Database all-in-one machine based on NVMe and distributed storage architecture

Abstract

The invention provides a database all-in-one machine based on NVMe and a distributed storage architecture, and belongs to the technical field of distributed storage, the database all-in-one machine comprises a distributed file system and a data disk, and an interface of the data disk is an NVMe interface; the distributed file system has a striping function and a mirroring function; and the distributed file system is used for combining the data disks into a disk array based on the striping function and the mirroring function, and storing data files to the disk array. According to the method, RAID0 is replaced by a striping function to improve the disk access performance, RAID1 is replaced by a mirroring function to realize a disk redundancy mechanism, namely, the hard RAID is realized by the striping function and the mirroring function, and meanwhile, various requirements of users on high performance, high data security and high reliability of the system are met; and high security and high availability of database issues are ensured.

Description

Technical Field

[0001] The present invention relates to the field of distributed storage technology, and in particular to a database all-in-one machine based on NVMe and distributed storage architecture. Background Art

[0002] The database machine achieves high system availability and data security at the database level, but for some critical business systems, there are strict requirements for fault switching time and data security, and the data redundancy mechanism based on the software level cannot fully meet business needs. Therefore, the traditional architecture of the database machine is equipped with a solid-state hard disk with a SATA3.0 interface at the hardware level, and uses a RAID array card to achieve different levels of disk redundancy and high-performance mechanisms, which improves the speed of disk data access while ensuring business data security and effective utilization of disk space.

[0003] With the continuous advancement of technology and the test of diversified market demand, the traditional SATA interface has gradually become a bottleneck that limits the performance of SSDs. In order to unleash the potential of SSDs, NVMe came into being. However, after the SATA interface is upgraded to NVMe, the performance can be greatly improved. However, since the NVMe disk is directly connected to the CPU, it cannot be managed through the RAID card, so different RAID levels cannot be created through the RAID card. At present, RAID can only be used for NVMe through soft RAID on the domestic platform. The implementation of soft RAID does not involve hardware, but is completely implemented by software imitating the disk array. The RAID logic is calculated by the CPU core. Therefore, the performance of soft RAID is affected by the operating system load and the host CPU performance. When processing large amounts of data, more CPU resources may be consumed, resulting in performance degradation. For database application scenarios with high performance requirements, soft RAID is not the best choice; in addition, soft RAID is relatively weak in data security. Since it relies on the stability and correct configuration of the operating system, once the operating system fails, RAID may lose its function. The fault recovery of soft RAID may be more complicated and require more manual intervention.

[0004] Therefore, there is an urgent need to provide a database all-in-one machine based on NVMe and distributed storage architecture to ensure high security and high availability at the storage level. Summary of the invention

[0005] In view of this, it is necessary to provide a database all-in-one machine based on NVMe and distributed storage architecture to solve the technical problem in the prior art that hard RAID cannot be implemented and different RAID levels are only created through soft RAID, resulting in low security and availability of the database all-in-one machine.

[0006] On the one hand, in order to solve the above technical problems, the present invention provides a database integrated machine based on NVMe and distributed storage architecture, including: a distributed file system and a data disk, the interface of the data disk is an NVMe interface; the distributed file system has striping function and mirroring function; The distributed file system is used to combine the data disks into a disk array based on the striping function and the mirroring function, and store data files in the disk array.

[0007] In a possible implementation, the data file includes an equally divided first data sub-file and a second data sub-file, the disk array includes a plurality of distributed disks, each of the distributed disks includes a first disk and a second disk, and the first disk and the second disk are used to store the first data sub-file and the second data sub-file, respectively.

[0008] In a possible implementation, the distributed file system includes a primary storage device and at least one backup storage device, and the primary storage device and each of the backup storage devices include the disk array.

[0009] In a possible implementation, the distributed file system includes an interface, and when the interface is called, the distributed file system is used to erase invalid data in the disk array.

[0010] In one possible implementation, when the interface is called, the distributed file system is specifically used to: obtain the logical block address of the data deleted by the database application layer, and send an ATA TRIM command and the logical block address to the disk array, and the disk array erases the invalid data in the disk array based on the ATA TRIM command and the logical block address.

[0011] In a possible implementation, the data application layer deleting data includes deleting a table space file, truncating a table space file, and deleting a table space.

[0012] In one possible implementation, the database integrated machine based on NVMe and distributed storage architecture also includes a disk compression ratio determination unit, which is used to obtain logical usage space and physical usage space based on space call instructions, and use the ratio of the logical usage space to the physical usage space as the disk compression ratio.

[0013] In a possible implementation, the database integrated machine based on NVMe and distributed storage architecture also includes a disk array management unit, which is used to expand, shrink, create, delete, monitor faults, and monitor space usage of the disk array.

[0014] In a possible implementation, the database integrated machine based on NVMe and distributed storage architecture also includes a physical disk management unit, and the physical disk management unit is used to monitor the status, usage space and data compression ratio of the data disk.

[0015] In a possible implementation, the data disk is a solid state drive.

[0016] The beneficial effects of the present invention are as follows: the database all-in-one machine based on NVMe and distributed storage architecture provided by the present invention, by setting up the database all-in-one machine to include a distributed file system with striping function and mirroring function, improves disk access performance by replacing RAID0 with striping function, and realizes disk redundancy mechanism by replacing RAID1 with mirroring function, that is, hardware RAID is realized by striping function and mirroring function, and at the same time meets the user's various needs for system high performance, high data security and high system reliability, ensuring high security and high availability of database issues. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 A schematic diagram of the structure of an embodiment of a database integrated machine based on NVMe and distributed storage architecture provided by the present invention; Figure 2 A schematic diagram of the structure of a disk array according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] It should be understood that the schematic drawings are not drawn to scale. The flowchart used in the present invention shows the operations implemented according to some embodiments of the present invention. It should be understood that the operations of the flowchart can be implemented out of order, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art, under the guidance of the content of the present invention, can add one or more other operations to the flowchart, and can also remove one or more operations from the flowchart. Some of the block diagrams shown in the accompanying drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0021] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0022] The present invention provides a database integrated machine based on NVMe and distributed storage architecture, which is described in detail below.

[0023] Figure 1 A schematic diagram of the structure of an embodiment of a database integrated machine based on NVMe and distributed storage architecture provided by the present invention is shown in FIG. Figure 1 As shown, the database integrated machine 10 based on NVMe and distributed storage architecture includes: a distributed file system 100 and a data disk 200, the interface of the data disk 200 is an NVMe interface; the distributed file system 100 has striping function and mirroring function; The distributed file system 100 is used to combine the data disks 200 into a disk array based on the striping function and the mirroring function, and store the data file 20 in the disk array.

[0024] Striping is a technique that evenly distributes data across two (or more) storage devices to improve overall performance and reduce retrieval time. Data in a disk striping setup can be stored across hard disk drives (HDDs) or solid-state drives (SSDs).

[0025] Mirroring refers to the technology whereby data is copied and written simultaneously to two different storage devices instead of being split. This feature is ideal for preventing data loss, as there is always a backup available on another disk.

[0026] Compared with the prior art, the database integrated machine 10 based on NVMe and distributed storage architecture provided in the embodiment of the present invention, by setting the database integrated machine 10 to include a distributed file system 100 with striping function and mirroring function, improves disk access performance by replacing RAID0 with striping function, and realizes disk redundancy mechanism by replacing RAID1 with mirroring function, that is: hardware RAID is realized by striping function and mirroring function, and at the same time meets the user's various needs for system high performance, high data security and high system reliability, ensuring high security and high availability of the database integrated machine.

[0027] In a specific embodiment of the present invention, the data file is equally divided into a first data sub-file and a second data sub-file through a striping function, and the first data sub-file and the second data sub-file are backed up through a mirroring function, that is, there are multiple identical first data sub-files and second data sub-files. Figure 2 As shown, the disk array includes a plurality of distributed disks, each of which includes a first disk and a second disk, and the first disk and the second disk are used to store a first data sub-file and a second data sub-file, respectively.

[0028] In a specific embodiment of the present invention, Figure 2 As shown, the data file includes eight data sub-files A, B, C, D, E, F, G, and H. The striping function divides the eight data sub-files into two parts. The first data sub-file includes A, C, E, and G, and the second data sub-file includes B, D, F, and H. The mirroring function mirrors the first data sub-file and the second data sub-file. Figure 2 Only two distributed disks are shown.

[0029] In a specific embodiment of the present invention, the distributed file system is the DAMO Distributed Storage Manager (DmASM) developed by DAMO.

[0030] To further improve the high availability and data security of the database integrated machine, in some embodiments of the present invention, the distributed file system 100 includes a primary storage device and at least one backup storage device, and the primary storage device and each backup storage device include a disk array.

[0031] In the embodiment of the present invention, the distributed file system 100 includes a primary storage device and at least one backup storage device, that is, the distributed file system 100 includes a primary and a backup cluster, which can prevent problems with the database server from affecting data normality and ensure high availability and data security of the database integrated machine.

[0032] Among them, the deployment method of the active-standby cluster is specifically as follows: automatic deployment of the active-standby cluster is realized in the distributed file system mirror environment. Based on the Ansible framework, the one-active-multiple-standby DAMO active-standby cluster can be flexibly deployed by configuring Ansible-playbook.

[0033] The specific implementation steps are as follows: 1. Deployment parameter configuration The user runs the active / standby cluster deployment parameter configuration script. This script is developed based on the Python 3 language and provides a visual window interactive filling method, allowing users to intuitively complete the operation parameter configuration required for automated deployment.

[0034] Running the script will enter the parameter filling visualization window. The parameters that need to be filled in are as follows: (1) Environmental related parameters ISO_FILE_NAME: DAMO database software iso file name HOME_DIR: DAMO database software installation directory USER_NAME: DAMO database software installation user DATA_DIR: DAMO database instance data file storage path STANDBY_NUM: Number of standby databases in the master and standby clusters MASTER_PUBLIC_IP: Public IP address of the master database STANDBY_PUBLIC_IP: Public IP address of the standby database MASTER_PRIVATE_IP: the intranet IP address of the master database cluster STANDBY_PRIVATE_IP: the private IP address of the standby cluster DW_NUM: Number of instances of the DAMO master and standby clusters SLEEP_TIME: Script execution sleep time CHECK_TIME_INTERVAL: Database log traceback time range IF_INSTALL_MONITOR_IN_YUNWEI: Whether to install a monitor in the all-in-one operation and maintenance node MONITOR_HOME_DIR: The home directory of the DAMO database software required by the monitor MONITOR_USER_NAME: The user name required by the monitor to install the DAMO database software (2) Cluster general parameters OGUID: Inter-cluster message identifier DB_NAME: database name EP_PORT_DB: Database port MAL_DW_PORT: daemon listening port MAL_INST_DW_PORT: The port where the database instance listens to the daemon process EP_PORT_DMAP: DmAP service port MAL_PORT: MAL system TCP listening port SYSTEM_TABLESPACE_SIZE: System tablespace size (M) ROLL_TABLESPACE_SIZE: Rollback tablespace size (M) MAIN_TABLESPACE_SIZE: User tablespace size (M) ONLINE_LOG_SIZE: Redo log size ONLINE_LOG_NUM: Number of Redo logs PAGE_SIZE: Database instance page size (k) EXTENT_SIZE: Database instance cluster size (number of pages) CASE_SENSITIVE: Is it case sensitive? CHARSET: database instance character set LENGTH_IN_CHAR: character length unit BLANK_PAD_MODE: Is the blank fill mode compatible with Oracle? ARCH_FILE_SIZE: Archive log size (M) ARCH_SPACE_LIMIT_SIZE: Archive log size limit (M) (3) DmASM image-specific parameters IF_ASM_MIRROR_ENV: Whether to use DmASM mirror environment ASM_MIRROR_NUM: Number of mirror copies DCRV_PATH: voting disk path DISK_NUM_OF_DCRV: Number of voting disks DG_NAME_OF_ARCH: Archive log disk group name DISK_NUM_OF_ARCH: Number of archive log disks DG_NAME_OF_DATA: Data disk group name DISK_NUM_OF_DATA: Number of data disks DG_NAME_OF_LOG: Redo log disk group name DISK_NUM_OF_LOG: Number of redo log disks EP_MAL_PORT: DmASM storage service MAL system listening port EP_PORT_CSS: CSS service listening port EP_PORT_ASM: ASM service listening port EP_SHM_KEY: ASM shared memory key CHECK_PORT: The port to check whether the instance is active ASM_RESTART_INTERVAL: ASM service restart interval (s) DB_RESTART_INTERVAL: Database service restart interval (s) AUTO_OPEN_CHECK: Instance timeout detection time (s) DATA_STRIPING: Data file striping granularity LOG_STRIPING: Log file striping granularity Click "save" to save the configured parameters to the parameter file path, and click "reset" to reconfigure the above parameters.

[0035] 2. Active / standby cluster deployment phase This automated deployment script uses the ansible-playbook module to split the DAMO master-slave cluster deployment into 35 yml script tasks. The method includes: (1) Use the group module of Ansible to create a database installation group; (2) Use the user module of Ansible to create a database installation user; (3) Use the file module of Ansible to create the directories required for cluster installation; (4) Use the copy module of Ansible to transfer the DAMO database iso file; (5) Use the script module of Ansible to generate the silent installation configuration file of the DAMO database; (6) Use the script module of Ansible to install the DAMO database on the primary and standby nodes; (7) Use the file module of Ansible to create a database instance configuration file directory; (8) If the current deployment environment uses the DmASM image, use the template module of Ansible to generate DmASM to initialize the ASM disk and create a system disk group execution script for each cluster node; (9) If the current deployment environment uses the DmASM image, use the template module of Ansible to generate the configuration file dmdcr_cfg.ini for formatting the dcrv disk, the MAL system configuration file dmasvrmal.ini used by ASM, and the input parameter configuration file dmdcr.ini used by the DmASM system tool; (10) If the current deployment environment uses the DmASM image, use the ansible shell module to generate the CSS and ASM service startup scripts; (11) If the current deployment environment uses the DmASM image, use the shell module of Ansible to start the ASM service of each node in the active and standby clusters; and use the wait_for module of Ansible to wait for the ASM service to start; (12) If the current deployment environment uses the DmASM image, use the ansible shell module to create data, archive, and redo log disk groups for each node in the cluster; (13) Use the template module of Ansible to generate the database instance initialization configuration file dminit.ini for each node in the cluster; (14) Use the ansible shell module to initialize the database instance on each node of the cluster; (15) Use the ansible shell module to generate database service startup scripts on each node in the cluster; (16) Use the script module of Ansible to turn on the archiving switch of the main library, and use the template module of Ansible to generate the archiving configuration file dmarch.ini for the main library; (17) Use the ansible shell module to start the main database service; (18) If it is not the first cluster on the integrated machine, use the Ansible tempalte module to generate the DmAP service port configuration file dmap.ini for the main database, use the Ansible script to modify the DmAP port value in the main database instance configuration file, and modify the port reading path in the main database DmAP service startup script; (19) Use the script module of ansible to start the main library DmAP service; (20) Use the script module of Ansible to back up the main database online; (21) Use the copy module of Ansible to transfer the backup set to each backup database node; (22) Use the script module of Ansible to restore the backup set of the main database on each backup database node; (23) Use the script module of Ansible to modify the parameter values ​​in the database instance configuration file of the master and standby database nodes; (24) Use the script module of Ansible to add the real-time archiving information of the backup database to the archiving configuration file of the main database; (25) Use the template module of Ansible to generate the archive configuration file dmarch.ini on the standby node; (26) Use the template module of Ansible to generate database instances on each node of the primary and standby databases using the MAL system configuration file dmmal.ini, the daemon configuration file dmawatcher.ini, and the monitor configuration file dmmonitor.ini; (27) Use the script module of Ansible to shut down the main database service and check whether the main database service has been shut down normally; (28) Use the shell module of Ansible to start the database service of the master and standby nodes in mount mode, and use the wait_dor module of Ansible to wait for the database service to start; (29) Use the script module of Ansible to set the inter-cluster message identifier of the primary and standby databases and set the database role; (30) Use the shell module of Ansible to generate a daemon service startup script for the primary and standby database nodes; (31) Use the Ansible shell module to start the daemon service of the primary and standby database nodes respectively; (32) If it is not the first cluster on the integrated machine, use the Ansible tempalte module to generate the DmAP service port configuration file dmap.ini for the standby database, use the Ansible script to modify the DmAP port value in the standby database instance configuration file, and modify the port reading path in the standby DmAP service startup script; (33) Use the script module of Ansible to start the DmAP service of the standby node.

[0036] At this point, the automated deployment of the master-slave cluster of the DAMO database all-in-one machine is completed.

[0037] In actual applications, if the distributed file system 100 does not support the TRIM mechanism, a large-scale write penalty will be generated when the disk is almost full, affecting the read and write performance of the database and reducing the service life of the data disk. In addition, the write speed limit mechanism will be triggered in advance, restricting the read and write performance of the database. In severe cases, it will cause the hard disk to crash and the system to freeze. To solve this technical problem, in some embodiments of the present invention, the distributed file system 100 includes an interface. When the interface is called, the distributed file system 100 is used to erase invalid data in the disk array.

[0038] That is, when the interface is called, the TRIM mechanism is automatically triggered in the embodiment of the present invention to reduce the occurrence of write amplification, avoid the hidden danger of disk overwrite caused by the long-term operation of the business system, reduce the impact on the upper-level database io performance, and extend the service life of the disk.

[0039] Among them, the TRIM mechanism is the ATA-8 instruction of the solid state drive (SSD). Its existence enables the file system to send a notification to the SSD to record the invalid pages generated during the modification or deletion process, and then uniformly recycle and erase all invalid pages at a certain interval, and erase and update the blocks where the invalid pages are located. On the one hand, enough space is reserved to avoid write amplification due to insufficient space. On the other hand, using TRIM, invalid pages are recycled and erased when IO is idle, avoiding the controller from moving these invalid data blocks, and these invalid data blocks are cleared in the GC (garbage collection) operation, thereby reducing the number of SSD particle erases and extending the service life of the solid state drive.

[0040] In a specific embodiment of the present invention, the specific process of the distributed file system 100 erasing invalid data in the disk array is as follows: The database application layer obtains the logical block address (LBA) of the deleted data, and sends the ATA TRIM command and the logical block address to the disk array. The disk array erases the invalid data in the disk array based on the ATA TRIM command and the logical block address.

[0041] The embodiment of the present invention can clear invalid data to be deleted in advance, so that the used capacity of the disk is closer to the actual value, and the disk compression ratio becomes visible, calculable, and monitorable, making it possible to effectively monitor the disk usage rate and accurately avoid the disk false overwrite problem.

[0042] In a specific embodiment of the present invention, deleting data at the data application layer includes deleting tablespace files, resizing tablespace files, and deleting tablespaces.

[0043] From the above description, it can be seen that one of the purposes of adding the TRIM mechanism is to make the disk compression ratio calculable. Therefore, in some embodiments of the present invention, for example Figure 1 As shown, the database integrated machine 10 based on NVMe and distributed storage architecture also includes a disk compression ratio determination unit 300, which is used to obtain logical usage space and physical usage space based on space call instructions, and uses the ratio of logical usage space to physical usage space as the disk compression ratio.

[0044] In a specific embodiment of the present invention, the space call instruction includes a logical use space call instruction and a physical use space call instruction.

[0045] Specifically: The logical space usage calling instruction is: SELECT TOTAL_MB - FREE_MB FROM V$ASMGROUP, unit: MB; The physical space usage call instruction is: sfx-status / dev / nvme2n1 Provisioned Capacity- Free Physical Space, unit: GB; or smartctl --all Namespace 1 Utilization in / dev / nvme2n1, unit: Bytes.

[0046] In order to provide enterprise-level user operation and maintenance personnel with a convenient and intelligent database integrated machine cloud pattern management method, so as to effectively improve the work efficiency and operation and maintenance costs of the operation and maintenance personnel, thereby ensuring a more stable and efficient operation of the database integrated machine, in some embodiments of the present invention, such as Figure 1 As shown, the database integrated machine 10 based on NVMe and distributed storage architecture also includes a disk array management unit 400 and a physical disk management unit 500; The disk array management unit 400 is used to expand, shrink, create, delete, monitor faults and monitor space usage of the disk array; The physical disk management unit 500 is used to monitor the status, used space and data compression ratio of the data disk.

[0047] The embodiment of the present invention implements the management of disk arrays and data disks by setting up a disk array management unit 400 and a physical disk management unit 500, covering multiple operation and maintenance management functions including disk expansion and disk compression ratio monitoring, which can reduce the operation and maintenance costs of enterprise personnel and improve the operation and maintenance management efficiency of database application services.

[0048] In a specific embodiment of the present invention, the data disk 200 is a solid state disk.

[0049] By configuring the data disk 200 as a solid state disk, the embodiment of the present invention can expand the storage capacity of the data integrated machine by utilizing the hardware compression function of the solid state disk, thereby effectively reducing the enterprise-level use cost.

[0050] To sum up, the database integrated machine based on NVMe and distributed storage architecture proposed in the embodiment of the present invention realizes the RAID10-level disk high-performance redundancy mechanism at the file system level through the mirror striping function; improves the logical available capacity of the physical disk by virtue of the disk hardware compression function, and makes the disk usage visible and closer to the actual value with the help of the trim mechanism of the distributed file system, thereby reducing the risk of disk overwrite and extending the service life of the disk; provides a background management system with multiple operation and maintenance functions including disk expansion and disk compression ratio monitoring, so as to realize user-friendly control of the details such as whole machine operation and maintenance and fault handling of the integrated machine.

[0051] The above is a detailed introduction to a database integrated machine based on NVMe and distributed storage architecture provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A database integrated machine based on NVMe and distributed storage architecture, characterized in that: include: A distributed file system and a data disk, wherein the interface of the data disk is an NVMe interface; the distributed file system has striping function and mirroring function; The distributed file system is used to combine the data disks into a disk array based on the striping function and the mirroring function, and store data files in the disk array.

2. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The data file includes a first data sub-file and a second data sub-file that are equally divided. The disk array includes a plurality of distributed disks. Each of the distributed disks includes a first disk and a second disk. The first disk and the second disk are used to store the first data sub-file and the second data sub-file, respectively.

3. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The distributed file system includes a primary storage device and at least one backup storage device, and the primary storage device and each backup storage device include the disk array.

4. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The distributed file system includes an interface. When the interface is called, the distributed file system is used to erase invalid data in the disk array.

5. The database integrated machine based on NVMe and distributed storage architecture according to claim 4, characterized in that: When the interface is called, the distributed file system is specifically used to: obtain the logical block address of the data deleted by the database application layer, and send an ATA TRIM command and the logical block address to the disk array, and the disk array erases the invalid data in the disk array based on the ATATRIM command and the logical block address.

6. The database integrated machine based on NVMe and distributed storage architecture according to claim 5, characterized in that: The data application layer deletes data including deletion of table space files, truncation of table space files and deletion of table space.

7. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The database integrated machine based on NVMe and distributed storage architecture also includes a disk compression ratio determination unit, which is used to obtain logical usage space and physical usage space based on a space call instruction, and use the ratio of the logical usage space to the physical usage space as the disk compression ratio.

8. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The database integrated machine based on NVMe and distributed storage architecture also includes a disk array management unit, which is used to expand, shrink, create, delete, monitor faults and monitor space usage of the disk array.

9. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The database integrated machine based on NVMe and distributed storage architecture also includes a physical disk management unit, which is used to monitor the status, usage space and data compression ratio of the data disk.

10. The database integrated machine based on NVMe and distributed storage architecture according to claim 1, characterized in that: The data disk is a solid state drive.

Images