CN116455730A - A method, system and controller for active/standby switching - Google Patents

Abstract

The embodiment of the invention discloses a master-slave switching method, a system and a controller, which relate to the field of data communication. The method comprises the following steps: the controller detects the abnormality of any proxy gateway and obtains a first bare metal server node taking any proxy gateway as a main proxy gateway according to the corresponding relation table; and the controller selects a new main proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to a preset rule, so that the new main proxy gateway processes the message sent by the first bare metal server node.

Description

A method, system and controller for active/standby switching

technical field

The invention relates to the field of data communication, in particular to a master-standby switching method, system and controller.

Background technique

With the continuous increase of data center bandwidth and the rise of cloud computing, the requirements for server network processing capabilities are getting higher and higher. At the same time, with complex technologies such as VXLAN (Virtual Extensible Local Area Network, virtual scalable local Network, software-defined network) promotion, need to spend more server CPU (Central Processing Unit, central processing unit) for processing. In this context, the smart network card came into being. After the smart network card is introduced on the server side, some business capabilities of the server can be offloaded to the smart network card for execution, which greatly improves the overall data processing capability of the server and frees up the CPU. computing power.

In the application process of smart network cards, for scenarios where smart network cards that do not support network offloading or standard network When the intelligent network card is uninstalled, the OVS (Open Virtual Switch, open virtual switch) component needs to run on an agent node other than the server for specialized network service processing. However, the reliability of these proxy nodes and how to restore the business after the proxy node fails are technical problems that need to be solved urgently in practical applications.

Contents of the invention

In order to solve the above-mentioned technical problems, the embodiment of the present invention provides a master/standby switchover method, system and controller, which are used to solve the technical problems of insufficient reliability of the proxy node and how to quickly restore the service after the proxy node is abnormal.

In the first aspect, the embodiment of the present invention provides an active/standby switchover method, which is applied to an active/standby switchover system, and the system includes a controller, at least two agent gateways, and at least one bare metal server node, and the controller Communicatively connected to the at least two proxy gateways, the at least two proxy gateways are communicatively connected to the at least one bare metal server node, and the master proxy gateway corresponding to the at least one bare metal server node is stored in the controller and the corresponding relationship table of the standby proxy gateway, the method includes:

The controller detects that any proxy gateway is abnormal, and obtains the first bare metal server node with any proxy gateway as the primary proxy gateway according to the correspondence table;

The controller elects a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway Processing the message sent by the first bare metal server node.

In the active/standby switching method provided by the embodiment of the present invention, the bare metal server node is not aware of the existence of the proxy gateway, and when the primary proxy gateway is abnormal, the bare metal server node can forward the business traffic to the new primary proxy gateway without awareness , so as to quickly restore customer business.

In a possible implementation manner, the controller elects a new master proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, including :

The controller obtains the number of IP addresses of the bare metal server nodes processed by each standby proxy gateway corresponding to the first bare metal server node, and selects the standby proxy gateway with the least number of IP addresses of the processed bare metal server nodes as the new main proxy gateway;

And/or, the controller acquires the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and selects the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway.

In a possible implementation manner, before the controller detects that any of the proxy gateways is abnormal, the method further includes:

The controller configures a master proxy gateway and a backup proxy gateway for each bare metal server node, and stores the correspondence between each bare metal server node and the master proxy gateway and backup proxy gateway in the local correspondence table; The controller configures the same virtual IP address and different virtual MAC addresses for all master proxy gateways;

The controller obtains the key information of each bare metal server node in the correspondence table, creates a first proxy flow table according to the key information of each bare metal server node, and sends it to the corresponding BMS node. The primary proxy gateway, wherein the first proxy flow table includes the virtual IP address and virtual MAC address of the primary proxy gateway, and the key information of the corresponding bare metal server node;

After receiving the first proxy flow table, each main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node and the The virtual IP address and the virtual MAC address construct a first ARP message, and send the first ARP message to the corresponding bare metal server node;

After receiving the first ARP message, the corresponding bare metal server node generates a local ARP entry, so that the corresponding bare metal server node communicates with the destination end according to the ARP entry.

In the active/standby switchover method provided by the embodiment of the present invention, there is no need for protocol interaction between the bare metal server node and the proxy gateway, and all proxy gateways use the same virtual IP address, so that the bare metal server Nodes are not aware of the existence of proxy gateways. At the same time, the number of proxy gateway nodes can be flexibly configured according to user needs.

In a possible implementation manner, the controller elects a new primary proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules Afterwards, the method also includes:

The controller acquires a second bare metal server node using the new primary proxy gateway as the primary proxy gateway according to the correspondence table, and creates a second proxy flow according to key information of the second bare metal server node table, and send it to the new main proxy gateway, the second proxy flow table includes the virtual IP address and virtual MAC address of the new main proxy gateway, and the key information of the second bare metal server node ;

After the new main proxy gateway receives the second proxy flow table, obtains the key information of the second bare metal server node in the second proxy flow table, and based on the second bare metal server node The key information and the virtual IP address and virtual MAC address of the new main agent gateway construct a second ARP message, and send the second ARP message to the second bare metal server node;

After receiving the second ARP message, the second bare metal server node updates a local ARP entry, so that the second bare metal server node communicates with the destination end according to the updated ARP entry.

In the second aspect, the embodiment of the present invention provides an active-standby switchover method, which is applied to a controller in an active-standby switchover system, the controller communicates with at least two proxy gateways, and the at least two proxy gateways Communicatively connected to at least one bare metal server node, the controller stores a corresponding relationship table between the primary agent gateway and the standby agent gateway corresponding to the at least one bare metal server node, and the method includes:

An abnormality of any proxy gateway is detected, and according to the corresponding relationship table, the first bare metal server node using the proxy gateway as the primary proxy gateway is obtained;

Elect a new master proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway processes the first bare metal server node. A message sent by a BMS node.

In a possible implementation manner, the electing a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules includes:

Obtain the number of IP addresses of the bare metal server nodes processed by each standby proxy gateway corresponding to the first bare metal server node, and select the standby proxy gateway with the least number of IP addresses of the processed bare metal server nodes as the new primary proxy gateway;

And/or, acquire the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and select the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway.

In a possible implementation manner, before the detection of the abnormality of any of the proxy gateways, the method further includes:

Configure a master proxy gateway and a backup proxy gateway for each bare metal server node, and store the corresponding relationship between each bare metal server node and the master proxy gateway and backup proxy gateway in the local correspondence table; The proxy gateway is configured with the same virtual IP address and different virtual MAC addresses;

Obtain the key information of each bare metal server node in the correspondence table, create a first proxy flow table according to the key information of each bare metal server node, and send it to the main proxy gateway corresponding to the bare metal server node so that after receiving the first proxy flow table, the main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node And the virtual IP address and the virtual MAC address construct a first ARP message, and send the first ARP message to the corresponding BMS node, triggering the corresponding BMS node according to the first ARP message An ARP message generates a local ARP entry, so that the corresponding bare metal server node communicates with the destination end according to the ARP entry;

Wherein, the first proxy flow table includes the virtual IP address and virtual MAC address of the main proxy gateway, and key information of the corresponding bare metal server node.

In a possible implementation manner, after selecting a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules, The method also includes:

According to the correspondence table, obtain the second bare metal server node with the new primary proxy gateway as the primary proxy gateway, and create a second proxy flow table according to the key information of the second bare metal server node, and download sending to the new primary proxy gateway, so that the new primary proxy gateway obtains the key information of the second bare metal server node in the second proxy flow table after receiving the second proxy flow table , and construct a second ARP message based on the key information of the second bare metal server node and the virtual IP address and virtual MAC address of the new master proxy gateway, and send the second ARP message to the The second bare metal server node triggers the second bare metal server node to update the local ARP entry according to the second ARP message, so that the second bare metal server node can communicate with the destination end according to the updated ARP entry communicate;

Wherein, the second proxy flow table includes the virtual IP address and virtual MAC address of the new master proxy gateway, and key information of the second bare metal server node.

In a third aspect, an embodiment of the present invention provides an active/standby switchover system, the system includes a controller, at least two agent gateways, and at least one bare metal server node, and the controller communicates with the at least two agent gateways The at least two proxy gateways are connected in communication with the at least one bare metal server node, and the controller is configured to execute the active/standby switching method as described in the second aspect above.

In a fourth aspect, an embodiment of the present invention provides a controller, including a processor, the processor is coupled to a memory, the memory stores a program, and when the program instructions stored in the memory are executed by the processor, the The active/standby switching method as described in the second aspect above.

In the active-standby switching method, system and controller provided by the embodiments of the present invention, the bare metal server node does not perceive the existence of the proxy gateway, and when the primary proxy gateway is abnormal, the bare metal server node can forward the business traffic to A new master proxy gateway to quickly restore customer business. The present invention does not require the protocol interaction between the bare metal server node and the proxy gateway, and all the proxy gateways use the same virtual IP address, so that the bare metal server node does not perceive the existence of the proxy gateway. At the same time, the number of proxy gateway nodes can be flexibly configured according to user needs.

Description of drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some of the present invention. Embodiments, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.

FIG. 1 is a schematic diagram of a system architecture of an example of an embodiment of the present invention;

FIG. 2 is a schematic diagram of the system architecture of another example of the embodiment of the present invention;

Fig. 3 is a schematic flow chart of an active/standby switching method according to an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The system architecture and business scenarios described in the embodiments of the present invention are to illustrate the technical solutions of the embodiments of the present invention more clearly, and do not constitute limitations on the technical solutions provided by the embodiments of the present invention. The evolution of the architecture and the emergence of new business scenarios, the technical solutions provided by the embodiments of the present invention are also applicable to similar technical problems.

The above method will be described in detail below in conjunction with specific embodiments.

With the application of smart network cards, especially when standard network cards or smart network cards that do not support network offloading are used on the BMS node of the data center, the OVS component needs to run on an agent node other than the server for specialized network business processing. However, the reliability of these proxy nodes and how to restore the business after the proxy node is abnormal are technical problems that need to be solved urgently in practical applications.

In order to solve the above technical problems, an embodiment of the present invention provides an active/standby switchover system, which includes a controller, at least two agent gateways, and at least one bare metal server node. Wherein, the controller communicates with at least two proxy gateways, and is mainly used for management control, such as flow table distribution, status detection, and control of ARP (Address Resolution Protocol, address resolution protocol) packet distribution. At least two proxy gateways communicate with at least one bare metal server node. The OVS component runs on the proxy gateway to process the business of the bare metal server node. The proxy gateway can be implemented by software and run on a certain server. It may also be implemented through hardware products such as smart network cards.

Example 1

As shown in Figure 1, it is a schematic diagram of the system architecture of the embodiment of the present invention, wherein, the controller communicates with n (n ≥ 2, n is a natural number) proxy gateways, and n proxy gateways communicate with m (m ≥ 1, m is natural number) BMS node communication links, under this system architecture, the controller configures a primary agent gateway and a backup agent gateway for each bare metal server node, wherein, one or more agent gateways can be configured for each bare metal server node based on user requirements The main proxy gateway, and configure at least one backup proxy gateway. For example, one main agent gateway and two standby agent gateways can be set for BMS1. Set up 1 primary proxy gateway, 3 backup proxy gateways, etc. for BMS2.

The controller stores the correspondence between the bare metal server nodes and the active and standby proxy gateways in a local correspondence table. According to the corresponding relationship table, the controller can obtain the information of the bare metal server node proxied by any proxy gateway, so as to send the corresponding proxy flow table to any proxy gateway, so as to realize the processing by any proxy gateway according to the proxy flow table The packet sent by the corresponding bare metal server node.

In the embodiment of the present invention, when the controller detects that any proxy gateway is abnormal, it obtains the first bare metal server node information with any proxy gateway as the primary proxy gateway according to the corresponding relationship table. The obtained first There may be only one bare metal server node, or there may be multiple ones. If there is only one first bare metal server node with any proxy gateway as the primary proxy gateway, the corresponding A new master proxy gateway is elected from the standby proxy gateways, so that the new master proxy gateway processes the message sent by the first bare metal server node. If there are multiple first BMS nodes with any proxy gateway as the primary proxy gateway, select a new primary proxy gateway for each first BMS node from the corresponding backup proxy gateways . Wherein, the active and standby proxy gateways of each first bare metal server node may be the same or different, and corresponding configurations need to be performed on the controller based on user requirements.

The embodiment of the present invention provides an active-standby switchover system. The controller configures the active-standby agent gateway for the bare metal server node, and selects a new active agent gateway for the bare metal server node after the main agent gateway is abnormal, so as to realize the bare metal server node. The metal server node does not need to be aware of the existence of the proxy gateway. When the primary proxy gateway is abnormal, the bare metal server can forward the business traffic to the new primary proxy gateway without awareness, so as to quickly restore customer services. The present invention does not require the interaction of protocols such as keep-alive between the bare metal server node and the proxy gateway, and the number of proxy gateway nodes can be flexibly configured according to user requirements.

Example 2

As shown in Figure 2, it is a schematic diagram of the system architecture of another exemplary active-standby switching system provided by the embodiment of the present invention, wherein the controller communicates with agent gateway 1, agent gateway 2, ... agent gateway n, and the agent gateway 1. Proxy gateway 2, ... proxy gateway n communicates with BMS1-BMSm through a service switch. The service switch configures the port connected to BMS1-BMSm as access mode, configures the VLAN of the port connected to BMS1 as VLAN1, and configures the port connected to BMS2 The VLAN is configured as VLAN 2, and so on, and the VLAN of the port connected to BMSm is configured as VLAN m. The service switch configures the ports linked with the NIC1 of the proxy gateway 1, proxy gateway 2, ... proxy gateway n as trunk mode, allowing VLAN 1-VLANm to pass through. The user sets the gateway IP address of BMS1-BMSm to IP-GW. Set the IP address of BMS1 to IP-1, set the IP address of BMS2 to IP-2, and so on, set the IP address of BMSm to IP-m. Set the MAC address of BMS1 to MAC1, the MAC address of BMS2 to MAC2, and so on, and the MAC address of BMSm to MACm. Set the management IP address of proxy gateway 1 to IP_MG1, the management IP address of proxy gateway 2 to IP_MG2, and so on, and the management IP address of proxy gateway n to IP_MGn. The controller configures the virtual IP addresses of proxy gateway 1 to proxy gateway n as the gateway IP addresses of the bare metal server nodes set by the user, that is, the controller sets the virtual IP addresses of proxy gateway 1 to proxy gateway n as IP-GW .

In this embodiment of the present invention, the controller configures a master proxy gateway and a backup proxy gateway for each bare metal server node, and stores the corresponding relationship between each bare metal server node and the master proxy gateway and at least one backup proxy gateway in the local Correspondence table.

In the embodiment of the present invention, the key information of the bare metal server node includes VLAN ID, IP address and MAC address, and these key information form a one-to-one correspondence with a bare metal server node, which can ensure the access security of the bare metal server node and accuracy. The controller monitors and manages the active and standby proxy gateways based on the management IP address information of the active and standby proxy gateways. Therefore, as an implementation of the embodiment of the present invention, at least the VLAN ID, IP address, and MAC address information of all bare metal server nodes are stored in the correspondence table, as well as the management IP address, virtual IP address, and virtual IP address of the active and standby agent gateways. MAC address information. OVS components are running on both the active agent gateway and the standby agent gateway.

In the system architecture shown in Figure 2, the controller configures the main agent gateway and the standby agent gateway for BMS1-BMSm respectively, and stores the corresponding relationship between BMS1-BMSm and the main agent gateway and the standby agent gateway in the local corresponding relationship table. For example, assuming m=2, that is, there are 2 BMS nodes: BMS1 and BMS2, and the controller configures 1 primary proxy gateway and 2 backup proxy gateways for BMS1 and BMS2 respectively. For example, the primary proxy gateway of BMS1 is set as Proxy gateway 1, standby proxy gateways are proxy gateway 2 and proxy gateway 3; set the main proxy gateway of BMS2 as proxy gateway 2, and the backup proxy gateways are proxy gateway 3 and proxy gateway 4.

The corresponding relationship between BMS1 and BMS2 configured by the controller and the active and standby proxy gateways is shown in Table 1 below.

Table 1

In the embodiment of the present invention, after the corresponding relationship between the proxy gateway and the bare metal server node is established in the controller, the proxy flow table needs to be delivered to the proxy gateway, so that the proxy gateway can process the message from the bare metal server node according to the proxy flow table , to establish communication between bare metal server nodes and other network communication devices. For example, after the controller establishes the corresponding relationship table shown in Table 1 above, according to the corresponding relationship table, it obtains the bare metal server node BMS1 with the proxy gateway 1 as the main proxy gateway, and sends the proxy flow for processing the message of BMS1 to the proxy gateway 1. surface. The proxy gateway 1 processes the message sent by the BMS1 according to the proxy flow table, so as to realize the communication between the BMS1 and the destination end.

It can be understood that the correspondence between bare metal server nodes and active/standby proxy gateways recorded in Table 1 is just an example. When there are more correspondences between bare metal server nodes and active/standby proxy gateways, you can refer to The above example establishes more correspondences between bare metal server nodes and active/standby proxy gateways in the correspondence table.

As an implementation of the embodiment of the present invention, the controller configures the same virtual IP address and different virtual MAC addresses for all the main proxy gateways, so that the bare metal server node cannot perceive the existence of multiple proxy gateways, but can Distinguish different proxy gateways, and the virtual IP address of the proxy gateway is only used in the interactive packets (such as proxy flow table, ARP broadcast packets, etc.), and is not actually configured on the interface, so there will be no IP Address conflict.

As an implementation method of the embodiment of the present invention, the controller obtains the information of the bare metal server nodes that need to be processed by each main proxy gateway according to the corresponding relationship table, and sends the proxy information of the bare metal server nodes that need to be processed to each main proxy gateway. Flow table, the proxy flow table includes the VLAN ID, MAC address and IP address information of the bare metal server node, and the virtual IP address and virtual MAC address information of the main proxy gateway.

When the main agent gateways are normal, each main agent gateway is only responsible for processing the packets of their corresponding bare metal server nodes.

Specifically, when the bare metal server node initially communicates, it sends an ARP Request broadcast message for the gateway address, and the ARP Request broadcast message is sent to all proxy gateways after passing through the service switch, but only the primary proxy gateway corresponding to the bare metal server node The message will be processed. After the main proxy gateway performs ARP proxy response to the ARP Request broadcast message, the bare metal server node generates the corresponding ARP entry locally after receiving the ARP reply message from the main proxy gateway, so as to establish the correspondence between the bare metal server node and its The corresponding relationship of the main proxy gateway of the BMS node, so that the packets sent from the bare metal server node will only be sent to the corresponding main proxy gateway.

When the BMS node needs to communicate with the destination end, the BMS node sends a service packet to the destination end. The source IP address of the service packet is the IP address of the BMS node, and the source MAC address is the BMS node The target MAC address is the virtual MAC address of the main agent gateway corresponding to the bare metal server node, and the destination IP address is the IP address of the target virtualization node, so that the service packet is sent to the target virtualization virtualization nodes to implement communication between bare metal server nodes and target virtualization nodes.

In the embodiment of the present invention, when the controller detects that any proxy gateway is abnormal, according to the above established correspondence table, obtains the corresponding bare metal server node with any proxy gateway as the primary proxy gateway, and according to the preset The rule elects a new master proxy gateway from the standby proxy gateways corresponding to the bare metal server nodes, so as to realize the active/standby switchover of the proxy gateways.

In the embodiment of the present invention, after the controller elects a new master proxy gateway, it needs to deliver to the new master proxy gateway a proxy flow table for processing messages of all second bare metal server nodes of its new proxy; the proxy The flow table includes the VLAN ID, MAC address, and IP address of the second bare metal server node, and the virtual IP address and virtual MAC address of the new master proxy gateway. After receiving the proxy flow table, the new main proxy gateway obtains the key information of the second bare metal server node in the proxy flow table, constructs the corresponding ARP message based on the key information, and sends the ARP message to the corresponding The second bare metal server node. The ARP message constructed here includes at least the VLAN ID, MAC address, and IP address of the second BMS node, and the virtual IP address and virtual MAC address of the main proxy gateway. For different second BMS nodes, the ARP The content of the message is different. Therefore, when the number of second bare metal server nodes newly proxied by the new master proxy gateway is multiple, it is necessary to construct multiple different ARP packets for key information of different second bare metal server nodes. After receiving the corresponding ARP message, the second bare metal server node parses and obtains information such as the virtual MAC address of the new main agent gateway, and updates the local ARP entry.

As a preferred implementation of the embodiment of the present invention, the controller can configure the same virtual IP address and different virtual MAC addresses for all the master proxy gateways; and carry the virtual IP addresses and virtual MAC addresses in the In the proxy flow table of the proxy gateway, when the primary proxy gateway node is abnormal, the bare metal server node can forward the business traffic to the new primary proxy node without awareness, so as to quickly restore customer services.

In the embodiment of the present invention, the bare metal server node does not need to be aware of the existence of the proxy gateway. When the proxy gateway node is abnormal, the bare metal server can forward the business traffic to the new proxy node without awareness, so as to quickly restore customer services .

Example 3

Based on the active-standby switchover system provided by the embodiment of the present invention, the embodiment of the present invention provides an active-standby switchover method, as shown in FIG. 3 , the method includes:

S301. The controller detects that any proxy gateway is abnormal, and obtains the first bare metal server node using the proxy gateway as the primary proxy gateway according to the correspondence table.

In this step, whether the controller detects whether the proxy gateway is abnormal can be judged through the connection established between the controller and the proxy gateway. For example, when the OVS connection between the controller and the proxy gateway is disconnected, it is considered that the proxy gateway is abnormal. Alternatively, when the controller fails to establish a connection with the proxy gateway, it is considered that the proxy gateway is abnormal. Alternatively, when the controller cannot communicate with the proxy gateway, it is considered that the proxy gateway is abnormal.

In this embodiment of the present invention, before the controller detects that any proxy gateway is abnormal, the controller configures a primary proxy gateway and a backup proxy gateway for each bare metal server node, and connects each bare metal server node with the primary proxy gateway The corresponding relationship with the standby proxy gateway is stored in the local corresponding relationship table. Wherein, the correspondence table at least stores key information such as the VLAN ID, IP address, and MAC address of the bare metal server node, and information such as the management IP address, virtual IP address, and virtual MAC address of the active and standby proxy gateways.

After the controller establishes the correspondence table between the bare metal server node and the active and standby agent gateways, it will send the agent flow table to each main agent gateway based on the correspondence table, so that the main agent gateway can process the corresponding bare metal server according to the agent flow table. message sent by the node.

Specifically, the controller obtains the key information of each bare metal server node in the correspondence table, and creates a first proxy flow table based on the key information of each bare metal server node, and sends it to the corresponding The main proxy gateway, wherein the first proxy flow table includes the virtual IP address and virtual MAC address of the main proxy gateway, and key information of the corresponding bare metal server node. The key information of the bare metal server node includes at least: VLAN ID, MAC address, and IP address.

For example, for "the main proxy gateway of BMS1 is proxy gateway 1, and the primary proxy gateway of BMS 2 is proxy gateway 2", the controller sends proxy flow table 1 to proxy gateway 1, and the proxy flow table 1 is used to control BMS1 The sent message is processed, and the message sent by BMS1 carries information such as "VLAN ID=1, src_ip=IP-1, src_mac=MAC1". The controller sends the proxy flow table 2 to the proxy gateway 2. The proxy flow table 2 is used to process the message sent by the BMS2. The message sent by the BMS2 carries "VLAN ID=2, src_ip=IP-2, src_mac= MAC2" and other information. The proxy flow table type may be ARP proxy and/or corresponding service processing, for example, the corresponding service processing may be an encapsulation tunnel or the like. It is important to note that the source IP addresses used in the ARP proxy flow table on all primary proxy gateways are virtual IP addresses. Since the virtual IP address only exists in the entry and is used when the packet is encapsulated, and is not directly configured on an interface, there is no problem of IP address conflict.

In the embodiment of the present invention, after receiving the first proxy flow table, each main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node and The virtual IP address and the virtual MAC address construct the first ARP packet, and send the first ARP packet to the corresponding bare metal server node. After receiving the first ARP message, the corresponding bare metal server node generates a local ARP entry, so that the corresponding bare metal server node communicates with the destination end according to the ARP entry.

It should be noted that, when all the main proxy gateways are normal, each main proxy gateway is only responsible for processing the packets of their corresponding bare metal server nodes.

Specifically, when the bare metal server node initially communicates, after sending an ARP Request broadcast message for the gateway address (the ARP Request broadcast message carries the VLAN ID and the gateway IP address), after the ARP Request broadcast message passes through the service switch It is sent to all proxy gateways, but only the main proxy gateway corresponding to the bare metal server node will process the message. After the main agent gateway performs ARP proxy for the ARP Request broadcast message, the bare metal server node generates the corresponding ARP entry locally after receiving the ARP response from the main agent gateway. Packets sent from the bare metal server node will only be sent to its corresponding master proxy gateway.

When the BMS node needs to communicate with the target virtualization node, the BMS node sends a service packet to the target virtualization node. The source IP address of the service packet is the IP address of the BMS node, and the source MAC address is The MAC address of the bare metal server node, the target MAC address is the virtual MAC address of the main proxy gateway corresponding to the bare metal server node, and the destination IP address is the IP address of the target virtualization node, so that after the service packet passes through the main proxy gateway, Send it to the target virtualization node to realize the communication between the bare metal server node and the target virtualization node.

In the embodiment of the present invention, the IP addresses used by each main agent gateway for ARP proxy are the virtual IP address IP-GW, and the MAC addresses used are respectively the virtual MAC addresses of each main agent gateway, such as the The virtual MAC address is MAC-AG1, the virtual MAC address of agent gateway 2 is MAC-AG2, ..., the virtual MAC address of agent gateway n is MAC-AGn.

As an example, taking the communication between the bare metal server node 1 (BMS1) and the virtualization node 1 in FIG. 2 as an example, the implementation process of the inter-node communication of the present invention is described in detail, as follows:

BMS1 sends an ARP Request broadcast message targeting the gateway address IP-GW. The destination MAC address of the ARP Request broadcast message is ffff.ffff.ffff, the source MAC address is MAC1, the destination IP address is IP-GW, and the source IP address is IP-1, after the service switch interconnected with BMS1 receives the ARP Request broadcast message from port P1, it floods all ports belonging to VLAN 1 (such as adding VLAN ID=1 from ports P1_1, P1_2, and P1_n after flooding flood out), and then all proxy gateways will receive the ARP request broadcast message carrying VLAN ID=1. Since only the main agent gateway 1 corresponding to BMS1 has a proxy flow table matching "VLAN ID=1, src_ip=IP-1, sec_mac=MAC1", only the main agent gateway 1 will process the ARP request broadcast message, and other agents The gateway will discard the received message.

The main proxy gateway 1 responds to the ARP Request broadcast message, and sends an ARP Replay message from the NIC1 interface. The destination MAC address of the ARP Replay message is MAC1, the source MAC address is MAC-AG1, the VLAN ID is 1, and the destination IP The address is IP-1, and the source IP address is IP-GW.

After the service switch receives the ARPReplay message from the P1_1 port connected to the main agent gateway 1, after querying the FDB entry, the VLAN TAG of the message is stripped off, and then sent out from the P1 port connected to BMS1;

After receiving the ARP Replay message, BMS1 locally generates the corresponding ARP entry.

BMS1 sends the first service message to virtualized node 1. The destination MAC address of the first service message is MAC-AG1, the source MAC address is MAC1, the destination IP address is the IP address of virtualized node 1, and the source IP address is IP-1;

After the service switch receives the first service message from the port P1 interconnected with BMS1, after querying the FDB entry, the first service message is added with VLAN ID=1 and sent out from the port P1_1 interconnected with the main agent gateway 1, In this way, only the main proxy gateway 1 can receive the first service message;

After the main proxy gateway 1 receives the first service message, it processes the first service message according to the corresponding flow table requirements (such as adding VXLAN tunnel information, etc.) and sends it out from the interface NIC2;

The service switch receives the processed first service message from the port (such as P2_1) interconnected with the NIC2 of the main proxy gateway 1, and sends it out from the corresponding port (such as port Pup) after being queried by the FDB entry or the routing table entry. , and finally reach the destination virtualization node 1. In this way, the communication between the BMS1 and the virtualization node 1 is realized.

As a preferred implementation of the embodiment of the present invention, the controller can set more than one backup proxy gateway for the bare metal server node, and when the primary proxy gateway is abnormal, a primary proxy gateway is elected from these backup proxy gateways to take over. The number of specific standby proxy gateways can be set according to user requirements.

As an implementation of the embodiment of the present invention, the controller configures the same virtual IP address and different virtual MAC addresses for all the main proxy gateways, so that the bare metal server node cannot perceive the existence of multiple proxy gateways, but can Different proxy gateways are distinguished, and the virtual IP address of the proxy gateway is only used in the exchanged packets and is not actually configured on the interface, so there will be no IP address conflict.

S302. The controller elects a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway handles the first bare metal server message sent by the node.

In this step, the controller elects a new primary proxy gateway from the standby proxy gateway, which may be judged based on the number of IP addresses of bare metal server nodes processed by the standby proxy gateway and the number of flow table entries, so that the corresponding standby proxy gateway A new master proxy gateway is elected from the gateways.

Specifically, the controller elects a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules, including:

In an implementation manner, the controller acquires the number of IP addresses of the BMS nodes processed by each backup proxy gateway corresponding to the first BMS node, and selects the backup proxy gateway with the least number of IP addresses of the BMS nodes processed as the new BMS node. The main agent gateway of ;

In another implementation manner, the controller acquires the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and selects the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway.

Optionally, a new master proxy gateway can also be elected based on the number of IP addresses of bare metal server nodes and the number of flow table entries processed by each standby proxy gateway corresponding to the bare metal server node, for example, from each corresponding standby proxy gateway Select a standby proxy gateway that handles the least number of BMS node IP addresses as the new primary proxy gateway. When the number of bare metal server node IP addresses processed in each corresponding standby proxy gateway is the same, select the standby proxy gateway with the least number of processed flow table entries from each corresponding standby proxy gateway as the new primary proxy gateway . Or, when the number of flow table entries processed in each corresponding standby proxy gateway is the same, select the standby proxy gateway with the least number of BMS node IP addresses processed from each corresponding standby proxy gateway as the new primary proxy gateway. proxy gateway.

For example, if the controller detects that proxy gateway 1 fails, it can obtain the BMS node with proxy gateway 1 as the primary proxy gateway based on the corresponding relationship table as BMS1, and the backup proxy gateways corresponding to BMS1 are proxy gateway 2 and proxy gateway 3. Compare the number of BMS node IP addresses processed by Proxy Gateway 2 and Proxy Gateway 3, and select the one with fewer IP addresses of BMS nodes processed (such as Proxy Gateway 3) as the new primary Proxy Gateway; if Proxy Gateway 2 If the number of IP addresses of bare metal server nodes processed by proxy gateway 3 is the same, determine the number of flow table entries of proxy gateway 2 and proxy gateway 3, and select the one with fewer flow table entries (such as proxy gateway 2) as the new master proxy gateway.

In the embodiment of the present invention, when the main proxy gateway is abnormal, after the controller selects a new main proxy gateway, the controller sends the new main proxy gateway a proxy flow table for processing messages corresponding to the bare metal server node. After a new primary proxy gateway is generated, the new primary proxy gateway will construct an ARP packet with specific characteristics and send it to the corresponding bare metal server node. All affected bare metal server nodes can update ARP entries in time to complete proxy gateway switching, so that subsequent packets of affected bare metal server nodes can be sent to the new proxy gateway, and services can be quickly restored.

Specifically, after the controller elects a new primary proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, the controller obtains the new primary proxy gateway according to the corresponding relationship table. The primary proxy gateway of the primary proxy gateway is the second bare metal server node of the primary proxy gateway, and creates a second proxy flow table based on the key information of the second bare metal server node, and sends it to the new primary proxy gateway, wherein the second proxy flow table The table includes the virtual IP address and virtual MAC address of the new main agent gateway, and key information of the second bare metal server node. The key information of the second bare metal server node may be information such as a VLAN ID, a MAC address, and an IP address. For example, if the main proxy gateway 1 of BMS1 and BMS3 fails, the controller selects proxy gateway 3 as the new primary proxy gateway of BMS1 and BMS3, then the controller sends proxy flow table 3 and proxy flow table to proxy gateway 3 4. Proxy flow table 3 is used to process packets matched with "VLAN ID=1, src_ip=IP-1, src_mac=MAC1", and proxy flow table 4 is used to process packets matching "VLAN ID=3, src_ip=IP -3, src_mac=MAC3" matches the message to be processed. That is, the controller issues the proxy flow table 3 for processing the message sent by BMS1 and the proxy flow table 4 for processing the message sent by BMS3 to the proxy gateway 3 . Among them, src_ip represents the source IP address, and src_mac represents the source MAC address.

After receiving the second proxy flow table, the new main proxy gateway obtains the key information of the second bare metal server node in the second proxy flow table, and based on the key information of the second bare metal server node and the virtual The IP address and the virtual MAC address construct a second ARP packet, and send the second ARP packet to the second bare metal server node. Among them, the ARP message includes at least the VLAN ID, IP address, and MAC address of the bare metal server node and the virtual IP address and virtual MAC address of the new primary agent gateway. For different bare metal server nodes, the content of the ARP message is different. .

In an optional implementation mode, the controller invokes the ARP generator on the new main proxy gateway to construct and generate a certain number of ARP packets according to a certain frequency, such as sending once per second, and sending a total of 3 times. The generated ARP message is sent out from the NIC1 interface of the new main agent gateway after the ARP conversion module constructs an ARPReplay message including the key information of the bare metal server node.

For example, assuming that the main proxy gateway 1 for BMS1 and BMS3 fails, and the new main proxy gateway selected is proxy gateway 3, then the ARP message converted by the ARP conversion module needs to generate two copies, each with the following characteristics:

ARP message 1: VLAN ID=1, arp_op=2, src_mac=MAC-AG3, dst_mac=MAC1, src_ip=IP-GW, dst_ip=IP-1.

ARP packet 2: VLAN ID=3, arp_op=2, src_mac=MAC-AG3, dst_mac=MAC3, src_ip=IP-GW, dst_ip=IP-3. Among them, arp_op indicates the type of ARP packet, src_mac indicates the source MAC address, dst_mac indicates the destination MAC address, src_ip indicates the source IP address, and dst_ip indicates the destination IP address.

Then send the two ARP packets through the specified NIC1 interface. Since the generated ARP packet has the virtual characteristics of the new primary proxy gateway (that is, virtual IP address and virtual MAC address, etc.) and carries the VLAN TAG of all new proxy BMS nodes of the new primary proxy gateway, The text will eventually reach all BMS nodes proxied by the new master proxy gateway.

After receiving the second ARP message, the second bare metal server node updates a local ARP entry, so that the second bare metal server node communicates with the destination end according to the updated ARP entry.

Specifically, after receiving an ARP packet whose destination MAC address is the local MAC address, the BMS node parses and obtains relevant information in the ARP packet, and then updates the local ARP entry, so that the The destination MAC address is updated to the virtual MAC address of the new primary proxy gateway, and the packets on the BMS node will only be sent to the new primary proxy gateway, thus completing proxy gateway switching.

In the active-standby switching method provided by the embodiment of the present invention, the bare metal server node does not need to be aware of the existence of the proxy gateway. When the primary proxy gateway is abnormal, the bare metal server can forward the business traffic to the new proxy gateway without awareness. So as to quickly restore customer business. The present invention does not require the protocol interaction between the bare metal server node and the proxy gateway, and all the proxy gateways use the same virtual IP address and different virtual MAC addresses, so that the bare metal server node does not perceive the proxy gateway exist. At the same time, the number of proxy gateway nodes can be flexibly configured according to user needs.

The embodiment of the present invention provides a master-standby switchover method, which is applied to a controller in a master-standby switchover system, where the controller communicates with at least two agent gateways, and the at least two agent gateways communicate with at least one bare metal The server node is connected in communication, and the controller stores a corresponding relationship table between the primary agent gateway and the standby agent gateway corresponding to at least one bare metal server node, and the method includes:

An abnormality of any proxy gateway is detected, and the first bare metal server node using the proxy gateway as the primary proxy gateway is obtained according to the corresponding relationship table.

Before any proxy gateway abnormality is detected, the method further includes: configuring a primary proxy gateway and a standby proxy gateway for each bare metal server node, and mapping each bare metal server node to the primary proxy gateway and the standby proxy gateway Stored in the local correspondence table; configure the same virtual IP address and different virtual MAC addresses for all master agent gateways. Obtain the key information of each bare metal server node in the corresponding relationship table, create a first proxy flow table based on the key information of each bare metal server node, and send it to the main proxy gateway corresponding to the bare metal server node, so as to After receiving the first proxy flow table, the main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node and the virtual IP address and virtual MAC address Construct the first ARP message, and send the first ARP message to the corresponding bare metal server node, trigger the corresponding bare metal server node to generate a local ARP entry according to the first ARP message, so that the corresponding bare metal server node Communicate with the destination end based on the ARP entry.

Wherein, the first proxy flow table includes the virtual IP address and virtual MAC address of the main proxy gateway, and key information of the corresponding bare metal server node.

Specifically, a new master proxy gateway is selected for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway can process the first bare metal server node sent message.

Wherein, according to preset rules, a new master proxy gateway is elected for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node, including:

A possible implementation method is to obtain the number of IP addresses of BMS nodes processed by each backup proxy gateway corresponding to the first BMS node, and select the backup proxy gateway with the least number of IP addresses of BMS nodes processed as new main proxy gateway;

Another possible implementation is to obtain the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and select the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway.

After electing a new primary proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, the method further includes:

According to the corresponding relationship table, obtain the second bare metal server node with the new main agent gateway as the main agent gateway, and create a second agent flow table based on the key information of the second bare metal server node, and send it to the new main agent gateway, so that after the new master proxy gateway receives the second proxy flow table, it obtains the key information of the second bare metal server node in the second proxy flow table, and based on the key information of the second bare metal server node and the new master The proxy gateway's virtual IP address and virtual MAC address construct a second ARP message, and send the second ARP message to the second BMS node, triggering the second BMS node to update the local ARP table according to the second ARP message entry, so that the second bare metal server node communicates with the destination end according to the updated ARP entry. Wherein, the second proxy flow table includes the virtual IP address and virtual MAC address of the new primary proxy gateway, and key information of the second bare metal server node.

In the active-standby switching method provided by the embodiment of the present invention, the bare metal server node does not need to be aware of the existence of the proxy gateway. When the primary proxy gateway is abnormal, the bare metal server can forward the business traffic to the new proxy gateway without awareness. So as to quickly restore customer business. The present invention does not require the protocol interaction between the bare metal server node and the proxy gateway, and all the proxy gateways use the same virtual IP address and different virtual MAC addresses, so that the bare metal server node does not perceive the proxy gateway exist. At the same time, the number of proxy gateway nodes can be flexibly configured according to user needs.

The embodiment of the present invention also provides a controller. Specifically, the controller includes a processor, the processor is coupled to a memory, and the memory stores a program. When the program instructions stored in the memory are executed by the processor During execution, the steps of the master-standby switchover method described in the foregoing embodiments are realized.

It can be understood that the embodiment of the present invention can also provide an electronic device, specifically, the electronic device includes: a processor, a memory, and a computer program stored in the memory and operable on the processor, the When the computer program is executed by the processor, the steps of the master/standby switchover method provided in the foregoing embodiments are implemented.

Wherein, the above-mentioned processor is the control center of the electronic equipment. By running or executing software programs and/or modules stored in the memory and calling data stored in the memory, it executes various functions of the electronic equipment and processes data, thereby controlling the electronic equipment. The device is monitored as a whole. A processor may include one or more processing units.

The memory can be used to store software programs as well as various data. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function, etc.; the data storage area may store user usage data, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage devices.

An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, multiple processes of the master-standby switching method provided in the above-mentioned embodiments are implemented. , and can achieve the same technical effect, in order to avoid repetition, it will not be repeated here. Among them, the computer-readable storage medium, such as random access memory (Random-Access Memory, RAM), read-only memory (Read-Only Memory, ROM), flash memory (Flash Memory), hard disk (Hard Disk Drive, HDD) , Solid State Drive (Solid-State Drive, SSD), optical disc, etc.

Those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of each module of the controller described above can refer to the corresponding process in the foregoing method embodiment, and will not be repeated here.

In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the above-described system embodiments are only illustrative, and may also be other similar systems omitting some devices, or similar system architectures in related fields, to implement the method process provided by the embodiments of the present invention, which will not be repeated here repeat.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions depart from the protection scope of the present invention.

Claims (10)

1. A master-standby switchover method, characterized in that it is applied to a master-standby switchover system, the system includes a controller, at least two agent gateways and at least one bare metal server node, the controller and the at least two The agent gateways are connected in communication, the at least two agent gateways are in communication connection with the at least one bare metal server node, and the controller stores the correspondence between the main agent gateway and the standby agent gateway corresponding to the at least one bare metal server node. relational table, the method includes: The controller detects that any proxy gateway is abnormal, and obtains the first bare metal server node with any proxy gateway as the primary proxy gateway according to the correspondence table; The controller elects a new master proxy gateway for the first bare metal server node from the backup proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway Processing the message sent by the first bare metal server node. 2. The method according to claim 1, wherein the controller elects a proxy gateway for the first bare metal server node from the standby agent gateway corresponding to the first bare metal server node according to a preset rule. New master proxy gateway, including: The controller obtains the number of IP addresses of the bare metal server nodes processed by each standby proxy gateway corresponding to the first bare metal server node, and selects the standby proxy gateway with the least number of IP addresses of the processed bare metal server nodes as the new main proxy gateway; And/or, the controller acquires the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and selects the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway. 3. The method according to claim 1 or 2, wherein, before the controller detects that any proxy gateway is abnormal, the method further comprises: The controller configures a master proxy gateway and a backup proxy gateway for each bare metal server node, and stores the correspondence between each bare metal server node and the master proxy gateway and backup proxy gateway in the local correspondence table; The controller configures the same virtual IP address and different virtual MAC addresses for all master proxy gateways; The controller obtains the key information of each bare metal server node in the correspondence table, creates a first proxy flow table according to the key information of each bare metal server node, and sends it to the corresponding BMS node. The primary proxy gateway, wherein the first proxy flow table includes the virtual IP address and virtual MAC address of the primary proxy gateway, and the key information of the corresponding bare metal server node; After receiving the first proxy flow table, each main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node and the The virtual IP address and the virtual MAC address construct a first ARP message, and send the first ARP message to the corresponding bare metal server node; After receiving the first ARP message, the corresponding bare metal server node generates a local ARP entry, so that the corresponding bare metal server node communicates with the destination end according to the ARP entry. 4. The method according to claim 3, wherein the controller selects the proxy gateway for the first bare metal server node from the standby proxy gateway corresponding to the first bare metal server node according to preset rules. After the new main agent gateway is released, the method also includes: The controller acquires a second bare metal server node using the new primary proxy gateway as the primary proxy gateway according to the correspondence table, and creates a second proxy flow according to key information of the second bare metal server node Table, and issued to the new master proxy gateway, wherein, the second proxy flow table also includes the virtual IP address and virtual MAC address of the new master proxy gateway; After the new main proxy gateway receives the second proxy flow table, obtains the key information of the second bare metal server node in the second proxy flow table, and based on the second bare metal server node The key information and the virtual IP address and virtual MAC address of the new main agent gateway construct a second ARP message, and send the second ARP message to the second bare metal server node; After receiving the second ARP message, the second bare metal server node updates a local ARP entry, so that the second bare metal server node communicates with the destination end according to the updated ARP entry. 5. A master-standby switchover method, characterized in that it is applied to a controller in a master-standby switchover system, the controller communicates with at least two proxy gateways, and the at least two proxy gateways communicate with at least one bare metal server The nodes are connected by communication, and the controller stores a corresponding relationship table between the primary agent gateway and the standby agent gateway corresponding to the at least one bare metal server node, and the method includes: An abnormality of any proxy gateway is detected, and according to the corresponding relationship table, the first bare metal server node using the proxy gateway as the primary proxy gateway is obtained; Elect a new master proxy gateway for the first bare metal server node from the standby proxy gateways corresponding to the first bare metal server node according to preset rules, so that the new master proxy gateway processes the first bare metal server node. A message sent by a BMS node. 6. The method according to claim 5, characterized in that, according to preset rules, a new primary agent gateway is selected for the first bare metal server node from the standby agent gateway corresponding to the first bare metal server node. Proxy gateways, including: Obtain the number of IP addresses of the bare metal server nodes processed by each standby proxy gateway corresponding to the first bare metal server node, and select the standby proxy gateway with the least number of IP addresses of the processed bare metal server nodes as the new primary proxy gateway; And/or, acquire the number of flow table entries processed by each standby proxy gateway corresponding to the first bare metal server node, and select the standby proxy gateway with the least number of processed flow table entries as the new primary proxy gateway. 7. The method according to claim 5 or 6, wherein, before the detection of the abnormality of any of the proxy gateways, the method further comprises: Configure a master proxy gateway and a backup proxy gateway for each bare metal server node, and store the corresponding relationship between each bare metal server node and the master proxy gateway and backup proxy gateway in the local correspondence table; The proxy gateway is configured with the same virtual IP address and different virtual MAC addresses; Obtain the key information of each bare metal server node in the correspondence table, create a first proxy flow table according to the key information of each bare metal server node, and send it to the main proxy gateway corresponding to the bare metal server node so that after receiving the first proxy flow table, the main proxy gateway obtains the key information of the corresponding bare metal server node in the first proxy flow table, and based on the key information of the corresponding bare metal server node And the virtual IP address and the virtual MAC address construct a first ARP message, and send the first ARP message to the corresponding BMS node, triggering the corresponding BMS node according to the first ARP message An ARP message generates a local ARP entry, so that the corresponding bare metal server node communicates with the destination end according to the ARP entry; Wherein, the first proxy flow table includes the virtual IP address and virtual MAC address of the main proxy gateway, and key information of the corresponding bare metal server node. 8. The method according to claim 7, wherein, in the standby proxy gateway corresponding to the first bare metal server node according to preset rules, a new proxy gateway is selected for the first bare metal server node After the master proxy gateway, the method further includes: According to the correspondence table, obtain the second bare metal server node with the new primary proxy gateway as the primary proxy gateway, and create a second proxy flow table according to the key information of the second bare metal server node, and download sending to the new primary proxy gateway, so that the new primary proxy gateway obtains the key information of the second bare metal server node in the second proxy flow table after receiving the second proxy flow table , and construct a second ARP message based on the key information of the second bare metal server node and the virtual IP address and virtual MAC address of the new master proxy gateway, and send the second ARP message to the The second bare metal server node triggers the second bare metal server node to update the local ARP entry according to the second ARP message, so that the second bare metal server node can communicate with the destination end according to the updated ARP entry communicate; Wherein, the second proxy flow table further includes the virtual IP address and the virtual MAC address of the new primary proxy gateway. 9. An active/standby switchover system, characterized in that the system includes a controller, at least two agent gateways and at least one bare metal server node, the controller communicates with the at least two agent gateways, the At least two proxy gateways are communicatively connected to the at least one bare metal server node, and the controller is configured to execute the active/standby switching method according to any one of claims 5-8. 10. A controller, characterized in that it comprises a processor, the processor is coupled to a memory, and the memory stores a program, and when the program instructions stored in the memory are executed by the processor, the process according to claim 5 is realized. - The active-standby switching method described in any one of 8.