Server cluster system and data replication method

Abstract

This application provides a server cluster system and a data replication method, which can be applied to the field of cluster backup technology. The server cluster system includes: a host determining the synchronous replication mode of multiple standby machines based on the mirror attributes and link information of multiple standby machines; based on the standby machine disk identifier corresponding to the data to be replicated, and according to the synchronous replication mode of the target standby machine, sending cluster information and the data to be replicated to the target standby machine in server replication mode via fiber optic channel, and sending cluster information and the data to be replicated to the target standby machine in disk replication mode via network channel and fiber optic channel respectively; any standby machine storing the data to be replicated to the target disk according to the cluster information; and, if the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate is greater than a first predetermined threshold, integrating the stored data and the data to be updated and replicated received during the polling period, and sending the integrated data to the backup server via fiber optic channel.

Description

Technical Field

[0001] This application relates to the field of cluster backup technology, and more specifically to a server cluster system and a data replication method. Background Technology

[0002] In a server cluster system consisting of multiple servers, these servers collaborate through network channels to build a redundant server cluster architecture. This ensures that the cluster continues to operate normally and communicate with application clients even when some server nodes fail. However, due to the inherent attenuation and transmission characteristics of network channels, it is necessary to utilize system software within the servers to monitor the transmitted data in real time to prevent packet loss and ensure the integrity of data backups. Summary of the Invention

[0003] In view of the above problems, this application provides a server cluster system and a data replication method.

[0004] According to a first aspect of this application, a server cluster system is provided, comprising: a host, which is communicatively connected to multiple standby machines via Fibre Channel and Network Channel, for determining a synchronous replication mode between the host and the multiple standby machines based on the mirror attributes and link information of the multiple standby machines; determining a target standby machine from the multiple standby machines based on the standby machine disk identifier corresponding to the data to be replicated, and sending cluster information and the data to be replicated to the target standby machine in server replication mode via Fibre Channel according to the synchronous replication mode of the target standby machine, and sending cluster information and the data to be replicated to the target standby machine in disk replication mode via Network Channel and Fibre Channel respectively; and multiple standby machines, each of which is used to store the data to be replicated to the target disk according to the cluster information, and, when the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate is greater than a first predetermined threshold, integrating the stored data and the data to be updated and replicated received during the polling period, and sending the integrated data to a backup server via Fibre Channel so that the backup server can back up the integrated data.

[0005] The second aspect of this application provides a data replication method, comprising: determining a synchronous replication mode between a master and multiple standby machines based on the mirror attributes and link information of multiple standby machines; determining a target standby machine from among the multiple standby machines based on the standby machine disk identifier corresponding to the data to be replicated; sending cluster information and replicated data to the target standby machine in server replication mode via fiber optic channel according to the synchronous replication mode of the target standby machine; and sending cluster information and replicated data to the target standby machine in disk replication mode via network channel and fiber optic channel respectively; so that the target standby machine stores the replicated data to the target disk according to the cluster information; and, when the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate of the target standby machine is greater than a first predetermined threshold, integrating the stored data and the replicated data to be updated received during the polling period to obtain integrated data, and sending it to the backup server via fiber optic channel.

[0006] According to embodiments of this application, the host in a server cluster system can determine the synchronization replication mode between each standby machine and the host based on the mirror attributes and link information of different standby machines. This allows for the targeted use of fiber optic channels with low loss and low data transmission risk to transmit data to be replicated to standby machines in server replication mode or disk replication mode, based on the synchronization replication mode of each standby machine. Furthermore, different channels can be used to transmit cluster information according to different modes, improving the data backup efficiency of the server cluster and reducing dependence on and pressure on network channels.

[0007] For the target standby machine that needs to store and back up the data to be replicated, with the asynchronous protection mode set as the standby protection mode, the data is integrated by taking into full account the current resource utilization rate and operating load of the target standby machine, thereby providing continuous and efficient data protection for the server cluster system through the backup server. This reduces the risk of data loss on the host and standby machines before receiving the erroneous data when the host and standby machines are in abnormal operation due to erroneous data, and provides efficient and complete historical data support for the host and standby machines to resume operation, maintaining normal operation with application clients. Attached Figure Description

[0008] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0009] Figure 1 A schematic diagram of a server cluster system according to an embodiment of this application is shown;

[0010] Figure 2 A schematic diagram of a server replication mode according to an embodiment of this application is shown;

[0011] Figure 3A schematic diagram of a disk copying mode according to an embodiment of this application is shown;

[0012] Figure 4 A schematic diagram illustrating data integration according to an embodiment of this application is shown;

[0013] Figure 5 A schematic diagram of the host protection mode and the backup protection mode according to an embodiment of this application is shown;

[0014] Figure 6 A schematic diagram of a host protection mode according to an embodiment of this application is shown;

[0015] Figure 7 A schematic diagram of an alternative protection mode according to an embodiment of this application is shown;

[0016] Figure 8 A schematic diagram of data transmission to be replicated in server replication mode according to an embodiment of this application is shown;

[0017] Figure 9 A schematic diagram illustrating protocol conversion using a bus adapter according to an embodiment of this application is shown;

[0018] Figure 10 A flowchart of a data copying method according to an embodiment of this application is shown. Detailed Implementation

[0019] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0020] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0021] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0022] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0023] In a server cluster system consisting of multiple servers, these servers collaborate through network channels to build a redundant server cluster architecture. This ensures that the cluster continues to operate normally and communicate with application clients even when some server nodes fail. However, due to the inherent attenuation and transmission characteristics of network channels, it is necessary to utilize system software within the servers to monitor the transmitted data in real time to prevent packet loss and ensure the integrity of data backups.

[0024] This application provides a server cluster system, including: a host, which communicates with multiple standby machines via fiber optic channels and network channels, and is used to determine the synchronous replication mode between the host and the multiple standby machines based on the mirror attributes and link information of the multiple standby machines; to determine a target standby machine from the multiple standby machines based on the standby machine disk identifier corresponding to the data to be replicated, and to send cluster information and the data to be replicated to the target standby machine in server replication mode via fiber optic channels according to the synchronous replication mode of the target standby machine, and to send cluster information and the data to be replicated to the target standby machine in disk replication mode via network channels and fiber optic channels respectively; multiple standby machines, any one of which is used to store the data to be replicated to the target disk according to the cluster information, and, when the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate is greater than a first predetermined threshold, integrate the stored data and the data to be updated and replicated received during the polling period, and send the integrated data to the backup server via fiber optic channels so that the backup server can back up the integrated data.

[0025] Figure 1 A schematic diagram of a server cluster system according to an embodiment of this application is shown.

[0026] like Figure 1 As shown, a server cluster system can include multiple servers capable of communicating with application clients. These servers can be divided into primary servers and multiple backup servers. The primary server and the backup servers can communicate with each other via Fibre Channel and network channels to replicate and store data, allowing any one of the backup servers to take over communication with the application client from the primary server.

[0027] Additionally, a server cluster system can include a backup server. The backup server communicates with the host and multiple standby servers via Fibre Channel to back up and protect the data on the host. When a failed host or standby server is preparing to recover, it can retrieve historical data through the backup server to ensure normal operation.

[0028] Specifically, the host can communicate with multiple standby machines via Fibre Channel and Network Channel. The host can determine the synchronization replication mode between itself and the multiple standby machines based on their mirroring attributes and link information. Based on the standby machine disk identifier corresponding to the data to be replicated, the target standby machine is selected from the multiple standby machines. Following the target standby machine's synchronization replication mode, cluster information and replication data are sent to the target standby machine in server replication mode via Fibre Channel, and cluster information and replication data are sent to the target standby machine in disk replication mode via both Network Channel and Fibre Channel.

[0029] Image attributes can include server cluster images and disk cluster images. Link information can include link latency and link length between the master and each standby machine, as well as link latency and link length between multiple standby machines.

[0030] Synchronous replication modes can include server replication mode and disk replication mode. When the standby machine's mirroring attribute is server cluster mirroring, the synchronous replication mode between the primary and standby machines can be server replication mode. When the standby machine's mirroring attribute is disk cluster mirroring, the synchronous replication mode between the primary and standby machines can be disk replication mode.

[0031] Multiple standby machines can all be in server replication mode, or all can be in disk replication mode, or some of the standby machines can be in server replication mode and others in disk replication mode.

[0032] In response to different standby machine replication modes, based on the data to be replicated generated by the current host, at least one target standby machine and its disk identifier are determined to be stored and replicated. Based on the target standby machine's replication mode, if the replication mode is server replication mode, the data to be replicated generated by the host and cluster information are synchronously sent to the target standby machine via Fibre Channel. If the replication mode is disk replication mode, the data to be replicated generated by the host is synchronously sent to the target standby machine via Fibre Channel, and the cluster information generated by the host is synchronously sent to the target standby machine via a network channel (IP network). This ensures that one or more target standby machines corresponding to the data to be replicated synchronously save the received data.

[0033] You can refer to this. Figure 1 , Figure 1The multiple standby machines shown include both those in server replication mode and those in disk replication mode. Therefore, the same fiber optic channel can be used to send data to be replicated and cluster information to the standby machines in server replication mode, and only send data to be replicated to the standby machines in disk replication mode.

[0034] Any of the multiple standby machines can be used to store the data to be replicated to the target disk according to the cluster information. When the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate is greater than the first predetermined threshold, the stored data and the data to be updated and replicated received during the polling period are integrated, and the integrated data is sent to the backup server through the fiber channel so that the backup server can back up the integrated data.

[0035] Each standby machine can include one or more disks. When any of the multiple standby machines receives cluster information and data to be replicated as the target standby machine, the target standby machine can store the data to be replicated to the target disk of the target standby machine according to the data type, storage requirements, storage disk, storage sector, and other information of the data to be replicated in the cluster information.

[0036] The host machine can determine the initiator of the current data communication with the backup server based on indicators such as the communication status between the target standby machine and the host machine, the heartbeat information of the target standby machine, and the heartbeat information of the host machine, thereby determining the asynchronous protection mode. Asynchronous protection modes can include standby protection mode and host protection mode. Standby protection mode is characterized by the standby machine acting as the initiator of data communication, sending data requiring backup protection to the backup server via Fibre Channel. Host protection mode is characterized by the host machine acting as the initiator of data communication, sending data requiring backup protection to the backup server via Fibre Channel. Data requiring secondary backup protection does not need to be sent synchronously to the backup server; instead, it can be sent asynchronously, allowing the backup server to perform simple backup protection.

[0037] If the asynchronous protection mode is determined to be the standby protection mode and the current resource utilization rate of the target standby machine is greater than the first predetermined threshold, considering the current operating load of the target standby machine, the stored data in the disk of the target standby machine can be periodically polled. If new data to be updated and copied is received during the polling period, the stored data and the data to be updated and copied can be integrated to obtain integrated data, and the integrated data can be sent to the backup server. In this way, the business data of the host can be protected without frequently sending the data that needs to be backed up to the backup server.

[0038] According to embodiments of this application, the host in a server cluster system can determine the synchronization replication mode between each standby machine and the host based on the mirror attributes and link information of different standby machines. This allows for the targeted use of fiber optic channels with low loss and low data transmission risk to transmit data to be replicated to standby machines in server replication mode or disk replication mode, based on the synchronization replication mode of each standby machine. Furthermore, different channels can be used to transmit cluster information according to different modes, improving the data backup efficiency of the server cluster and reducing dependence on and pressure on network channels.

[0039] For the target standby machine that needs to store and back up the data to be replicated, with the asynchronous protection mode set as the standby protection mode, the data is integrated by taking into full account the current resource utilization rate and operating load of the target standby machine, thereby providing continuous and efficient data protection for the server cluster system through the backup server. This reduces the risk of data loss on the host and standby machines before receiving the erroneous data when the host and standby machines are in abnormal operation due to erroneous data, and provides efficient and complete historical data support for the host and standby machines to resume operation, maintaining normal operation with application clients.

[0040] Figure 2 A schematic diagram of a server replication mode according to an embodiment of this application is shown.

[0041] like Figure 2 As shown, multiple backup machines are in server replication mode, so the master machine can transmit the data to be replicated and cluster information to any backup machine through the Fibre Channel, and use the network channel as a backup channel. In the event of an anomaly in the Fibre Channel, the network channel can be used to transmit the data to be replicated and cluster information.

[0042] Figure 3 A schematic diagram of a disk copying mode according to an embodiment of this application is shown.

[0043] like Figure 3 As shown, multiple standby machines are in disk replication mode. The host machine can include its own disk and mirror disks corresponding to multiple standby machines after mirroring. Thus, the host machine can transmit data to be replicated to any standby machine through Fibre Channel, transmit cluster information to any standby machine through Network Channel, and use Network Channel as a backup channel. In the event of failure of Fibre Channel, Network Channel can be used to transmit data to be replicated.

[0044] According to an embodiment of this application, when it is necessary to protect the data stored in the target disk of the target backup machine, any backup machine that is the target backup machine can determine the polling period of multiple sectors located in the target disk based on the priority of the target disk and the data type of the data to be copied determined from the cluster information, when the asynchronous protection mode is the standby protection mode and the resource utilization rate is greater than a first predetermined threshold.

[0045] According to an embodiment of this application, the data status of multiple sectors in the target disk is polled separately according to the polling period of multiple sectors.

[0046] According to an embodiment of this application, when data to be updated is received during the polling period of multiple sectors, the stored data in multiple sectors is updated based on the storage method of the data to be updated, thereby obtaining updated data.

[0047] According to an embodiment of this application, the updated and copied data is integrated, and the integrated data is sent to the backup server via Fibre Channel.

[0048] Each standby machine can include one or more disks, and each disk can include multiple sectors. Once the target standby machine receives cluster information and the data to be replicated, it can determine the target disk and target sector from among the multiple disks based on the cluster information.

[0049] If the current resource utilization rate of the target standby machine is determined to be greater than a first predetermined threshold, the historical modification frequency of the target sector can be obtained. Based on the priority of the target disk, a basic polling period corresponding to the target disk can be determined, serving as the basic constraint for each sector. Based on indicators such as the historical modification frequency, the priority of multiple sectors within the target disk, the data type of the data to be copied, and the weight corresponding to each indicator, a polling coefficient corresponding to each sector is determined. Then, based on the polling coefficient, the basic constraint, and a dynamic polling adjustment factor adjusted in real-time according to the changing resource utilization rate, the polling period corresponding to each sector within the target disk is jointly determined.

[0050] For example, the target standby machine may include 3 disks. The first disk is the current target disk, and the priority of the second disk is greater than the priority of the first disk, which is greater than the priority of the third disk. The first disk contains 3 sectors. Based on the priority of the first disk, and according to the overall polling period of the target standby machine (60 seconds), the basic polling period corresponding to the first disk is determined to be 20 seconds.

[0051] Then, the historical modification frequency of the first disk is obtained. The historical modification frequency, the priority of multiple sectors within the target disk, and the data type of the data to be copied are normalized to obtain frequency parameters, priority parameters for each sector, and data type parameters for the data to be copied. The parameters corresponding to each sector are multiplied by their corresponding weights to calculate the polling coefficients: 0.36 for the first sector, 0.43 for the second sector, and 0.21 for the third sector. This determines the current polling timeframes: 7.2s for the first sector, 8.6s for the second sector, and 4.2s for the third sector.

[0052] The storage methods for the data to be updated and replicated can include storage, overwriting, modification, and deletion. After determining the polling period for each sector, the data storage and modification status of multiple sectors within the target disk is polled. If, during the polling period, data to be updated and replicated is received from the host, or if the data status of already stored data changes, the data in the sector is updated based on the storage method of the data to be updated and replicated. This updated and replicated data from multiple sectors is then integrated and sent to the backup server.

[0053] According to embodiments of this application, when the asynchronous protection mode is standby protection mode and the resource utilization rate is greater than a first predetermined threshold, the target standby machine can determine the polling period for each sector in the target disk based on priority and other indicator parameters. It then polls the sectors based on different polling periods, integrates the updated and replicated data within the polling period, and sends the integrated data to the backup server. This fully considers the operating load of the target standby machine, reduces the number of communications and data size with the backup server, and while utilizing the backup server to provide continuous, efficient, and complete data protection for the server cluster system, it reduces the operating load of the standby machine and improves the operational security of the server cluster system.

[0054] Figure 4 A schematic diagram illustrating data integration according to an embodiment of this application is shown.

[0055] like Figure 4As shown, when any standby machine, acting as the target standby machine, receives data to be updated during the polling period of multiple sectors, it determines the sector identifier and storage method corresponding to the data to be updated from the updated cluster information. Based on the sector identifier corresponding to the data to be updated, at least one target sector is determined from the multiple sectors, and the stored data in the at least one target sector is updated to obtain the updated replicated data. If there are multiple target sectors, and at least two target sectors stop polling simultaneously, the updated replicated data of the two target sectors are integrated to obtain integrated data, which is then sent to the backup server.

[0056] When the target standby machine receives the data to be updated and replicated, it confirms the target disk corresponding to the data to be updated and replicated. If the target disk corresponding to the data to be updated and replicated are the same disk as the target disk in the polling period of multiple sectors, it determines the target sector to be stored and updated from multiple sectors based on the received update cluster information, and updates the data.

[0057] Since the start time of the polling period for each sector can be different, when there are multiple target sectors corresponding to the data to be updated and replicated, the polling time of each sector can be used to integrate the updated and replicated data in the target sectors that have stopped polling at the same time. For the updated and replicated data in the target sectors that have not stopped polling at the same time, the target standby machine can send it to the backup server separately after the polling stops.

[0058] For example, the target disk contains 5 sectors. During the polling period, sectors 1, 2, 4, and 5 each received and updated data at least once. The polling period for sector 1 is 10 seconds, for sector 2 it is 12 seconds, for sector 3 it is 9 seconds, for sector 4 it is 7 seconds, and for sector 5 it is 15 seconds.

[0059] The polling period for each sector starts at a different time. The first sector uses an overwrite storage method, overwriting the previously stored data with the received data to be updated (1→5). The second sector uses a storage method, overwriting the previously stored data with the received data to be updated (1→5→3). The third sector uses a storage method, storing the received data to be updated (1→1,2). The fourth sector uses a storage method that deletes the third bit of data, deleting the previously stored third bit of data in the target sector (5,3,9→5,3). The fifth sector uses an overwrite storage method, overwriting the previously stored data with the received data to be updated (10,18→10,7).

[0060] The fourth sector stops polling first, then the second and fifth sectors stop polling simultaneously, and finally the first sector stops polling. Thus, the fourth sector, which stops polling first, can send the updated copy data (5, 3) to the backup server via the fiber channel. Then, the updated copy data of the second and fifth sectors, which stop polling simultaneously, are integrated to obtain integrated data with sector identifiers (second sector 1, 2; fifth sector 10, 7), and the integrated data is sent to the backup server via the fiber channel. Finally, the first sector, which stops polling, can send the updated copy data (3) to the backup server via the fiber channel so that the backup server can store it in batches.

[0061] According to embodiments of this application, update replication data in multiple sectors with simultaneous polling stop times are integrated and then sent to the backup server. For update replication data in sectors without simultaneous polling stop times, data is sent to the backup server individually based on the polling stop time. This reduces the number of communications and data size with the backup server, lowers the operating load of the backup machine, and improves the operational security of the server cluster system.

[0062] Furthermore, during the data integration process, data from multiple disks that simultaneously stop polling can also be integrated. That is, at the same stop polling time, the updated and replicated data of all sectors in all disks within the target standby machine are integrated and then sent to the backup server.

[0063] If the resource utilization rate of the target standby machine exceeds the third predetermined threshold and the host resource utilization rate of the host exceeds the second predetermined threshold, the target standby machine can generate overload information based on its current state and notify the host through the network channel. Upon receiving the overload information from the target standby machine, the host can negotiate protection with the target standby machine. For the next data to be replicated associated with the current target standby machine, the next data to be replicated is split according to the ratio corresponding to the resource utilization rate of the current target standby machine and the host resource utilization rate. The host and the target standby machine respectively perform asynchronous transmission of the split data to be replicated with the backup server. For example, if the next data to be replicated associated with the current target standby machine is 500MB, according to the ratio corresponding to the resource utilization rate of the current target standby machine (65%) and the host resource utilization rate (38%), the next data to be replicated is split into data of 184MB (500 / (65%+38%)*65%) corresponding to the target standby machine and data of 316MB (500 / (65%+38%)*38%) corresponding to the host. After the host sends all the data to be copied to the target standby machine, the host sends 316MB of data to be copied to the backup server via Fibre Channel, and the target standby machine sends 184MB of data to be copied to the backup server via Fibre Channel. At the same time, the above data integration can be performed before the host and the target standby machine send the data to the backup server.

[0064] This can reduce the data transmission load between the primary and backup servers in many ways, and improve protection efficiency while protecting the data to be replicated.

[0065] Figure 5 A schematic diagram of the host protection mode and the backup protection mode according to an embodiment of this application is shown.

[0066] like Figure 5 As shown, in a server cluster system consisting of a host and multiple backup servers, there may be backup servers with asynchronous protection mode set to host protection mode, and there may also be backup servers with asynchronous protection mode set to standby protection mode. When it is necessary to determine the asynchronous protection mode, the host can also determine the asynchronous protection mode corresponding to the backup server based on the host resource utilization rate and the operating status of the target backup server.

[0067] If no heartbeat information is received from the target standby machine within the predetermined time period, the asynchronous protection mode is determined to be the master protection mode, the data to be replicated is sent to the backup server, and a fault log is generated based on the abnormal working status of the target standby machine.

[0068] If a heartbeat message from the target standby machine is received within a predetermined time period and the host resource utilization rate is greater than the second predetermined threshold, the asynchronous protection mode is determined to be the standby protection mode. Cluster information and data to be replicated are sent to the target standby machine in server replication mode via fiber optic channel. Cluster information and data to be replicated are also sent to the target standby machine in disk replication mode via network channel and fiber optic channel, respectively, so that the target standby machine can send the stored data to the backup server via fiber optic channel according to the resource utilization rate.

[0069] The primary and multiple backup machines can acquire heartbeat information in real time via Fibre Channel or network channel. If no heartbeat information is received from the backup machine within a predetermined period, it can be confirmed that the backup machine is in a state of temporary disconnection. In this case, the primary machine needs to directly send the data to be replicated to the backup server through Fibre Channel and generate a fault log based on the backup machine in a state of temporary disconnection, which will be sent to the external switch to prompt maintenance and repair of the backup machine.

[0070] If a heartbeat message from the standby machine is received within the predetermined time period, it can be confirmed that the standby machine is currently in normal working condition. If the host resource utilization rate is greater than the second predetermined threshold, the target standby machine selectively sends data to the backup server synchronously through the fiber optic channel based on the standby machine's resource utilization rate, so that the backup server can perform backup protection.

[0071] Alternatively, disregarding the host's resource utilization, and confirming the target standby machine's normal operation, the target standby machine can be prioritized as the data initiator between the standby machine and the backup server, directly confirming the asynchronous protection mode as the standby protection mode. When the host's resource utilization is less than or equal to a second predetermined threshold, the host can monitor the target standby machine's resource utilization in real time. If the target standby machine's resource utilization is greater than the host's, the host can temporarily send data directly to the backup server until the target standby machine's resource utilization is less than or equal to the host's. When the target standby machine's resource utilization is less than or equal to the host's, the target standby machine can send data to the backup server.

[0072] Figure 6 A schematic diagram of a host protection mode according to an embodiment of this application is shown.

[0073] like Figure 6 As shown, when multiple backup machines are in an abnormal working state, it is confirmed that all backup machines are in master protection mode, and the master machine can directly send the data to be copied to the backup server through the fiber optic channel.

[0074] Figure 7 A schematic diagram of an alternative protection mode according to an embodiment of this application is shown.

[0075] like Figure 7 As shown, when multiple backup machines are in normal working condition and the host resource utilization rate is greater than the second predetermined threshold, it is confirmed that multiple backup machines are in standby protection mode. The host can directly send the data to be copied to the backup machine through the fiber optic channel, and then the backup machine sends it to the backup server.

[0076] According to the embodiments of this application, the host comprehensively determines the asynchronous protection mode of the target standby machine based on the operating status of the standby machine and the host's own host resource occupancy status. For different asynchronous protection modes, the host determines the data initiator between the standby machine and the backup server, thereby enabling efficient data communication with the backup server.

[0077] According to embodiments of this application, the host can also be used to identify at least one first target standby in server replication mode and at least one second target standby in disk replication mode when there are multiple target standby machines, based on the synchronous replication mode of the multiple target standby machines and the standby machine disk identifier.

[0078] According to an embodiment of this application, based on the link information between at least one first target standby machine and the host machine, a portion of the data to be copied and cluster information corresponding to at least one first target standby machine are sent to at least one first target standby machine via a fiber optic channel, so that at least one first target standby machine stores a portion of the data to be copied to the corresponding target disk according to the partition address in the cluster information.

[0079] According to an embodiment of this application, based on the standby disk identifier of at least one second target standby machine, at least one target mirror disk corresponding to at least one second target standby machine is determined from the host, and the status of at least one target mirror disk is monitored.

[0080] According to an embodiment of this application, in response to at least one target mirror disk being in normal working condition, another portion of the data to be copied corresponding to at least one second target standby machine is sent to at least one second target standby machine via a fiber optic channel, cluster information corresponding to at least one second target standby machine is sent to at least one second target standby machine via a network channel, and the attribute information of at least one target mirror disk located in the host is updated.

[0081] In cases where multiple target standby machines exist in both server replication mode and disk replication mode, the multiple target standby machines are divided into at least one first target standby machine and at least one second target standby machine.

[0082] For at least one primary target standby machine in server replication mode, the host can directly send the data to be replicated and cluster information to at least one primary target standby machine via Fibre Channel.

[0083] The host's disk space contains multiple mirror disks corresponding to multiple standby machines in disk replication mode. For at least one second target standby machine in disk replication mode, the host can determine at least one target mirror disk from the multiple mirror disks based on the standby disk identifier of the at least one second target standby machine. Depending on the working status of the at least one target mirror disk, if the at least one target mirror disk is in normal working condition, the data to be replicated is transmitted via Fibre Channel, while cluster information needs to be transmitted to the at least one second target standby machine via network channel.

[0084] In disk replication mode, the communication between the standby and primary machines can refer to the initiator and target model and communication method in Fibre Channel (FC) technology. After determining the synchronous replication mode, the primary machine can first confirm the disk authorization of the standby machine in disk replication mode through the network channel. If the disk authorization is confirmed, the primary machine can retrieve the attribute information of each disk through the network channel, thereby mapping the disks of the standby machine in disk replication mode to the primary machine via FC technology. Then, the primary machine needs to use the computer system's dynamic disk or a technology such as Linux's LVM at the operating system level to form a mirrored disk group with the disks mapped from the standby machine.

[0085] For standby machines in disk replication mode, Fibre Channel cannot transmit status information and metadata of the high-availability cluster. Therefore, it is necessary to add status monitoring and corresponding anomaly handling related to Fibre Channel and disk mirroring to the cluster software of both the primary and standby machines, such as monitoring of disk software RAID (such as Windows Dynamic Disk or Linux LVM).

[0086] According to embodiments of this application, for a first target standby machine in server replication mode, the host can use a fiber optic channel with low loss and low data transmission risk to send the data to be replicated and cluster information to it. For a second target standby machine in disk replication mode, the host can also use a fiber optic channel with low loss and low data transmission risk to send the data to be replicated to it. At the same time, in conjunction with a network channel, cluster information is sent to the second target standby machine, thereby improving the data backup efficiency of the server cluster and reducing the dependence on and pressure on the network channel.

[0087] Figure 8 A schematic diagram of data transmission to be replicated in server replication mode according to an embodiment of this application is shown.

[0088] like Figure 8As shown, when there are multiple target backup machines, the host can also determine the hierarchical transmission information and transmission parameters for the multiple target backup machines based on the host resource utilization rate, the number of multiple target backup machines, the link distance between the multiple target backup machines and the host, and the link distance between the multiple target backup machines. Based on the multiple transmission parameters and hierarchical transmission information, the host determines the data transmission strategy for the multiple target backup machines and at least one data packet to be replicated corresponding to the data transmission strategy. According to the transmission order in the data transmission strategy, at least one data packet to be replicated and cluster information are sent sequentially to the multiple target backup machines through the fiber optic channel.

[0089] In the case of multiple primary target backup machines, the host can determine the hierarchical transmission information based on the constraint that the number of primary transmissions is close to the number of hierarchical transmissions, or the constraint of preset hierarchical levels, according to the number of multiple primary target backup machines.

[0090] For example, in the case of 5 primary target backup machines, the host can determine that the hierarchical transmission information is 3 primary transmissions + 2 secondary transmissions. That is, firstly, the data to be copied is sent to the first primary target backup machine of the primary transmission via the fiber optic channel. Then, while sending the data to be copied to the third primary target backup machine of the primary transmission via the fiber optic channel, the aforementioned first primary target backup machine that has already received the data to be copied can send the data to be copied to the second primary target backup machine of the secondary transmission via the fiber optic channel. And while sending the data to be copied to the fifth primary target backup machine of the primary transmission via the fiber optic channel, the aforementioned third primary target backup machine that has already received the data to be copied can send the data to be copied to the fourth primary target backup machine of the secondary transmission via the fiber optic channel.

[0091] Then, the host can normalize the host resource utilization, the link distance between the host and multiple target backup machines, the link distance between the multiple target backup machines, and the data size stored in each target backup machine to obtain resource parameters, host link parameters corresponding to each target backup machine, backup machine link parameters corresponding to each target backup machine, and data parameters corresponding to each target backup machine. Based on the product of the multiple parameters for each target backup machine, the transmission parameters corresponding to each target backup machine are determined.

[0092] Once the host has determined the transmission parameters and hierarchical transmission information corresponding to each first target standby machine, a data transmission strategy for multiple first target standby machines can be determined based on multiple transmission parameters and hierarchical transmission information. Based on the data transmission strategy, the data to be replicated corresponding to the first target standby machine is packetized, thereby ensuring that at least one data packet to be replicated matches the data transmission strategy.

[0093] For example, the transmission parameters for the first target backup machine can be 0.14, the second target backup machine can be 0.25, the third target backup machine can be 0.3, the fourth target backup machine can be 0.12, and the fifth target backup machine can be 0.19. Therefore, the data transmission strategy is determined as follows: first, send the data to be replicated and cluster information to the third target backup machine; then send the data to be replicated and cluster information to the second target backup machine; simultaneously, the third target backup machine sends the data to be replicated and cluster information to the fourth target backup machine; finally, send the data to be replicated and cluster information to the fifth target backup machine; and simultaneously, the second target backup machine sends the data to be replicated and cluster information to the first target backup machine. Thus, for the first and second target backup machines, the third and fourth target backup machines, and the fifth target backup machine, the data to be replicated is packetized, resulting in three data packets to be replicated.

[0094] After the data to be replicated is processed into packets, it can be transmitted sequentially, sending at least one data packet to be replicated and cluster information to multiple primary target backup machines through fiber optic channels.

[0095] According to an embodiment of this application, when the host sends the data to be replicated and cluster information to the first target standby machine in server replication mode, it can determine the data transmission strategy based on the number of the first target standby machines and link parameters, etc. Based on the data transmission strategy, data and information are transmitted to each first target standby machine quickly and efficiently, thereby improving the communication efficiency between the host and the standby machine, as well as the efficiency of the standby machine in data replication.

[0096] Figure 9 A schematic diagram illustrating protocol conversion using a bus adapter according to an embodiment of this application is shown.

[0097] like Figure 9 As shown, both the host and multiple backup machines can include bus adapters. The host can also be used to perform protocol conversion on another portion of the data to be copied corresponding to at least one second target backup machine in response to at least one target mirror disk being in normal working condition, using the bus adapter within the host to obtain target data to be copied conforming to the Fibre Channel protocol. Then, the target data to be copied is sent to at least one second target backup machine via Fibre Channel, so that the bus adapter within at least one second target backup machine can be used to perform protocol parsing and storage of the target data to be copied.

[0098] When using the bus adapter to perform protocol conversion on the data to be copied, verification information can be inserted to generate target data that conforms to the Fibre Channel protocol. Upon receiving the target data to be copied, the bus adapter in the second target backup machine can perform protocol parsing, then perform basic verification based on the verification information. If the verification is successful, the parsed data to be copied is stored on the target disk.

[0099] Meanwhile, the primary and backup machines can also contain cluster software, operating systems, network interface cards (NICs), and protected business application processes. The cluster software can monitor the operating system, NICs, protected business application processes, and local disks in real time to obtain the communication and operational status of various software, NICs, and links.

[0100] According to an embodiment of this application, a bus adapter adapted to Fibre Channel is used to perform protocol conversion on the transmitted data to be copied, thereby matching the transmission characteristics of Fibre Channel so as to send the data to be copied to the backup machine using Fibre Channel.

[0101] The aforementioned bus adapter can also be used in the primary target standby and primary machine in server replication mode. Furthermore, regardless of whether it's the primary target standby in server replication mode or the secondary target standby in disk replication mode, the data to be replicated can be transmitted via the network channel in the event of a Fibre Channel failure.

[0102] According to embodiments of this application, multiple standby machines can also be used to: in response to not receiving cluster information or host heartbeat information from the host within a predetermined time period, the multiple standby machines generate initial broadcast information based on their respective image attributes and link information and broadcast it through the network channel.

[0103] According to an embodiment of this application, based on multiple initial broadcast messages received, an update host and an update synchronization replication mode between the update host and multiple update standby hosts are determined from at least one standby host including server image attributes and disk image attributes. The update host is used to generate a stop operation command and send it to the host through a network channel.

[0104] The standby and primary machines can communicate bidirectionally via Fibre Channel or network channel. If more than a predetermined number of standby machines fail to receive cluster information or heartbeat information from the primary machine within a predetermined time period, each of these standby machines can broadcast the primary machine's abnormality information to other standby machines via the network channel. If multiple standby machines are aware of the primary machine's abnormality, each standby machine can generate an initial broadcast message based on its own mirror attributes and link information, and broadcast it to other standby machines via the network channel.

[0105] When each standby machine receives the initial broadcast information, the update master can be determined from at least one standby machine that has both server mirroring and disk mirroring attributes, based on the mirroring attributes of other standby machines in the initial broadcast information, and the update synchronization replication mode between the update master and multiple update standby machines can be determined.

[0106] If none of the multiple standby machines possess both server mirroring and disk mirroring attributes, i.e., if all standby machines have only a single server mirroring attribute or disk mirroring attribute, the update master can be determined from at least one standby machine that only includes server mirroring or disk mirroring attributes.

[0107] Once the update host is confirmed, it can send a stop command to the original abnormal host to prevent multiple hosts from running simultaneously, which could lead to abnormal situations such as split-brain.

[0108] According to embodiments of this application, when a primary host malfunctions, multiple backup hosts can confirm each other via network or fiber optic channels and broadcast the changes. This allows for the rapid selection of a new primary host from among the backup hosts to replace the malfunctioning primary host, improving the robustness and stability of the server cluster system. Simultaneously, after a new primary host is identified, the updating primary host sends a stop command to the malfunctioning primary host. After the malfunctioning primary host stops operating, the updating primary host takes over the work between the original primary host and the application clients, preventing a split-brain scenario in the server cluster system.

[0109] According to embodiments of this application, the multiple standby machines can also be used to: for any one of the multiple standby machines, determine at least one candidate standby machine including server mirroring attributes and disk mirroring attributes based on the mirroring attributes of the any standby machine and the mirroring attributes of other standby machines determined from multiple initial broadcast messages; determine a temporary update host for any standby machine based on the link distance between the at least one candidate standby machine and other standby machines; and generate and broadcast candidate broadcast messages based on the standby disk identifier of the temporary update host.

[0110] According to an embodiment of this application, multiple backup machines determine the backup machine that appears most frequently in the multiple candidate broadcast messages as the update master, and determine the backup machines other than the update master as multiple update backup machines.

[0111] According to an embodiment of this application, an update synchronization replication mode between the update master and the multiple update standby machines is determined based on the mirror attributes of the multiple update standby machines.

[0112] After receiving initial broadcast information from other standby machines, any standby machine determines the mirroring attributes of the other standby machines from the initial broadcast information. Based on the mirroring attributes of the other standby machines and its own mirroring attributes, at least one candidate standby machine is determined. Then, based on the link distances between the multiple standby machines determined from the initial broadcast information and each candidate host, shortest distance optimization and normalization are performed to determine the initial communication distance parameter corresponding to each candidate host. The initial communication distance parameter corresponding to each candidate host is then adjusted using the link distance between any standby machine and each candidate host to obtain multiple target communication distance parameters. The candidate standby machine corresponding to the largest target communication distance parameter is selected as the temporary update host for any standby machine.

[0113] After each standby machine has its corresponding temporary update host, candidate broadcast information is generated based on the standby machine disk identifier of the temporary update host and broadcast to other standby machines. If each standby machine receives candidate broadcast information from other standby machines, the standby machine appearing most frequently among the candidate broadcast information can be determined as the update host.

[0114] According to embodiments of this application, multiple backup machines can first determine the candidate backup machine corresponding to themselves based on link information and mirror attributes, and then the candidate backup machine that appears most frequently is selected as the update backup machine, thereby selecting the preferred update host under a relatively fair principle.

[0115] Figure 10 A flowchart of a data copying method according to an embodiment of this application is shown.

[0116] like Figure 10 As shown, the data copying method of this embodiment may include operations S1010 to S1020.

[0117] When operating S1010, the synchronization replication mode between the master and multiple standby machines is determined based on the mirror attributes and link information of multiple standby machines.

[0118] In operation S1020, based on the standby disk identifier corresponding to the data to be replicated, the target standby machine is determined from multiple standby machines. According to the synchronous replication mode of the target standby machine, cluster information and replication data are sent to the target standby machine in server replication mode via fiber optic channel, and cluster information and replication data are sent to the target standby machine in disk replication mode via network channel and fiber optic channel respectively. This allows the target standby machine to store the replicated data to the target disk according to the cluster information. If the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate of the target standby machine is greater than the first predetermined threshold, the stored data and the replication data to be updated received during the polling period are integrated to obtain integrated data, which is then sent to the backup server via fiber optic channel.

[0119] According to embodiments of this application, the synchronization replication mode between each standby machine and the primary machine is first determined based on the mirror attributes and link information of different standby machines. This facilitates the targeted sending of data to be replicated and cluster information to standby machines in different modes, improving the data backup efficiency of the server cluster and reducing dependence on and pressure on network channels. Simultaneously, standby machines can integrate data based on current resource utilization and fully consider operational load. This utilizes the backup server to reduce the risk of data loss before the primary and standby machines receive erroneous data when they are in abnormal operation due to erroneous data. It also provides an efficient and complete historical data foundation for the restored primary and standby machines, maintaining normal operation with application clients.

[0120] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0122] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

[0123] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A server cluster system, characterized in that, The system includes: The host machine communicates with multiple backup machines via fiber optic channels and network channels. It determines the synchronous replication mode between the host and the backup machines based on the mirroring attributes and link information of the backup machines. Based on the backup machine disk identifier corresponding to the data to be replicated, it identifies a target backup machine from among the backup machines. Based on the host's resource utilization and the target backup machine's operating status, it determines the asynchronous protection mode corresponding to the backup server. If no backup machine heartbeat information is received from the target backup machine within a predetermined time period, the asynchronous protection mode is determined to be the host protection mode, and the data to be replicated is sent to the backup server. If a backup machine heartbeat information is received from the target backup machine within the predetermined time period and the host's resource utilization is greater than a second predetermined threshold, the asynchronous protection mode is determined to be the standby protection mode. Following the target backup machine's synchronous replication mode, it sends cluster information and the data to be replicated to the target backup machine in server replication mode via fiber optic channels, and sends the cluster information and the data to be replicated to the target backup machine in disk replication mode via network channels and the fiber optic channels, respectively. Multiple standby machines are provided. Each standby machine is used to store the data to be replicated to the target disk according to the cluster information. When the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate is greater than a first predetermined threshold, the stored data and the data to be updated and replicated received during the polling period are integrated, and the integrated data is sent to the backup server through the fiber channel so that the backup server can back up the integrated data.

2. The system according to claim 1, characterized in that, The standby unit is also used for: When the asynchronous protection mode is the backup protection mode and the resource utilization rate is greater than the first predetermined threshold, the polling period of multiple sectors located in the target disk is determined according to the priority of the target disk and the data type of the data to be copied determined from the cluster information. Based on the polling period of the multiple sectors, the data status of the multiple sectors in the target disk is polled respectively. If the data to be updated is received during the polling period of the multiple sectors, the stored data in the multiple sectors is updated based on the storage method of the data to be updated to obtain the updated data; The updated and replicated data is integrated, and the integrated data is sent to the backup server via the fiber optic channel.

3. The system according to claim 2, characterized in that, The standby unit is also used for: If the data to be updated and replicated is received during the polling period of the multiple sectors, the sector identifier corresponding to the data to be updated and the storage method of the data to be updated and replicated are determined from the update cluster information; Based on the sector identifier corresponding to the data to be updated and copied, at least one target sector is determined from the plurality of sectors, and the stored data in the at least one target sector is updated to obtain the updated and copied data; If there are multiple target sectors and at least two target sectors stop polling simultaneously, the updated replication data of the two target sectors are integrated to obtain and send the integrated data to the backup server.

4. The system according to claim 1, characterized in that, The host is also used for: If the asynchronous protection mode is determined to be the host protection mode, a fault log is generated based on the abnormal working state of the target standby machine; If the asynchronous protection mode is determined to be the standby protection mode, the target standby machine will send the stored data to the backup server through the fiber optic channel according to the resource utilization rate.

5. The system according to claim 1, characterized in that, The host is also used for: When there are multiple target backup machines, at least one first target backup machine in the server replication mode and at least one second target backup machine in the disk replication mode are identified based on the synchronous replication mode of the multiple target backup machines and the backup machine disk identifier. Based on the link information between the at least one first target standby machine and the host, a portion of the data to be copied corresponding to the at least one first target standby machine and the cluster information are sent to the at least one first target standby machine through the fiber channel, so that the at least one first target standby machine stores the portion of the data to be copied to the corresponding target disk according to the partition address in the cluster information; Based on the standby disk identifier of the at least one second target standby machine, at least one target mirror disk corresponding to the at least one second target standby machine is determined from the host, and the status of the at least one target mirror disk is monitored. In response to the at least one target mirror disk being in normal working condition, another portion of the data to be copied corresponding to the at least one second target standby machine is sent to the at least one second target standby machine through the fiber channel, the cluster information corresponding to the at least one second target standby machine is sent to the at least one second target standby machine through the network channel, and the attribute information of the at least one target mirror disk located in the host is updated.

6. The system according to claim 5, characterized in that, The host is also used for: When there are multiple first target backup machines, the hierarchical transmission information and transmission parameters of the multiple first target backup machines are determined based on the host resource utilization rate, the number of the multiple first target backup machines, the link distance between the multiple first target backup machines and the host, and the link distance between the multiple first target backup machines. Based on multiple transmission parameters and the hierarchical transmission information, the data transmission strategy of the multiple first target backup machines and at least one data packet to be copied corresponding to the data transmission strategy are determined; According to the transmission order in the data transmission strategy, the at least one data packet to be copied and the cluster information are sequentially sent to the plurality of first target backup machines through the fiber optic channel.

7. The system according to claim 5, characterized in that, The host and the plurality of backup units all include a bus adapter, and the host is further used for: In response to the at least one target mirror disk being in normal working condition, the host uses a bus adapter to perform protocol conversion on another part of the data to be copied corresponding to the at least one second target backup machine to obtain target data to be copied that conforms to the Fibre Channel protocol. The target data to be copied is sent to the at least one second target backup machine through the fiber optic channel, so that the bus adapter in the at least one second target backup machine can be used to perform protocol parsing and storage of the target data to be copied.

8. The system according to claim 1, characterized in that, The multiple backup units are also used for: In response to not receiving the cluster information or host heartbeat information from the host within a predetermined time period, the multiple standby machines generate initial broadcast information based on their respective mirror attributes and link information and broadcast it through the network channel; Based on multiple initial broadcast messages received, from at least one standby machine including server mirror attributes and disk mirror attributes, an update master and an update synchronization replication mode between the update master and multiple update standby machines are determined. The update master is used to generate a stop operation command and send it to the master machine through the network channel.

9. The system according to claim 8, characterized in that, The multiple backup units are also used for: For any one of the plurality of backup machines: Based on the mirroring attributes of any standby machine and the mirroring attributes of other standby machines determined from the plurality of initial broadcast messages, at least one candidate standby machine is determined, including server mirroring attributes and disk mirroring attributes. Based on the link distance between the at least one candidate standby machine and other standby machines, a temporary update host for any of the standby machines is determined; Based on the standby disk identifier of the temporary update host, generate candidate broadcast information and broadcast it; The multiple backup machines, based on the received multiple candidate broadcast messages, determine the backup machine mentioned most frequently in the multiple candidate broadcast messages as the update host, and determine the backup machines other than the update host as the multiple update backup machines; Based on the mirror attributes of the multiple update standby machines, the update synchronization and replication mode between the update master and the multiple update standby machines is determined.

10. A data replication method, applied to the server cluster system according to any one of claims 1-9, characterized in that, The method includes: Based on the image attributes and link information of multiple standby machines, the synchronization replication mode between the master machine and the multiple standby machines is determined; Based on the standby disk identifier corresponding to the data to be replicated, a target standby machine is determined from the plurality of standby machines. According to the synchronous replication mode of the target standby machine, cluster information and replicated data are sent to the target standby machine in server replication mode via fiber optic channel. The cluster information and replicated data are also sent to the target standby machine in disk replication mode via network channel and the fiber optic channel, respectively. This allows the target standby machine to store the replicated data to the target disk according to the cluster information. If the asynchronous protection mode is determined to be the standby protection mode and the resource utilization rate of the target standby machine is greater than a first predetermined threshold, the stored data and the replicated data to be updated received during the polling period are integrated to obtain integrated data, which is then sent to the backup server via the fiber optic channel.

Images