A method and system for constructing a highly reliable heterogeneous storage resource pool

Abstract

The present application provides a method and system for constructing a highly reliable heterogeneous storage resource pool. The method for constructing the highly reliable heterogeneous storage resource pool includes the following steps: virtualizing heterogeneous storage devices; expanding the heterogeneous storage space; optimizing the data reading and writing of the heterogeneous storage pool; and designing the reliability of the heterogeneous storage pool. In the above technical solution, online expansion of the storage resource pool of heterogeneous storage devices is achieved, while keeping the user upper-layer application interface unchanged, and the convenience and reliability of constructing the storage resource pool are realized.

Description

Background Art

[0002] There are mainly two types of existing storage devices, namely disk array storage and distributed cluster storage. During the storage usage process, as user data increases, there will be a need for capacity expansion. Expanding the original storage is the simplest and quickest solution. Disk array storage expands by vertically extending disk enclosures, and distributed cluster storage expands by horizontally adding nodes. There are certain limitations to the expansion of both types of storage. Disk array expansion increases the number of disks while maintaining the original storage processor. Due to limited storage heads, bottlenecks are extremely likely to occur. Although distributed cluster storage systems claim to support a large number of nodes and large capacities, expansion is difficult due to reasons such as hardware discontinuation or software compatibility. When the original storage cannot be expanded, expansion can only be achieved by purchasing new storage devices. However, it is difficult to integrate the storage space of the original storage device with that of the existing storage device, and a unified storage space cannot be provided to users. Upper-layer applications must identify the old and new storage types and develop corresponding usage interfaces according to the storage, causing inconvenience to users. Summary of the Invention

[0003] This application provides a method and system for constructing a highly reliable heterogeneous storage resource pool to achieve online expansion of the storage resources of heterogeneous storage devices while keeping the upper-layer application interfaces of users unchanged, realizing the convenience and reliability of constructing the storage resource pool.

[0004] In the first aspect, a method for constructing a highly reliable heterogeneous storage resource pool is provided, including the following steps:

[0005] Virtualize heterogeneous storage devices;

[0006] Expand heterogeneous storage spaces;

[0007] Optimize data reading and writing of the heterogeneous storage pool;

[0008] Design the reliability of the heterogeneous storage pool.

[0009] In the above technical solution, by virtualizing heterogeneous storage devices; expanding heterogeneous storage spaces; optimizing data reading and writing of the heterogeneous storage pool; designing the reliability of the heterogeneous storage pool; online expansion of the storage resources of heterogeneous storage devices is achieved while keeping the upper-layer application interfaces of users unchanged, realizing the convenience and reliability of constructing the storage resource pool.

[0010] In a specific feasible implementation, the step of virtualizing heterogeneous storage devices specifically includes:

[0011] Unify the management of heterogeneous storage devices;

[0012] Perform multipath mapping on the virtualized block domain.

[0013] In a specific feasible implementation, heterogeneous storage devices are uniformly managed through the iSCSI protocol.

[0014] In a specific feasible implementation, the embedded expansion method strategy and the combined expansion method strategy are used to expand the heterogeneous storage space.

[0015] In a specific feasible implementation, the embedded expansion method strategy embeds the virtualized storage device as a whole into the virtualized block domain corresponding to the existing file system to increase the capacity of the virtualized block storage pool.

[0016] In a specific feasible implementation, the combined expansion method strategy takes the virtualized storage device as one or more independent virtualized block domains, constructs one or more virtualized block storage pools correspondingly, and adds the virtualized block storage pool to the file system as an additional data pool.

[0017] In a specific feasible implementation, the method for optimizing the data reading and writing of the heterogeneous storage pool includes designing a dynamic sub-tree partitioning distribution method for the file system, a data writing load balancing method, and a striping dynamic distribution optimization method.

[0018] In a specific feasible implementation, a reliability design method for the metadata management node is adopted to design the reliability of the heterogeneous storage pool.

[0019] In a specific feasible implementation, the reliability design method for the metadata management node includes:

[0020] Designing a negotiation network protocol between multiple nodes;

[0021] Designing the broadcast message format of multiple nodes;

[0022] Designing the main node management strategy;

[0023] Designing the primary and standby node switching strategy.

[0024] In a second aspect, a highly reliable heterogeneous storage resource pool construction system is provided, including:

[0025] A virtualization module for virtualizing heterogeneous storage devices;

[0026] An expansion module for expanding the heterogeneous storage space;

[0027] A read / write optimization module for optimizing the data reading and writing of the heterogeneous storage pool;

[0028] A reliability design module for designing the reliability of the heterogeneous storage pool.

[0029] In the above technical solution, by virtualizing heterogeneous storage devices, expanding heterogeneous storage spaces, optimizing the data reading and writing of heterogeneous storage pools, and designing the reliability of heterogeneous storage pools, the online expansion of the storage resource pool of heterogeneous storage devices is achieved. At the same time, the upper-layer application interfaces of users remain unchanged, and the convenience and reliability of constructing the storage resource pool are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flowchart of a method for constructing a highly reliable heterogeneous storage resource pool provided by an embodiment of the present application;

[0031] Figure 2 It is a structural block diagram of a system for constructing a highly reliable heterogeneous storage resource pool provided by an embodiment of the present application;

[0032] Figure 3 It is a schematic diagram of the virtualization design of heterogeneous storage devices provided by an embodiment of the present application;

[0033] Figure 4 It is a schematic diagram of a method for expanding the storage space of a file system provided by an embodiment of the present application;

[0034] Figure 5 It is a schematic diagram of the principle of DNS load balancing provided by an embodiment of the present application;

[0035] Figure 6 It is a schematic diagram of the message format of multi-node negotiation provided by an embodiment of the present application;

[0036] Figure 7 It is a schematic diagram of the state machine of the metadata management master node provided by an embodiment of the present application;

[0037] Figure 8 It is a schematic diagram of the state machine of the metadata management standby node provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The present application will be further described in detail below with reference to the drawings and embodiments. Through these descriptions, the features and advantages of the present application will become more clearly defined.

[0039] The special term "exemplary" here means "serving as an example, embodiment, or illustration". Any embodiment described as "exemplary" here does not have to be construed as superior to or better than other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0040] In addition, the technical features involved in different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0041] To facilitate the understanding of the method and system for constructing a highly reliable heterogeneous storage resource pool provided by the embodiments of the present application, the application scenario thereof will be described first. The method and system for constructing a highly reliable heterogeneous storage resource pool provided by the embodiments of the present application are used to realize the online expansion of the storage resource pool of heterogeneous storage devices, while keeping the upper-layer application interfaces of users unchanged, realizing the convenience and reliability of constructing the storage resource pool. There are mainly two types of existing storage devices, namely disk array storage and distributed cluster storage. During the storage usage process, as user data increases, there will be a demand for capacity expansion, and expanding the original storage is the simplest and quickest solution. Disk array storage expands by vertically expanding disk enclosures, and distributed cluster storage expands by horizontally adding nodes. There are certain limitations in the expansion of both types of storage. Disk array expansion increases the number of disks while maintaining the original storage processor. Due to the limited storage head, bottlenecks are likely to occur easily. Although the distributed cluster storage system claims to support a large number of nodes and large capacity, expansion is difficult due to reasons such as hardware discontinuation or software compatibility. In the case where the original storage cannot be expanded, it is only possible to expand by purchasing new storage devices. However, it is difficult to integrate the storage space of the original storage device with that of the existing storage device, and a unified storage space cannot be provided to users. The upper-layer application must identify the old and new storage and develop corresponding usage interfaces according to the storage, causing inconvenience to users. Therefore, the embodiments of the present application provide a method and system for constructing a highly reliable heterogeneous storage resource pool to realize the online expansion of the storage resource pool of heterogeneous storage devices, while keeping the upper-layer application interfaces of users unchanged, realizing the convenience and reliability of constructing the storage resource pool. The following will be described in detail with specific drawings by way of embodiments.

[0042] Reference Figures 1 to 8 , Figure 1 is the flowchart of the method for constructing a highly reliable heterogeneous storage resource pool provided by the embodiments of the present application; Figure 2 is the structural block diagram of the system for constructing a highly reliable heterogeneous storage resource pool provided by the embodiments of the present application; Figure 3 is the schematic diagram of the virtualization design of heterogeneous storage devices provided by the embodiments of the present application; Figure 4 is the schematic diagram of the method for expanding the storage space of the file system provided by the embodiments of the present application; Figure 5 is the schematic diagram of the DNS load balancing principle provided by the embodiments of the present application; Figure 6 is the schematic diagram of the multi-node negotiation message format provided by the embodiments of the present application; Figure 7 is the schematic diagram of the main node state machine of metadata management provided by the embodiments of the present application; Figure 8 is the schematic diagram of the standby node state machine of metadata management provided by the embodiments of the present application.

[0043] In Figure 1 , the embodiments of the present application provide a method for constructing a highly reliable heterogeneous storage resource pool, including the following steps:

[0044] Virtualize heterogeneous storage devices;

[0045] Expand the capacity of heterogeneous storage spaces;

[0046] Optimize the data reading and writing of heterogeneous storage pools;

[0047] Design the reliability of heterogeneous storage pools.

[0048] In the above technical solution, by virtualizing heterogeneous storage devices; expanding the capacity of heterogeneous storage spaces; optimizing the data reading and writing of heterogeneous storage pools; designing the reliability of heterogeneous storage pools; the online expansion of the storage resource pool of heterogeneous storage devices is realized, and at the same time, the upper-layer application interface of the user remains unchanged, realizing the convenience and reliability of the construction of the storage resource pool.

[0049] In a specific feasible implementation scheme, the steps of virtualizing heterogeneous storage devices specifically include:

[0050] Unify the management of heterogeneous storage devices;

[0051] Perform multipath mapping on the virtualized block domain.

[0052] In a specific feasible implementation scheme, heterogeneous storage devices are unifiedly managed through the iSCSI protocol.

[0053] In a specific feasible implementation scheme, the embedded expansion method strategy and the combined expansion method strategy are used to expand the capacity of heterogeneous storage spaces.

[0054] In a specific feasible implementation scheme, the embedded expansion method strategy integrally embeds the virtualized storage device into the virtualized block domain corresponding to the existing file system to increase the capacity of the virtualized block storage pool.

[0055] In a specific feasible implementation scheme, the combined expansion method strategy regards the virtualized storage device as one or more independent virtualized block domains, correspondingly constructs one or more virtualized block storage pools, and adds the virtualized block storage pool to the file system as an additional data pool.

[0056] In a specific feasible implementation scheme, the methods for optimizing the data reading and writing of heterogeneous storage pools include designing a dynamic sub-tree partitioning distribution method for the file system, a data writing load balancing method, and a striping dynamic distribution optimization method.

[0057] In a specific feasible implementation scheme, the reliability of heterogeneous storage pools is designed by using a reliability design method for metadata management nodes.

[0058] In a specific feasible implementation scheme, the reliability design method for metadata management nodes includes:

[0059] Design the negotiation network protocol between multiple nodes;

[0060] Design the broadcast message format for multiple nodes;

[0061] Design the master node management strategy;

[0062] Design the master - standby node switching strategy.

[0063] Refer to Figures 3 to 8 , the high - reliability heterogeneous storage resource pool construction method specifically includes:

[0064] I. Virtualization design of heterogeneous storage devices

[0065] Refer to Figure 3 , design the virtualization method of heterogeneous storage devices. Through the standard iSCSI protocol structure, multiple sets of heterogeneous disk arrays and distributed cluster storage devices are uniformly managed. The heterogeneous storage devices are taken over locally through the iSCSI protocol. All heterogeneous storage spaces form the heterogeneous layer of the system, and one or more virtualized block domain storage pools are constructed. Through the sharding technology, the larger storage space is first sliced into 1G - sized data blocks. The sliced data blocks can be used in different storage pools, and all data blocks form the sharding layer of the system. Through the redundancy technology, the sharded storage is formed into single - copy, multi - copy or erasure code according to the user's redundancy requirements. The file - layer virtualization method slices the unified storage space into multiple file systems according to different user capacity requirements to provide file services for users. The storage space of the already managed storage devices is uniformly managed, uniformly scheduled, and uniformly allocated for use to construct a shared distributed file system storage resource pool.

[0066] The storage device contains multiple nodes. Multiple disks are generated by mapping the storage space through distributed nodes. The same storage space mapped by multiple paths is virtualized into a virtualized block domain using multi - path management, which is convenient for direct access without caring about the status of each link. Each path can map a SCSI hard disk and each has a SCSI address, which consists of initiator ID, bus ID, target ID, and LUN ID. In actual networking, the initiator ID generally corresponds to the host HBA port, the bus ID is generally fixed at 0 in the SAN - mapped hard disk, the target ID generally corresponds to the storage array controller port, and the LUN ID represents the sequence number of the LUN in the target. The multi - path software borrows the WWN of the SCSI device to distinguish whether the hard disks mapped by different paths are the same LUN. When the WWNs are the same, it means that the SCSI devices mapped by the two paths are the same LUN.

[0067] II. Design for expanding heterogeneous storage space

[0068] Reference Figure 4 , two expansion methods, namely embedded expansion and combined expansion, for expanding the shared distributed storage resource pool are proposed. Both methods keep the NFS mount directory on the user side unchanged, and from the user's perspective, the available storage space under the mounted directory is increased.

[0069] The embedded expansion method integrally embeds the virtualized storage device into the virtualized block domain corresponding to the existing file system to increase the capacity of the virtualized block storage pool, and then increase the storage capacity of the distributed file system. During the expansion process, based on the real-time running status of each set of distributed cluster storage devices and the pre-set policy parameters, data migration is dynamically planned and executed to transfer some data from the storage device with higher load to the device with lower load, so as to ensure the optimal distribution of data in the storage system and improve the overall performance and efficiency at the same time.

[0070] The combined expansion method takes the virtualized storage device as one or more independent virtualized block domains, constructs one or more virtualized block storage pools correspondingly, and adds the virtualized block storage pool as an additional data pool to the file system to expand the storage capacity. After the virtualized block storage pool is constructed, activate the virtualized block storage pool that needs to be used according to the demand to make it active, so as to perform the expansion operation in units of resource pools. And those resource pools that are not currently in the active state can be set to the sleep mode, and they do not participate in the system's rebalancing calculation in this state to save resources. However, once there is a data access demand pointing to the data in these sleeping resource pools, the system can respond quickly. By means of efficient compression processing technology, while quickly providing data access, it also significantly reduces the overall storage space consumption.

[0071] III. Heterogeneous Storage Pool Data Read / Write Optimization Design

[0072] Reference Figure 5, a dynamic subtree partitioning and distribution algorithm for a RADOS-based file system is proposed. Based on the dynamic subtree partitioning method, in order for the file system to achieve optimal scalability and performance, metadata and business data are separated and uniformly stored in the RADOS layer. First, the data is split into objects of the same size, and each object has a unique ID. Then, the ID of the object is mapped to storage nodes using a hashing algorithm to ensure that the data is evenly distributed and stored on the correct nodes. To further improve data security and fault tolerance, each object is replicated on multiple nodes to achieve redundant backup of the data. The placement and replication of data are managed through a scalable storage allocation algorithm to ensure the efficient distribution and access of data in the cluster. The metadata and information of the entire cluster are maintained through a Cluster Map, including the status of nodes, the distribution of data, and the location of replicas, etc., to ensure the consistency and stability of the cluster. By monitoring the load of cluster nodes in real time, the subtree distribution is dynamically adjusted among different nodes. And the data distribution can be dynamically adjusted according to the load. Thus, the load balance of all nodes on the file system is guaranteed.

[0073] A method for allocating data write requests based on DNS load balancing is proposed. When the file system is successfully mounted on all nodes, in order to ensure load balance among nodes, DNS load balancing is designed as the initial allocation method for data write requests. Multiple IP addresses are configured for the same host name in the DNS server. When answering DNS queries, the DNS server returns different resolution results in sequence according to the IP addresses of the host records in the DNS file for each query, guiding the client's access to different machines, so that different clients access different servers. This mechanism enables each node to have an equal opportunity to process write requests, effectively avoiding the overload of a single node and improving the stability and response speed of the overall system.

[0074] A striping dynamic distribution optimization method is proposed. To further optimize the striping effect, the number of stripes to be allocated on each disk is dynamically calculated and adjusted according to the actual performance parameters of the disks (such as I / O speed, latency), the remaining storage space size, and the current load condition. When the number of stripes is large, the data is more likely to be written to the disks with more stripes by adjusting the weight factor. This dynamic adjustment mechanism ensures that the data can be evenly and efficiently distributed on all available disks, thus avoiding the situation where some disks are overloaded while others are idle due to data skew. When a large number of data stripes are detected on a certain disk, according to the automatic monitoring of the number of stripes on each disk, the new write requests are allocated to the disks with relatively fewer stripes currently by adjusting the weight factor to balance the load of each disk. This method maintains the balanced distribution of data among disks, can maximize the storage and processing capabilities of each disk, further improve the overall performance and resource utilization rate of the system, and enhance the parallel access capability and throughput of the storage pool.

[0075] IV. Reliability Design of Heterogeneous Storage Pool

[0076] Reference Figures 6 to 8 , a reliability design method for metadata management nodes is proposed. The metadata management node is used to store the metadata of the storage pool. The metadata management nodes adopt a primary and standby design, and multi-frame interconnection is achieved between the nodes through 10 Gigabit Ethernet. Each node is configured with a single ID and the corresponding frame IP address. Since multi-node negotiation needs to be supported, the negotiation between multiple nodes no longer uses the unicast method, but uses multicast messages for negotiation. The multicast address is 224.0.0.1, and the multicast port defaults to UDP port 30002 (configurable). Any node defaults to the primary node when joining the node cluster, and only switches to the standby node when the primary node cannot be found. In the normal running state, the standby node does not send messages, and the primary node broadcasts messages once per second. The multi-node broadcast message format is as Figure 6 shown, where:

[0077] ID: The ID configured for the current node, which is a unique value in the node cluster, and different nodes cannot be configured with the same ID.

[0078] role: The status of the current node, including the management primary node (1: HE_PRIMARY), the management standby frame (2: HE_STANDBY).

[0079] ip_num: The number of IP addresses included in the message, arranged in the form of an array in ip_address.

[0080] ip_address: The management IP address (binary format) of the current node, which can include multiple IP addresses.

[0081] The metadata management master node is responsible for the configuration of the entire machine. All configuration tasks are uniformly managed by the master node, and the configured configuration information is sent to all standby nodes. The state machine of the master node is as follows Figure 7 As shown, under normal circumstances, the status of the current node is broadcast once per second to prevent other nodes from preempting as the master node. When receiving a message from another node indicating it as the master node, arbitration is performed by comparing the node IDs. The box with the smaller node ID is the new management master node.

[0082] The metadata management standby node forwards the management message to the master node for management. Under normal circumstances, it does not send multi-frame negotiation multicast messages. Only when it tries to preempt as the master node when it cannot receive the master node message will it send a message. The state machine of the standby node is as follows Figure 6 As shown:

[0083] (1) The standby node checks every second whether it has received the broadcast message sent by the master node in the last 3 seconds. If it can normally receive the broadcast message from the master node, it only performs the check action; if it cannot receive the broadcast message from the master node, it executes step 2.

[0084] (2) Check whether it has received a message from a standby node with an ID smaller than its own in the last 3 seconds. If it can receive it, it sleeps for 1 second and then executes step 1 again; otherwise, it executes step 3.

[0085] (3) Check whether it has received the broadcast message sent by the master node in the last 6 seconds. If it can receive it, this frame sends a broadcast message, sleeps for 1 second, and then executes step 1; otherwise, it executes step 4.

[0086] (4) Since the master node message cannot be received within 6 seconds and the ID of this frame is the smallest in the normal operation frames, this node switches to the master node.

[0087] In this embodiment, a method for constructing a highly reliable heterogeneous storage resource pool is provided, which realizes the fusion use of the file system storage spaces of multiple sets of heterogeneous disk arrays and distributed cluster storage devices, enabling users to achieve reliable, efficient, shared read and write access to the distributed file system storage space without changing the user-side usage strategy.

[0088] In the above technical solution, a unified storage resource pool is constructed for heterogeneous storage devices, and a method for constructing a highly reliable heterogeneous storage resource pool is provided; by using the method of this application, under the condition that the existing storage devices remain in the in-use state unchanged, there is no need to stop the storage services in use by users, realizing the online expansion of the heterogeneous storage device storage resource pool, and at the same time keeping the user upper-layer application interface unchanged, achieving the convenience and reliability of the construction of the storage resource pool.

[0089] In Figure 2 , this application embodiment provides a highly reliable heterogeneous storage resource pool construction system, including:

[0090] A virtualization module for virtualizing heterogeneous storage devices;

[0091] An expansion module for expanding the heterogeneous storage space;

[0092] A read / write optimization module for optimizing the data read / write of the heterogeneous storage pool;

[0093] A reliability design module for designing the reliability of the heterogeneous storage pool.

[0094] In this embodiment, a system is provided that can integrate multiple heterogeneous storage devices and provide a unified storage resource pool. Whether it is expanding the original storage device or adding new storage devices, the user-side usage interface remains unchanged, achieving fast and imperceptible expansion and solving the inconvenience of data access after user expansion.

[0095] In the above technical solution, by virtualizing heterogeneous storage devices; expanding the heterogeneous storage space; optimizing the data read / write of the heterogeneous storage pool; designing the reliability of the heterogeneous storage pool; online expansion of the storage resource pool of heterogeneous storage devices is achieved, while keeping the user upper-layer application interface unchanged, realizing the convenience and reliability of the construction of the storage resource pool.

[0096] Those skilled in the art of the present technology know that the present application can be implemented as a system, a method, or a computer program product.

[0097] Therefore, the present disclosure can be specifically implemented in the following forms: it can be completely hardware, can also be completely software (including firmware, resident software, microcode, etc.), or can also be a combination of hardware and software, generally referred to as "circuit", "module" or "system" in this article. In addition, in some embodiments, the present application can also be implemented in the form of a computer program product in one or more computer-readable media, which contain computer-readable program codes.

[0098] Any combination of one or more computer-readable media may be employed. The computer-readable media may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example - but not limited to - an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present document, a computer-readable storage medium may be any tangible medium that contains or stores a program which can be used by or in connection with an instruction execution system, apparatus, or device.

[0099] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application. On this basis, various substitutions and improvements can be made to the present application, and all of these fall within the protection scope of the present application.

Claims

1. A method for constructing a highly reliable heterogeneous storage resource pool, characterized in that, Including the following steps: Virtualize heterogeneous storage devices; Expand the capacity of heterogeneous storage spaces; Use the embedded expansion method strategy and the combined expansion method strategy to expand the capacity of heterogeneous storage spaces; The embedded expansion method strategy embeds the virtualized storage devices as a whole into the virtualized block domain corresponding to the existing file system, increasing the capacity of the virtualized block storage pool; The combined expansion method strategy takes the virtualized storage devices as one or more independent virtualized block domains, correspondingly constructs one or more virtualized block storage pools, and adds the virtualized block storage pool as an additional data pool to the file system; Optimize the data reading and writing of heterogeneous storage pools; Design the reliability of heterogeneous storage pools.

2. The method for constructing a highly reliable heterogeneous storage resource pool according to claim 1, wherein The steps for virtualizing heterogeneous storage devices specifically include: Unify the management of heterogeneous storage devices; Perform multipath mapping on the virtualized block domain.

3. The method for constructing a highly reliable heterogeneous storage resource pool according to claim 2, wherein Unify the management of heterogeneous storage devices through the iSCSI protocol.

4. The method for constructing a highly reliable heterogeneous storage resource pool according to claim 3, wherein The methods for optimizing the data reading and writing of heterogeneous storage pools include designing a dynamic sub-tree partitioning distribution method for the file system, a data writing load balancing method, and a striped dynamic distribution optimization method.

5. The method for constructing a highly reliable heterogeneous storage resource pool according to claim 4, wherein Design the reliability of heterogeneous storage pools using the reliability design method for metadata management nodes.

6. The method for constructing a highly reliable heterogeneous storage resource pool according to claim 5, wherein The reliability design method for metadata management nodes includes: Design the negotiation network protocol between multiple nodes; Design the broadcast message format for multiple nodes; Design the master node management strategy; Design the master-slave node switching strategy.

7. A high-reliability heterogeneous storage resource pool construction system, characterized in that Including: A virtualization module for virtualizing heterogeneous storage devices; An expansion module for expanding the capacity of heterogeneous storage spaces; Use the embedded expansion method strategy and the combined expansion method strategy to expand the capacity of heterogeneous storage spaces; The embedded expansion method strategy embeds the virtualized storage devices as a whole into the virtualized block domain corresponding to the existing file system, increasing the capacity of the virtualized block storage pool; The combined expansion method strategy takes the virtualized storage devices as one or more independent virtualized block domains, correspondingly constructs one or more virtualized block storage pools, and adds the virtualized block storage pool as an additional data pool to the file system; A read-write optimization module for optimizing the data reading and writing of heterogeneous storage pools; A reliability design module for designing the reliability of heterogeneous storage pools.

Images