threads receives that write order handles respectively according to following situation on this network storage control end node minicomputer interface module:
(a): if writing position is before the data block pointer curr that thread ssd_workers arrives when pre-treatment, illustrate that pointer curr has pointed to the zone that this thread ssd_workers has finished dealing with, the zone bit of this data block bitmap correspondence directly is set so that when thread ssd_workers startup next time, as a token of handle for the data block of X;
(b): if writing position after pointer curr, is then handled respectively according to following situation:
If to have pointed to was the data block that non-O is X originally, only need the piece of this data block number is recorded in the new_dirty_list table, corresponding marker bit is not set;
Constantly visited at T1 if pointed to originally by write operation, the data block X that between T1 and T2, is visited by write operation again, then the zone bit of this data block is emptied, this piece content and piece number are recorded in the COW chained list, again the content that the writes O class data virtual flash disk VD that writes direct, piece number is recorded in the new_dirty_list table;
Step (2.3): at T2 constantly, all data blocks are traverse scanning, locking virtual flash disk VD, and the corresponding positions of all data blocks in bitmap in the mark new_dirty_list table, and empty this chained list, release virtual flash disk VD; At last all data blocks in the COW table are write the backup disk relevant position, the dormancy of ssd_workers thread;
Step (2.4): for the data block that enters VD later on, the corresponding position of a mark;
Step (3): the virtual memory disk module carries out data in the following manner and recovers: from the 0th sector of backup disk, read in the data of data block size in order successively and put into the data block of VD correspondence;
2. according to claim: described backup and restoration methods based on virtual flash disk is characterized in that, asynchronous scanning thread is to be waken up by system clock at interval every regular time, or is being reached threshold value by the number of data block to be written and waken up by the minicomputer interface module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610113704 CN1955939A (en) | 2006-10-13 | 2006-10-13 | Backup and recovery method based on virtual flash disk |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610113704 CN1955939A (en) | 2006-10-13 | 2006-10-13 | Backup and recovery method based on virtual flash disk |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1955939A true CN1955939A (en) | 2007-05-02 |
Family
ID=38063280
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610113704 Pending CN1955939A (en) | 2006-10-13 | 2006-10-13 | Backup and recovery method based on virtual flash disk |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1955939A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101458655B (en) * | 2007-12-12 | 2011-01-05 | 成都市华为赛门铁克科技有限公司 | Internal memory management method, system and internal memory control device |
| CN102024044A (en) * | 2010-12-08 | 2011-04-20 | 华为技术有限公司 | Distributed file system |
| CN102331949A (en) * | 2011-10-12 | 2012-01-25 | 华为技术有限公司 | Methods for generating and restoring memory snapshot of virtual machine, device and system |
| CN102460415A (en) * | 2009-05-13 | 2012-05-16 | 惠普开发有限公司 | System for virtual disks version control |
| CN101840362B (en) * | 2009-10-28 | 2012-05-30 | 创新科存储技术有限公司 | Method and device for achieving copy-on-write snapshot |
| CN102662799A (en) * | 2012-04-13 | 2012-09-12 | 华为技术有限公司 | Data backup method, server and hot backup system |
| CN104102750A (en) * | 2013-04-11 | 2014-10-15 | 深圳市腾讯计算机系统有限公司 | Operating data recording method, device and server |
| CN105376269A (en) * | 2014-08-11 | 2016-03-02 | 中国电信股份有限公司 | Virtual machine storage system, virtual machine storage system realizing method and virtual machine storage system realizing device |
| CN107797770A (en) * | 2017-11-07 | 2018-03-13 | 深圳神州数码云科数据技术有限公司 | A kind of synchronous method and device of Disk State information |
| CN109450974A (en) * | 2018-10-08 | 2019-03-08 | 郑州云海信息技术有限公司 | A kind of shared equipment high-speed data transmission method and system |
| CN110837442A (en) * | 2019-11-14 | 2020-02-25 | 北京京航计算通讯研究所 | KVM virtual machine backup system based on dirty data bitmap and network block equipment |
| CN110837441A (en) * | 2019-11-14 | 2020-02-25 | 北京京航计算通讯研究所 | KVM virtual machine backup method based on dirty data bitmap and network block equipment |
| CN111352579A (en) * | 2018-12-24 | 2020-06-30 | 深圳市茁壮网络股份有限公司 | Data storage method and device |
-
2006
- 2006-10-13 CN CN 200610113704 patent/CN1955939A/en active Pending
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101458655B (en) * | 2007-12-12 | 2011-01-05 | 成都市华为赛门铁克科技有限公司 | Internal memory management method, system and internal memory control device |
| CN102460415A (en) * | 2009-05-13 | 2012-05-16 | 惠普开发有限公司 | System for virtual disks version control |
| CN101840362B (en) * | 2009-10-28 | 2012-05-30 | 创新科存储技术有限公司 | Method and device for achieving copy-on-write snapshot |
| CN102024044A (en) * | 2010-12-08 | 2011-04-20 | 华为技术有限公司 | Distributed file system |
| CN102024044B (en) * | 2010-12-08 | 2012-11-21 | 华为技术有限公司 | Distributed file system |
| CN102331949A (en) * | 2011-10-12 | 2012-01-25 | 华为技术有限公司 | Methods for generating and restoring memory snapshot of virtual machine, device and system |
| CN102331949B (en) * | 2011-10-12 | 2014-11-05 | 华为技术有限公司 | Methods for generating and restoring memory snapshot of virtual machine, device and system |
| US9507672B2 (en) | 2011-10-12 | 2016-11-29 | Huawei Technologies Co., Ltd. | Method, apparatus, and system for generating and recovering memory snapshot of virtual machine |
| CN102662799A (en) * | 2012-04-13 | 2012-09-12 | 华为技术有限公司 | Data backup method, server and hot backup system |
| CN102662799B (en) * | 2012-04-13 | 2015-01-21 | 华为技术有限公司 | Data backup method, server and hot backup system |
| CN104102750B (en) * | 2013-04-11 | 2018-04-27 | 深圳市腾讯计算机系统有限公司 | Operation data recording method, device and server |
| CN104102750A (en) * | 2013-04-11 | 2014-10-15 | 深圳市腾讯计算机系统有限公司 | Operating data recording method, device and server |
| CN105376269A (en) * | 2014-08-11 | 2016-03-02 | 中国电信股份有限公司 | Virtual machine storage system, virtual machine storage system realizing method and virtual machine storage system realizing device |
| CN105376269B (en) * | 2014-08-11 | 2019-11-26 | 中国电信股份有限公司 | Virtual machine storage system and its implementation and device |
| CN107797770A (en) * | 2017-11-07 | 2018-03-13 | 深圳神州数码云科数据技术有限公司 | A kind of synchronous method and device of Disk State information |
| CN107797770B (en) * | 2017-11-07 | 2020-08-21 | 深圳神州数码云科数据技术有限公司 | Method and device for synchronizing disk state information |
| CN109450974A (en) * | 2018-10-08 | 2019-03-08 | 郑州云海信息技术有限公司 | A kind of shared equipment high-speed data transmission method and system |
| CN109450974B (en) * | 2018-10-08 | 2021-08-20 | 郑州云海信息技术有限公司 | High-speed data transmission method and system for shared equipment |
| CN111352579A (en) * | 2018-12-24 | 2020-06-30 | 深圳市茁壮网络股份有限公司 | Data storage method and device |
| CN111352579B (en) * | 2018-12-24 | 2023-11-10 | 深圳市茁壮网络股份有限公司 | Data storage method and device |
| CN110837442A (en) * | 2019-11-14 | 2020-02-25 | 北京京航计算通讯研究所 | KVM virtual machine backup system based on dirty data bitmap and network block equipment |
| CN110837441A (en) * | 2019-11-14 | 2020-02-25 | 北京京航计算通讯研究所 | KVM virtual machine backup method based on dirty data bitmap and network block equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1955939A (en) | Backup and recovery method based on virtual flash disk | |
| US11249951B2 (en) | Heuristic interface for enabling a computer device to utilize data property-based data placement inside a nonvolatile memory device | |
| US20210303407A1 (en) | Management of volume snapshots in a data storage system | |
| US10073656B2 (en) | Systems and methods for storage virtualization | |
| CN110795206B (en) | System and method for facilitating cluster-level caching and memory space | |
| US8219776B2 (en) | Logical-to-physical address translation for solid state disks | |
| US9342256B2 (en) | Epoch based storage management for a storage device | |
| US8166233B2 (en) | Garbage collection for solid state disks | |
| US8307154B2 (en) | System and method for performing rapid data snapshots | |
| US9778860B2 (en) | Re-TRIM of free space within VHDX | |
| US8868869B2 (en) | Enhanced copy-on-write operation for solid state drives | |
| CN1945537A (en) | Method for realizing high speed solid storage device based on storage region network | |
| US9535628B2 (en) | Memory system with shared file system | |
| CN104903872A (en) | Systems, methods, and interfaces for adaptive persistence | |
| US8924675B1 (en) | Selective migration of physical data | |
| US8572338B1 (en) | Systems and methods for creating space-saving snapshots | |
| CN1704903A (en) | Snapshot backup method | |
| CN1770115A (en) | Recovery operations in storage networks | |
| CN101030165A (en) | Magnetic disk space management and managing system | |
| US20190042462A1 (en) | Checkpointing for dram-less ssd | |
| US11269771B2 (en) | Storage device for improving journal replay, operating method thereof, and electronic device including the storage device | |
| US11119657B2 (en) | Dynamic access in flash system | |
| CN1716176A (en) | Data transmission device having the shape of a standard 3.5' disk | |
| CN105335220A (en) | Transaction recovery method and apparatus for storage device | |
| CN1564517A (en) | Memory-network memory-magnetic disc high speed reliable storage system and its reading/writing method |
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 |