CN104391802A - Streamline pool metadata node refreshing consistency protecting method - Google Patents
Streamline pool metadata node refreshing consistency protecting method Download PDFInfo
- Publication number
- CN104391802A CN104391802A CN201410678460.XA CN201410678460A CN104391802A CN 104391802 A CN104391802 A CN 104391802A CN 201410678460 A CN201410678460 A CN 201410678460A CN 104391802 A CN104391802 A CN 104391802A
- Authority
- CN
- China
- Prior art keywords
- node
- metadata
- tree
- write
- refreshing
- 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
Abstract
The invention provides a streamline pool metadata node refreshing consistency protecting method, belonging to the technical field of data storage. The streamline pool metadata node refreshing consistency protecting method comprises the steps of: backing up metadata cache mirror images and logs to a redundant controller; when inserting and deleting B+Tree nodes of metadata, using a copy-on-write mode from root nodes to leaf nodes, firstly refreshing and writing each B+Tree updated node while refreshing and writing and then refreshing and writing super-level blocks pointing to the root nodes; mirroring the B+Tree node inserting and deleting operation of the metadata to an opposite-end redundant controller by adopting a log mode, recording newly added and deleted leaf nodes, and recording B+Tree refreshing and writing time; when a controller at a main control end is ineffective, inserting and deleting information again by virtue of the logs and a version number to recover data. As the metadata are mirrored to the redundant controller remotely by virtue of the log mode, metadata consistency is ensured.
Description
Technical field
The present invention relates to data storage technology, refer more particularly to the metadata processing method of automatically simplifying configuration in high-end disk array, is that one simplifies pond metadata node refreshing method for protecting consistency specifically.
Background technology
High-end disk array adopts multiple controller Redundancy Design, the High Availabitity service storing data is provided, the Premium Features of its storage pool simplify configuration automatically, storage resources is adopted " when writing, to distribute " strategy, by changing the resource distribution according to need of storage system, improve disk storage space utilization factor, reach the lower deployment cost reducing storage system and the object economized on resources.
Have the storage pool abbreviation of automatically simplifying configuration feature and simplify pond, simplify pond and be structurally divided into data field and meta-data region, data field is for storing data.Meta-data region includes storage pool superblock, metadata bitmap, data bitmap, and logical volume information etc., to certainly simplifying storage pool organizer and governor.
For simplifying the metadata space in pond, data space, logical volume data block map information all adopt B+Tree structure organization to manage, connected by block pointer between B+Tree node, the advantage of this B+Tree organizational form is that the inquiry number of plies is few, efficiency is higher, be loaded into internal memory and do not need conversion, but when inserting B+Tree node, deletion action time, again to misplace movement to the child node pointer of B+Tree multilayer node or leaf node content, there is following problem:
When 1, increasing deletion of node, to metadata change amount, (each node is metadata spatial data block size comparatively greatly, >=4KB), if write in a synchronous manner, postpone greatly, affect performance, if write in an asynchronous manner, when controller power down or inefficacy can cause data loss problem;
2, in renewal process, generation systems power down or controller inefficacy can cause tree construction entanglement, loss, because each insertion/deletion of node may relate to the dislocation of multilayer node and to change and node takes up room and can be greater than 512 byte sector sizes, controller power down or moment of failure can not ensure to write completely.
Summary of the invention
The invention process proposes one and simplifies pond metadata node and refresh method for protecting consistency, effectively can solve the two problems that above-mentioned B+Tree node city and deletion action are brought.It comprises:
1, by metadata cache mirror image and Log backup on redundant manipulator, asynchronous write, to improve readwrite performance;
2, from root node to leaf node, use " copy-on-write " mode by when the B+Tree node city of metadata, deletion, first write with a brush dipped in Chinese ink B+Tree each more new node when writing with a brush dipped in Chinese ink, finally write with a brush dipped in Chinese ink the superblock pointing to root node again.Superblock takes up room enough little, and each pointer is in 512 bytes range, power-off at no point in the update process, can ensure superblock or record the pointer of new B+Tree, or record the pointer of old B+Tree, all have complete consistance;
3, the leaf node adopt log mode to be mirrored to opposite terminal controller the B+Tree node city of metadata, deletion action, record is newly-increased, deleted, and record B+Tree and write with a brush dipped in Chinese ink number of times (version number).When main control end controller lost efficacy, by daily record and version number, to reinsert, deletion information, recovered data.
Metadata B+Tree nodal operation, by log mode Remote Switched Port Analyzer to redundant manipulator, ensures metadata consistency.The child node pointer of the B+Tree multilayer node that metadata B+Tree nodal operation brings or leaf node content can be solved again to misplace the problem of movement, and lose problem at the metadata B+Tree nodal operation middle controller B+Tree brought that lost efficacy.
Accompanying drawing explanation
Fig. 1 is more new metadata B+Tree node schematic diagram.
Embodiment
More detailed elaboration is carried out to content of the present invention below:
By metadata cache mirror image and Log backup on redundant manipulator.
From root node to leaf node, use " copy-on-write " mode by when the B+Tree node city of metadata, deletion, first write with a brush dipped in Chinese ink B+Tree each more new node when writing with a brush dipped in Chinese ink, finally write with a brush dipped in Chinese ink the superblock pointing to root node again.
The leaf node adopt log mode to be mirrored to opposite end redundant manipulator the B+Tree node city of metadata, deletion action, record is newly-increased, deleted, and record B+Tree and write with a brush dipped in Chinese ink number of times.When main control end controller lost efficacy, by daily record and version number, to reinsert, deletion information, recovered data.
The embodiment of the present invention provides one to simplify pond metadata node to refresh method for protecting consistency, can effectively solve to simplify pond metadata B+Tree node data and write with a brush dipped in Chinese ink consistance, improves storage pool data integrity.
Below with reference to the explanation of Fig. 1 example.
The logical volume 32 of storage pool pool 0 increases block and maps, logical block is 40, then insert node (key=40) to mapping tree, first backup log is to mirrored-redundant controller (controller B), the pool that record is corresponding, lv operation and parameter (insert, key=40, value=xxx), then B+TREE is upgraded, from root node to leaf node, all newly distribute meta data block, avoid using original meta data block, and until B+TREE node is write with a brush dipped in Chinese ink just can discharge old meta data block.If when main control end controller lost efficacy, by daily record and version number, to reinsert, deletion information, recover data.
Through above operation, the operations such as B+Tree node city deletion can ensure consistency problem.
This design can the memory property of storage array, improves the reliability of simplifying storage pool operation, improves product competitive power commercially.
Claims (4)
1. simplify pond metadata node and refresh a method for protecting consistency, it is characterized in that, comprising:
By metadata cache mirror image and Log backup on redundant manipulator;
From root node to leaf node, use copy-on-write mode by when the B+Tree node city of metadata, deletion, first write with a brush dipped in Chinese ink B+Tree each more new node when writing with a brush dipped in Chinese ink, finally write with a brush dipped in Chinese ink the superblock pointing to root node again;
The leaf node adopt log mode to be mirrored to opposite end redundant manipulator the B+Tree node city of metadata, deletion action, record is newly-increased, deleted, and record B+Tree and write with a brush dipped in Chinese ink number of times; When main control end controller lost efficacy, by daily record and version number, to reinsert, deletion information, recovered data.
2. according to the requirement of right 1, it is characterized in that, also comprise:
By metadata cache mirror image and Log backup on redundant manipulator, adopt asynchronous write, to improve readwrite performance.
3. according to the requirement of right 1, it is characterized in that, also comprise:
From root node to leaf node, copy-on-write mode is used by when the B+Tree node city of metadata, deletion, all newly meta data block is distributed from root node to leaf node, avoid using original meta data block, and until B+TREE node is write with a brush dipped in Chinese ink just can discharge old meta data block.
4. according to the requirement of right 1, it is characterized in that, also comprise:
Metadata B+Tree node city, deletion action are recorded in the redundant manipulator of opposite end by log mode, comprise, the type of record operation is inserted or deletes, and is which storage pool, logical volume, data block.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410678460.XA CN104391802A (en) | 2014-11-24 | 2014-11-24 | Streamline pool metadata node refreshing consistency protecting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410678460.XA CN104391802A (en) | 2014-11-24 | 2014-11-24 | Streamline pool metadata node refreshing consistency protecting method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104391802A true CN104391802A (en) | 2015-03-04 |
Family
ID=52609709
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410678460.XA Pending CN104391802A (en) | 2014-11-24 | 2014-11-24 | Streamline pool metadata node refreshing consistency protecting method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104391802A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105718217A (en) * | 2016-01-18 | 2016-06-29 | 浪潮(北京)电子信息产业有限公司 | Method and device for maintaining data consistency of thin provisioning database |
CN109271570A (en) * | 2018-10-30 | 2019-01-25 | 郑州云海信息技术有限公司 | A kind of method of metadata management inquiry |
CN109976667A (en) * | 2016-01-26 | 2019-07-05 | 华为技术有限公司 | A kind of mirror image management method, apparatus and system |
CN110134340A (en) * | 2019-05-23 | 2019-08-16 | 苏州浪潮智能科技有限公司 | A kind of method, apparatus of metadata updates, equipment and storage medium |
CN113486399A (en) * | 2021-07-14 | 2021-10-08 | 上海瓶钵信息科技有限公司 | Data storage method and system based on RISC-V architecture |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609335A (en) * | 2012-01-12 | 2012-07-25 | 浪潮(北京)电子信息产业有限公司 | Device and method for protecting metadata by copy-on-write |
CN103150267A (en) * | 2013-02-21 | 2013-06-12 | 浪潮电子信息产业股份有限公司 | Cache synchronization method among multiple controllers |
CN103559188A (en) * | 2013-08-19 | 2014-02-05 | 曙光信息产业股份有限公司 | Metadata management method and system |
-
2014
- 2014-11-24 CN CN201410678460.XA patent/CN104391802A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609335A (en) * | 2012-01-12 | 2012-07-25 | 浪潮(北京)电子信息产业有限公司 | Device and method for protecting metadata by copy-on-write |
CN103150267A (en) * | 2013-02-21 | 2013-06-12 | 浪潮电子信息产业股份有限公司 | Cache synchronization method among multiple controllers |
CN103559188A (en) * | 2013-08-19 | 2014-02-05 | 曙光信息产业股份有限公司 | Metadata management method and system |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105718217A (en) * | 2016-01-18 | 2016-06-29 | 浪潮(北京)电子信息产业有限公司 | Method and device for maintaining data consistency of thin provisioning database |
CN105718217B (en) * | 2016-01-18 | 2018-10-30 | 浪潮(北京)电子信息产业有限公司 | A kind of method and device of simplify configuration storage pool data sign processing |
CN109976667A (en) * | 2016-01-26 | 2019-07-05 | 华为技术有限公司 | A kind of mirror image management method, apparatus and system |
CN109976667B (en) * | 2016-01-26 | 2021-09-14 | 华为技术有限公司 | Mirror image management method, device and system |
CN109271570A (en) * | 2018-10-30 | 2019-01-25 | 郑州云海信息技术有限公司 | A kind of method of metadata management inquiry |
CN110134340A (en) * | 2019-05-23 | 2019-08-16 | 苏州浪潮智能科技有限公司 | A kind of method, apparatus of metadata updates, equipment and storage medium |
CN110134340B (en) * | 2019-05-23 | 2020-03-06 | 苏州浪潮智能科技有限公司 | Method, device, equipment and storage medium for updating metadata |
CN113486399A (en) * | 2021-07-14 | 2021-10-08 | 上海瓶钵信息科技有限公司 | Data storage method and system based on RISC-V architecture |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11741048B2 (en) | Distributed write journals that support fast snapshotting for a distributed file system | |
US10795788B2 (en) | Remote data replication method and system | |
US7861049B2 (en) | Methods and apparatus for archiving digital data | |
CN100592251C (en) | Storage virtual computer system and management data snapshot image method in storage system | |
US11023318B1 (en) | System and method for fast random access erasure encoded storage | |
CN105718217B (en) | A kind of method and device of simplify configuration storage pool data sign processing | |
US8806115B1 (en) | NVRAM data organization using self-describing entities for predictable recovery after power-loss | |
US9501398B2 (en) | Persistent storage device with NVRAM for staging writes | |
US9946655B2 (en) | Storage system and storage control method | |
US11138156B2 (en) | Continuous data management system and operating method thereof | |
CN103049349B (en) | Based on Snapshot Method and the system of Bedding storage | |
CN106951375B (en) | Method and device for deleting snapshot volume in storage system | |
CN104391802A (en) | Streamline pool metadata node refreshing consistency protecting method | |
KR101569428B1 (en) | File processing method and storage device | |
CN102033793A (en) | Snapshot method and solid-state hard disk | |
US10628298B1 (en) | Resumable garbage collection | |
CN103226965B (en) | Based on the audio/video data access method of time bitmap | |
CN104462389A (en) | Method for implementing distributed file systems on basis of hierarchical storage | |
CN103164528B (en) | A kind of index establishing method of audio, video data | |
CN103617097A (en) | File recovery method and file recovery device | |
CN103473258A (en) | Cloud storage file system | |
CN102479118A (en) | Incremental backup method for data blocks | |
KR101456104B1 (en) | Method, system for dual buffering file management with non-volatile memory and mass storage device using the same | |
CN104142943A (en) | Database expansion method and database | |
CN110134551B (en) | Continuous data protection method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150304 |