Data synchronization methods, systems, electronic devices and media for distributed clusters

Abstract

This invention discloses a data synchronization method, system, electronic device, and medium for a distributed cluster. The data synchronization method includes: detecting whether all nodes have the same expected update time for the next incremental data update at the current moment; the expected update time is determined based on the moment when the data source performed an incremental data update in the previous moment; for nodes with an expected update time, using the expected update time as the determination time for the node to perform an incremental data update in the next moment; when the current moment reaches the determination time, controlling each node to obtain the latest incremental data from the data source to update the historical data of each node. This invention uses the update time of the data source's next incremental update as the determination time for each node's next data update, ensuring that the update time of each node remains consistent, guaranteeing consistent data updates across all nodes in the distributed cluster, and reducing latency between nodes and the data source.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a data synchronization method, system, electronic device, and medium for a distributed cluster. Background Technology

[0002] In a distributed cluster, each node processes business logic indiscriminately, making it easy to scale horizontally to handle increased traffic. Therefore, it is widely used for data processing on the internet. Because each node processes business requests indiscriminately and the distribution of these requests is random, it is necessary to ensure that the data on each node in a masterless distributed cluster is updated at approximately the same time, so that users receive consistent data results when accessing applications within the cluster.

[0003] However, the timing of updates initiated by different nodes is random, and different nodes do not communicate with each other, resulting in inconsistent data update times for different nodes. This leads to inconsistent data output by different nodes within a certain time period, which may cause some nodes to return old data and others to return new data when users access applications within the cluster. This can not only cause failures in the entire distributed cluster system, but also affect the user experience. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the defect of inconsistent data update time points of different nodes in the prior art, which leads to data inconsistency within the nodes, and to provide a data synchronization method, system, electronic device and medium for distributed clusters.

[0005] The present invention solves the above-mentioned technical problems through the following technical solution:

[0006] According to a first aspect of the present invention, a data synchronization method for a distributed cluster is provided, the distributed cluster including a data source and multiple nodes, the data source being communicatively connected to each of the nodes, the data synchronization method comprising:

[0007] Detect whether all the nodes have the same expected update time for the next incremental data update at the current moment; the expected update time is determined based on the moment when the data source performed the incremental data update in the previous moment.

[0008] For the node with the expected update time, the expected update time is used as the determination time for the node to perform incremental data update at the next moment;

[0009] When the current time is detected to have reached the determination time, each node is controlled to obtain the latest incremental data from the data source to update the historical data of each node.

[0010] Preferably, the data synchronization method further includes:

[0011] For the nodes that do not have the expected update time, obtain the historical update time of the previous incremental data update for each node;

[0012] The determination time for incremental data updates of the node is based on the historical update time and configuration period.

[0013] Preferably, the data synchronization method further includes:

[0014] If the current time has not reached the determination time, each node is controlled to enter a waiting state.

[0015] Preferably, the data synchronization method further includes:

[0016] When the current time is detected to have reached the determination time, each node is also controlled to obtain the latest update time and incremental data update cycle from the data source;

[0017] The expected update time for the next incremental data update is calculated based on the latest update time and the incremental data update cycle.

[0018] According to a second aspect of the present invention, a data synchronization system for a distributed cluster is provided, the distributed cluster including a data source and multiple nodes, the data source being communicatively connected to each of the nodes, and the data synchronization system including a judgment module, a first determination module, and a first control module.

[0019] The judgment module is used to detect whether all the nodes have the same expected update time for the next incremental data update at the current time; the expected update time is determined based on the time when the data source performed the incremental data update in the previous moment.

[0020] The first determining module is used to determine the next moment for incremental data update of the node for the node that has the expected update time.

[0021] The first control module is used to detect when the current time reaches the determination time, and control each node to obtain the latest incremental data from the data source to update the historical data of each node.

[0022] Preferably, the data synchronization system further includes an acquisition module and a second determination module:

[0023] The acquisition module is used to acquire the historical update time of each node that performed incremental data update at the previous moment for nodes that do not have the expected update time.

[0024] The second determining module is used to determine the determination time for incremental data update of the node at the next moment based on the historical update time and configuration period.

[0025] Preferably, the data synchronization system further includes a second control module:

[0026] The second control module is used to control each node to be in a waiting state when it detects that the current time has not reached the determination time.

[0027] Preferably, the data synchronization system further includes a third control module and a computing module:

[0028] The third control module is used to detect when the current time reaches the determination time, and also to control each node to obtain the latest update time and incremental data update cycle from the data source;

[0029] The calculation module is used to calculate the expected update time of the next incremental data update based on the latest update time and the incremental data update cycle.

[0030] According to a third aspect of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and for running on the processor, wherein the processor executes the computer program to implement the distributed cluster data synchronization method of the present invention.

[0031] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the data synchronization method of the distributed cluster of the present invention.

[0032] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of the present invention.

[0033] The positive and progressive effects of this invention are as follows:

[0034] Within a distributed cluster, each node communicates with the data source, using the next incremental update time from the data source as the determining moment for each node's next data update. This ensures that each node retrieves the latest incremental data from the data source at the same time and updates its historical data accordingly, guaranteeing consistent update times across all nodes in the masterless distributed cluster. Furthermore, more timely data updates reduce latency between nodes and the data source, ensuring relative synchronization between the data source and each node. Consequently, when users access different nodes within the distributed cluster, they not only receive the same data results but also relatively up-to-date data, improving the user experience. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating the data synchronization method for a distributed cluster according to Embodiment 1 of the present invention.

[0036] Figure 2 This is a schematic diagram of the structure of a distributed cluster according to Embodiment 1 of the present invention.

[0037] Figure 3 This is a schematic diagram of a data synchronization process of a distributed cluster according to Embodiment 1 of the present invention.

[0038] Figure 4 This is a schematic diagram of the structure of the distributed cluster data synchronization system of Embodiment 2 of the present invention.

[0039] Figure 5 This is a schematic diagram of the electronic device according to Embodiment 3 of the present invention. Detailed Implementation

[0040] The present invention will be further illustrated by way of embodiments below, but the present invention is not limited to the scope of the embodiments described herein.

[0041] Example 1

[0042] This embodiment provides a data synchronization method for a distributed cluster, such as... Figure 1 As shown, this data synchronization method is applied to a distributed cluster, which includes a data source and multiple nodes, such as... Figure 2 As shown, the data source communicates with each node individually, but the nodes do not communicate with each other. As an optional implementation, nodes can be application machines with no difference in role or code logic. Data synchronization refers to each node retrieving data from the data source and caching it in its local memory, ensuring that the cached data remains consistent across nodes. In this embodiment, the distributed cluster can include one or multiple data sources, with each node communicating with only one data source. The data on each node connected to each data source needs to be synchronized. Furthermore, it should be noted that the nodes involved in this embodiment are running, normal nodes. Nodes that are not ready or are in a stopped state will not provide services externally as available nodes.

[0043] In this embodiment, the data source is a medium that provides data access, such as a database, a third-party service interface, and Elasticsearch (ES, a data engine). As an optional implementation, the data source itself can support periodically fetching incremental data for nodes to read. Alternatively, if the data source is not periodically updated, a middleware layer can be built between the data source and the nodes. This middleware collects incremental data from the data source and then periodically packages it into incremental data packets for nodes to read. This middleware can be software programs or computing engines with pre-defined scheduled tasks.

[0044] See Figure 1 The data synchronization method includes the following steps:

[0045] S11. Check if all nodes have the same expected update time for the next incremental data update at the current moment. If yes, proceed to step S12; otherwise, proceed to step S13.

[0046] The expected update time is determined based on the time when the data source last performed an incremental data update. For example, assuming the data source's update cycle is one minute and the last incremental data update occurred at 10:00, the expected update time is 10:01. In this embodiment, when a node reads incremental data from the data source, it knows the timestamp and cycle of the last update. Based on the data source's update time and cycle, it calculates the expected update time for the next update. Since all nodes connect to the same data source, the expected update times for all nodes are the same, ensuring that each node performs incremental updates at the same expected update time.

[0047] As an optional implementation, one situation where a node does not have the expected update time at a certain moment is when the node has just started and has not yet had time to read the data source, so it cannot read the last update time and update cycle of the data source, and naturally there is no expected update time. As another optional implementation, another situation is that the node has been running for a long time, but at a certain moment the node has not been able to read the update time of the data source normally. This situation will also cause the node to not have the expected update time at a certain moment.

[0048] S12. For nodes with an expected update time, the expected update time is used as the determination time for incremental data update of the node at the next moment.

[0049] In this embodiment, the determination time refers to determining whether a node should enter the incremental data update process based on whether the current time has reached the determination time.

[0050] S13. For nodes that do not have an expected update time, obtain the historical update time of the previous incremental data update for each node.

[0051] S14. Determine the timing for incremental data updates at the next node based on historical update times and configuration cycles.

[0052] In this embodiment, the configuration period refers to the update period configured for each node. The configuration periods for different nodes can be the same or different. When a node does not have an expected update time, it is allowed to automatically follow its own period to perform the incremental data update process, thereby avoiding the node's historical data not being updated for a long time, and avoiding the sudden increase in the amount of data updated later, which would increase the burden on the node, etc.

[0053] In this embodiment, if the current time has not reached the determination time, step S15 is executed; if the current time has reached the determination time, steps S16, S17 and S18 can be executed.

[0054] As an optional implementation, for newly started nodes, the above-mentioned determination process is not performed. Instead, an incremental data update is performed directly at the moment of startup to read the last update time and update cycle of the data source. At the next moment after startup, the node can have the expected update time of the next incremental data update, so that nodes with different startup times can keep their data synchronized as soon as possible.

[0055] S15. When it is detected that the current time has not reached the determination time, control each node to be in a waiting state.

[0056] As an optional implementation, for nodes that do not have an expected update time, step S11 can be repeated after a short period of time. For example, if the node reads the last update time of the data source normally at the next moment, the expected update time can be calculated, so that the node can have the same expected update time as other nodes. Then the time can be determined to be changed to the expected update time.

[0057] S16. When the current time is detected to have reached the judgment time, control each node to obtain the latest incremental data from the data source in order to update the historical data of each node.

[0058] Newly generated or changed data is called incremental data, and the latest incremental data is the incremental data from the most recent update of the data source. As an optional implementation, incremental data updates can be performed based on the update timestamp of the data source. In the data source, each row of data typically generates an update timestamp at the time of its most recent creation or change, and the required incremental data is extracted based on this timestamp. Alternatively, incremental data updates can be performed based on the update log, which includes a series of update records such as additions, modifications, and deletions. Operations are then performed on each update record in the update log to obtain the incremental data after the data source update and before the update.

[0059] S17. When the current time is detected to have reached the judgment time, each node is also controlled to obtain the latest update time and incremental data update cycle from the data source.

[0060] In this embodiment, the incremental data update cycle is the data source update cycle, and the latest update time is the time when the data source last performed an update. When it is detected that the current time has reached the determination time, steps S16 and S17 can be performed simultaneously.

[0061] S18. Calculate the expected update time for the next incremental data update based on the latest update time and the incremental data update cycle.

[0062] As an optional implementation, the calculated expected update time is saved locally on each node as the expected update time for the next incremental data update of the node.

[0063] The following example illustrates the data synchronization process of this implementation. (See also:) Figure 3 In one possible implementation: Assume that the data value with id (unique identifier) ​​1 in the data source, node A, and node B is x. The data source updates the data value with id 1 from x to y at second 0 (e.g., 10:00), and the update time of the data source is 60 seconds. At second 3, node A starts, thus directly and actively obtaining the latest incremental data from the data source, and updates the data value with id 1 in node A from x to y. Since the latest update time of the data source is read at second 0, and the incremental data update cycle is 60 seconds, the expected update time is obtained as second 60 (i.e., 10:01). At second 5, node B starts, and similarly updates the data value with id 1 in node B from x to y, and the expected update time is also calculated to be second 60. At second 60, the data source updates the data value with id 1 from y to z. At this time, node A and node B simultaneously perform incremental data updates, and calculate the same expected update time as second 120. And so on. After this, node A and node B can perform incremental data updates at the same expected update time.

[0064] like Figure 3 As shown, apart from a brief difference in the data output by nodes A and B during the initial 3 and 5 seconds after startup, the data on different nodes is basically updated synchronously. Furthermore, the time delay between each node obtaining incremental data from the data source is almost negligible, ensuring that the data on each node is more up-to-date. Without this data synchronization process, the timing of incremental data updates by nodes A and B, which started at different times, might have remained staggered indefinitely.

[0065] Within a distributed cluster, each node communicates with the data source, using the next incremental update time from the data source as the determining moment for each node's next data update. This ensures that each node retrieves the latest incremental data from the data source at the same time and updates its historical data accordingly, guaranteeing consistent update times across all nodes in the masterless distributed cluster. Furthermore, more timely data updates reduce latency between nodes and the data source, ensuring relative synchronization between the data source and each node. Consequently, when users access different nodes within the distributed cluster, they not only receive the same data results but also relatively up-to-date data, improving the user experience.

[0066] Example 2

[0067] This embodiment provides a data synchronization system for a distributed cluster. This system is applied to a distributed cluster, which includes a data source and multiple nodes, such as... Figure 2 As shown, the data source communicates with each node individually, but the nodes do not communicate with each other. As an optional implementation, nodes can be application machines with no difference in role or code logic. Data synchronization refers to each node retrieving data from the data source and caching it in its local memory, ensuring that the cached data remains consistent across nodes. In this embodiment, the distributed cluster can include one or multiple data sources, with each node communicating with only one data source. The data on each node connected to each data source needs to be synchronized. Furthermore, it should be noted that the nodes involved in this embodiment are running, normal nodes. Nodes that are not ready or are in a stopped state will not provide services externally as available nodes.

[0068] In this embodiment, the data source is a medium that provides data access, such as a database, a third-party service interface, and Elasticsearch. As an optional implementation, the data source itself can support periodically pulling incremental data for nodes to read; as another optional implementation, if the data source is not updated periodically, a middleware layer can be built between the data source and the nodes. The middleware collects the incremental data from the data source and then periodically packages it into incremental data packets for nodes to read. This middleware can be a software program or computing engine with pre-set scheduled tasks.

[0069] See Figure 4 The data synchronization system includes a judgment module 21, a first determination module 22, an acquisition module 23, a second determination module 24, a second control module 25, a first control module 26, a third control module 27, and a calculation module 28.

[0070] The judgment module 21 is used to detect whether all nodes have the same expected update time for the next incremental data update at the current moment. If so, the first determination module 22 is called; otherwise, the acquisition module 23 is called. The expected update time is determined based on the time when the data source performed the incremental data update in the previous moment. For example, assuming the update cycle of the data source is one minute and the time when the data source performed the incremental data update in the previous moment was 10:00, then the expected update time is 10:01.

[0071] In this embodiment, when a node reads incremental data from a data source, it can know the timestamp and period of the last update of the data source. Based on the update time and update period of the data source, it can calculate the expected update time of the next data source. Since each node is connected to the same data source, the expected update time of each node is the same, so that each node performs incremental updates at the same expected update time.

[0072] As an optional implementation, one situation where a node does not have the expected update time at a certain moment is when the node has just started and has not yet had time to read the data source, so it cannot read the last update time and update cycle of the data source, and naturally there is no expected update time. As another optional implementation, another situation is that the node has been running for a long time, but at a certain moment the node has not been able to read the update time of the data source normally. This situation will also cause the node to not have the expected update time at a certain moment.

[0073] The first determining module 22 is used to determine whether a node will perform incremental data updates for nodes with expected update times. In this embodiment, the determining time refers to determining whether a node should enter the incremental data update process based on whether the current time has reached the determining time.

[0074] The acquisition module 23 is used to acquire the historical update time of each node that performed incremental data update at the previous moment for nodes that do not have an expected update time.

[0075] The second determining module 24 is used to determine the timing for incremental data updates of a node based on historical update times and configuration periods. In this embodiment, the configuration period refers to the update period configured for each node; different nodes may have the same or different configuration periods. When a node does not have an expected update time, it automatically follows its own period to perform the incremental data update process, thereby avoiding the node's historical data not being updated for a long time and the resulting large amount of data being updated suddenly later, which would increase the node's burden, etc.

[0076] In this implementation, if the current time has not reached the determination time, the second control module 25 is invoked; if the current time has reached the determination time, the first control module 26, the third control module 27, and the calculation module 28 can be invoked.

[0077] As an optional implementation, for newly started nodes, the above-mentioned determination process is not performed. Instead, an incremental data update is performed directly at the moment of startup to read the last update time and update cycle of the data source. At the next moment after startup, the node can have the expected update time of the next incremental data update, so that nodes with different startup times can keep their data synchronized as soon as possible.

[0078] The second control module 25 is used to control each node to be in a waiting state when it detects that the current time has not reached the determination time. As an optional implementation, for nodes that do not have an expected update time, the second control module 25 can make them repeat step S11 after waiting for a short period of time. For example, assuming that the node can read the last update time of the data source normally in the next time, the expected update time can be calculated, so that the node can have the same expected update time as other nodes. Then the determination time can be changed to the expected update time.

[0079] The first control module 26 is used to control each node to obtain the latest incremental data from the data source when the current time reaches the determination time, so as to update the historical data of each node. Among them, the newly generated or changed data is the incremental data, and the latest incremental data is the incremental data when the data source last performed an update.

[0080] As an optional implementation, the first control module 26 can perform incremental data updates based on the update timestamp of the data source. In the data source, each row of data usually generates an update timestamp at the time of the most recent creation or change of data. The required incremental data is extracted based on this update timestamp. As another optional implementation, the first control module 26 can also perform incremental data updates based on the update log. The update log includes a series of update records such as additions, modifications, and deletions. Then, operations are performed on each update record in the update log to obtain the incremental data after the data source is updated and before the update.

[0081] The third control module 27 is used to control each node to obtain the latest update time and incremental data update cycle from the data source when the current time is detected to have reached the determination time. In this embodiment, the incremental data update cycle is the update cycle of the data source, and the latest update time is the time when the data source last performed an update. When the current time is detected to have reached the determination time, the first control module 26 and the third control module 27 can be called synchronously.

[0082] The calculation module 28 is used to calculate the expected update time of the next incremental data update based on the latest update time and the incremental data update cycle. As an optional implementation, the calculation module 28 saves the calculated expected update time to the local storage of each node as the expected update time of the node's next incremental data update.

[0083] Example 3

[0084] This embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the program, it implements the data synchronization method of the distributed cluster in Embodiment 1.

[0085] like Figure 5 The electronic device 30 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0086] Electronic device 30 may be in the form of a general-purpose computing device, such as a server device. Components of electronic device 30 may include, but are not limited to: at least one processor 31, at least one memory 32, and a bus 33 connecting different system components (including memory 32 and processor 31).

[0087] Bus 33 includes a data bus, an address bus, and a control bus.

[0088] The memory 32 may include volatile memory, such as random access memory (RAM) 321 and cache memory 322, and may further include read-only memory (ROM) 323.

[0089] The memory 32 may also include a program tool 325 having a set (at least one) of program modules 324, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0090] The processor 31 executes various functional applications and data processing by running computer programs stored in the memory 32, such as the data synchronization method of the distributed cluster in Embodiment 1 of the present invention.

[0091] Electronic device 30 can also communicate with one or more external devices 34. This communication can be performed via input / output (I / O) interface 35. Furthermore, model generation device 30 can also communicate with one or more networks via network adapter 36. Figure 5 As shown, network adapter 36 communicates with other modules of the model-generating device 30 via bus 33. It should be understood that, although... Figure 5 Unless otherwise specified, the device 30 generated in conjunction with the model may use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0092] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0093] Example 4

[0094] This embodiment provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the data synchronization method of the distributed cluster in Embodiment 1.

[0095] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0096] In an alternative embodiment, the present invention can also be implemented as a program product comprising program code, wherein when the program product is run on a terminal device, the program code is used to cause the terminal device to perform the steps of implementing the data synchronization method of the distributed cluster of Embodiment 1.

[0097] The program code for executing the present invention can be written in any combination of one or more programming languages. The program code can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0098] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but all such changes and modifications fall within the scope of protection of the present invention.

Claims

1. A data synchronization method for a distributed cluster, characterized in that, The distributed cluster includes a data source and multiple nodes, wherein the data source is communicatively connected to each of the nodes, and the data synchronization method includes: Detect whether all the nodes have the same expected update time for the next incremental data update at the current moment; the expected update time is determined based on the moment when the data source performed the incremental data update in the previous moment. For the node with the expected update time, the expected update time is used as the determination time for the node to perform incremental data update at the next moment; When the current time is detected to have reached the determination time, each node is controlled to obtain the latest incremental data from the data source to update the historical data of each node; The data synchronization method further includes: For the nodes that do not have the expected update time, obtain the historical update time of the previous incremental data update for each node; The determination time for incremental data update of the node is determined based on the historical update time and configuration period; The data synchronization method further includes: When the current time is detected to have reached the determination time, each node is also controlled to obtain the latest update time and incremental data update cycle from the data source; The expected update time for the next incremental data update is calculated based on the latest update time and the incremental data update cycle.

2. The data synchronization method for a distributed cluster as described in claim 1, characterized in that, The data synchronization method further includes: If the current time has not reached the determination time, each node is controlled to enter a waiting state.

3. A distributed cluster data synchronization system, characterized in that, The distributed cluster includes a data source and multiple nodes, with the data source communicating with each node. The data synchronization system includes a judgment module, a first determination module, and a first control module. The judgment module is used to detect whether all the nodes have the same expected update time for the next incremental data update at the current time; the expected update time is determined based on the time when the data source performed the incremental data update in the previous moment. The first determining module is used to determine the next moment for incremental data update of the node for the node that has the expected update time. The first control module is used to detect when the current time reaches the determination time, and control each node to obtain the latest incremental data from the data source to update the historical data of each node; The data synchronization system further includes an acquisition module and a second determination module: The acquisition module is used to acquire the historical update time of each node that performed incremental data update at the previous moment for nodes that do not have the expected update time. The second determining module is used to determine the determination time for incremental data update of the node at the next moment based on the historical update time and the configuration period; The data synchronization system also includes a third control module and a computing module: The third control module is used to detect when the current time reaches the determination time, and also to control each node to obtain the latest update time and incremental data update cycle from the data source; The calculation module is used to calculate the expected update time of the next incremental data update based on the latest update time and the incremental data update cycle.

4. The distributed cluster data synchronization system as described in claim 3, characterized in that, The data synchronization system also includes a second control module: The second control module is used to control each node to be in a waiting state when it detects that the current time has not reached the determination time.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the data synchronization method for a distributed cluster as described in any one of claims 1-2.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the data synchronization method for a distributed cluster as described in any one of claims 1-2.

Images