CN113364633B - Container cluster dynamic construction method oriented to high-mobility environment - Google Patents
Container cluster dynamic construction method oriented to high-mobility environment Download PDFInfo
- Publication number
- CN113364633B CN113364633B CN202110677750.2A CN202110677750A CN113364633B CN 113364633 B CN113364633 B CN 113364633B CN 202110677750 A CN202110677750 A CN 202110677750A CN 113364633 B CN113364633 B CN 113364633B
- Authority
- CN
- China
- Prior art keywords
- node
- cluster
- management
- container
- working
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
- H04L41/0668—Management of faults, events, alarms or notifications using network fault recovery by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0823—Errors, e.g. transmission errors
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Abstract
The invention provides a container cluster dynamic construction method for a high-mobility environment, which comprises the following steps: firstly, a container deployment server is started, a container cluster is dynamically constructed according to the connection condition of adjacent servers in a network, or a single-computer cluster is independently constructed or a multi-computer cluster is added, and simultaneously the roles in the cluster are divided into a management node, a management succession node and a working node, and a corresponding cluster management control module group is loaded according to the roles; and then the server starts to work, and the dynamic adjustment of the container cluster is realized through the cooperation of the management control module group among all the roles in the cluster according to the connection change of the adjacent server in the network, or a new server node is added, or a combined multi-computer cluster is added, or a server node is withdrawn and reconstructed into a new single-computer cluster or a multi-computer cluster. The method is simple and easy to implement, can provide a highly available container cluster construction means for a high-mobility environment, and can support organization and application of a container cloud platform environment.
Description
Technical Field
The invention belongs to the technical field of computers, and particularly relates to a container cluster dynamic construction method for a high-mobility environment.
Background
With the rapid development of the fields of internet of things and internet of vehicles, the edge computing technology is continuously broken through, various cloud services are extended to the edge, the scale and complexity of an information system are further improved, and challenges are brought to the service organization and application of the edge environment.
In the comprehensive military and civil market, a large number of computing processing devices are used for building various information systems in the edge environment, but the devices have uneven software and hardware resources, different performances and even poorer performances, wherein terminal devices are particularly prominent, and limited computing, storage and network resources can restrict the functional services which can be integrally deployed and applied in various places. The cloud service can be provided by intensively applying server resources in the edge environment to construct a cluster for hopefully bringing more services with guaranteed quality to various end devices. Furthermore, server resources of motor vehicles, ships, aircrafts and the like can be properly utilized and the system is suitable for supporting a business system of the motor vehicles, the ships, the aircrafts and the like, the application requirement of a high-mobility scene is met, the resource utilization efficiency and the system availability are improved, the method can be applied to a container cluster dynamic construction technology under the high-mobility condition of the edge environment, and becomes a necessary consideration factor in the overall planning of an information system software service platform.
At present, a dynamic construction method for a container cluster under a high-mobility condition of an edge environment is not available, so that self-organized dynamic construction of a container deployment operation server cluster is designed comprehensively, high availability of clustered resources is realized, application risks of an information system caused by server hardware faults, software system problems and artificial misoperation are reduced, and the problem is a problem which needs to be solved urgently in current information system construction.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to solve the technical problem of the prior art, provides a container cluster dynamic construction method facing to a high-mobility environment, and can realize self-organization adaptive construction of a container cluster deployment operation environment under the high-mobility condition on the basis of an edge application scene.
The technical scheme is as follows: the container cluster dynamic construction method facing the high-mobility environment comprises the following steps:
when a container deployment server (referring to a computer server for deploying containers) is started, dynamically constructing a container cluster (referring to a server cluster for deploying containers) according to the connection condition of adjacent servers in a network, or independently constructing a single-machine cluster (referring to a container cluster constructed by a single server) or joining a multi-machine cluster (referring to a container cluster constructed by more than one server), and simultaneously dividing roles in the container cluster into a management node, a management succession node and a working node, and loading a corresponding cluster management control module group according to the roles;
after the container deployment server starts working, the dynamic adjustment of the container cluster is realized according to the connection change of adjacent servers in the network and the cooperation of the management control module group among all the roles in the cluster, or a new server node is added, or a combined multi-machine cluster is added, or a server node is withdrawn and reconstructed into a new single-machine cluster or a multi-machine cluster.
Further, the container (in the computer field, a container refers to a series of processes isolated and isolated from other parts of a computer system) refers to a software entity capable of being created, started, stopped, deleted, and paused when a (container) image is run, wherein the image is a special file system including all contents required for the container to run: programs, runtime environments, system tools, system libraries, and settings.
Further, the network refers to an internal (e.g., vehicle, ship, airplane, etc.) and an external (e.g., vehicle, ship, airplane, etc.) network of a mobile platform (e.g., vehicle, ship, airplane, etc.) to which the container deployment server belongs, wherein the internal (e.g., vehicle, ship, airplane, etc.) network is mostly connected in a wired manner, and the external (e.g., vehicle, ship, airplane, etc.) network is connected in a wireless manner and conforms to the IP communication standard.
The working nodes are service nodes which can only be assigned to execute tasks in the container cluster, the cluster management control module groups which are correspondingly loaded by the working nodes comprise node state report monitoring modules, and the node state report monitoring modules are used for reporting the working state of the local machine to the container cluster management nodes in real time and receiving the responses of the management nodes.
The management node has the functions of a working node and also has the master control scheduling capability of a container cluster, the management node is provided with a correspondingly loaded cluster management control module which comprises a cluster node composition management module, a cluster node state monitoring module, a cluster expansion authentication module and an adjacent cluster searching module, wherein the cluster node composition management module is used for managing a cluster node composition list of the container cluster, and the cluster node composition list comprises IP addresses and role information of all nodes;
the cluster node state monitoring module is used for collecting the working states of all the working nodes in the container cluster in real time;
the cluster extension authentication module is used for monitoring the access of a new node of an authentication container cluster to the network and the synchronization of other container clusters;
and the adjacent cluster searching module is used for searching adjacent other container clusters in the network in real time.
Further, the management successor node is a working node, but can directly become a management node of a new cluster after the management node in the container cluster fails to be able to be lost, and the setting of the cluster management control module group correspondingly loaded by the management successor node includes: the node status report listening module required as a working node and the cluster node (specific to the management successor node) constitute a backup handover module,
the cluster node composition backup handover module is used for receiving all node information of cluster composition issued by the cluster management node of the record container and transitionally handing over the cluster management node.
Further, when the container deployment server is started, dynamically constructing a container cluster according to the connection condition of adjacent servers in the network, or independently constructing a single-machine cluster or joining a multi-machine cluster, and simultaneously dividing roles in the container cluster into a management node, a management replacement node and a working node, and loading a corresponding cluster management control module group according to the roles, specifically comprising:
step a1, detecting time range T of network adjacent environment smax In time sequence t, the local server (referred to as the container deployment server that is started, described in the local view) is started s Sending a node network access application to the adjacent (referring to other container deployment servers possibly existing in the network) at an upper timing until receiving a network access permission or t of an adjacent container cluster management node cluster expansion authentication module s ≥T smax : when receiving the network access permission, indicating that the local server node passes the authentication and is added to a cluster node of an adjacent cluster as a work node role to form a list to form a new multi-computer cluster, and setting the local server as the work node role at the moment; otherwise, the local server is independently constructed into a single machine cluster and set as a management node role;
step a2, if the container cluster added by the local server is originally a stand-alone cluster: the management node of the newly constructed container cluster changes the role of the new network access node into a management take-over node through a cluster node composition management module, simultaneously sends the information of the management node and the management take-over node in a cluster node composition list (of the container cluster) to all working nodes (including the management take-over node), receives and updates the information of the management node and the management take-over node recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management take-over node changes the role of the working node into the role of the management take-over node;
step a3, the management node of the newly constructed container cluster sends the cluster node composition list to the management succession node, and the backup handover module composed of the cluster nodes of the management succession node receives the information of the backup storage cluster node composition list.
After the container deployment server starts working, according to connection change of adjacent servers in a network and cooperation of management control module groups among all roles in the cluster, the container cluster is dynamically adjusted, or a new server node is added, or a combined multi-computer cluster is added, or a server node is withdrawn to be reconstructed into a new single-computer cluster or a multi-computer cluster, and the method specifically comprises the following steps:
step b1, when monitoring and authenticating that there is a non-networked server node in the network, adding a new node into the container cluster;
b2, when other container clusters exist in the vicinity of the network by searching and authenticating, according to a preset strategy, one container cluster is selected to join and merge to the other container cluster, and a new cluster is constructed by combination;
and step b3, reconstructing the container cluster as a new cluster when the state monitoring finds that the nodes in the container cluster exit.
Step b1 includes:
b1-1, the cluster extension authentication module of the container cluster management node monitors the external server node network access application through the network, and verifies the system verification code attached to the application when the node network access application is monitored;
step b1-2, after the system verification code passes the authentication, adding the networking application node as a work node role to a cluster node composition list, recording the IP address and role information of the networking application node, simultaneously replying networking permission to the networking application node, and setting the networking application node as a work node role when receiving the networking permission;
step b1-3, if the container cluster before adding the node is originally a stand-alone cluster: the management node of the newly constructed container cluster changes the role of the new network access node into a management succession node through a cluster node composition management module, simultaneously sends the management node and the management succession node information in a cluster node composition list to all working nodes (including the management succession node), receives and updates the management node and the management succession node information recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management succession node changes the role of the working node into the role of the management succession node;
step b1-4, the management node of the newly constructed container cluster sends the cluster node composition list to the management succession node, and the backup handover module composed of the cluster nodes of the management succession node receives the information of the backup storage cluster node composition list.
Step b2 includes:
b2-1, the adjacent cluster searching module of the container cluster management node inquires whether other container clusters exist in the network at regular time, and submits the response to the cluster expansion authentication module to verify the system verification code attached to the response of other container clusters;
step b2-2, after the system verification code passes the authentication, starting container cluster merging, and calculating the cluster merging relationship according to the following formula:
wherein N is a 、N b Respectively representing the number of nodes formed by the container cluster A and the number of nodes formed by the cluster B; i is a 、I b Respectively representing an IP address representation value of a container cluster A management node and an IP address representation value of a cluster B management node (namely four numbers of the IP address are used as integers after combination conversion from high bytes to low bytes from front to back); g s 1 indicates that this cluster is a merge side, G s 0 means that the present cluster is the merging party;
b2-3, after the cluster merging relation is determined, the merge side container cluster management node obtains the cluster node composition list from the cluster node composition management module, and sends the cluster node composition list information to the adjacent cluster searching module of the merge side container cluster management node;
step b2-4, when the merger container cluster management node receives the cluster node composition list of the merger container cluster, all the nodes of the merger are used as working nodes to be handed over to the cluster node composition management module to perform cluster node composition list merging, if the container cluster before merging is single-machine cluster (the condition is not satisfied, the merger container cluster is a multi-machine cluster and is an existing management replacement node, so the operation of the second half section is not needed), the role of the original management node of the merger is changed into a management replacement node, the management node and the management replacement node information in the cluster node composition list are issued to all the working nodes (including the management replacement node), the node state report monitoring module of all the working nodes receives the management node and the management replacement node information recorded by updating the node, wherein the working node matched with the IP address of the management replacement node sets the change of the self role as the management replacement node angle node Color;
b2-5, the management node of the newly constructed container cluster sends the cluster node composition list to the management take-over node, and the backup handover module composed of the cluster nodes of the management take-over node receives the information of the backup storage cluster node composition list;
step b3 includes:
b3-1, all the working nodes of the container cluster report the working state of the machine to the cluster node state monitoring module of the management node at regular time, and correspondingly receive the response of the management node;
step b3-2, when the management node receives the working state abnormity (such as failure, pause, etc.) of any working node or fails to receive the working state of any working node overtime, it indicates that the corresponding working node exits the container cluster; when the management successor node does not receive the response of the management node after overtime, the management node is indicated to exit the container cluster;
step b3-3, when the non-management successor work nodes in the container cluster quit, deleting the quitted non-management successor work nodes to reconstruct the quitted non-management successor work nodes into a new cluster;
step b3-4, when the management successor node in the container cluster exits, deleting the exited management successor node and replacing the management successor node to reconstruct the management successor node into a new cluster;
step b3-5, when the management node in the container cluster exits, replacing the management node to reconstruct as a new cluster;
wherein, step b3-3 includes:
step b3-3-1, the container cluster management node deletes the quitted non-management successor working node from the cluster node composition list through the cluster node composition management module;
b3-3-2, the cluster node composition management module sends the modified cluster node composition list to the management succession node, and the backup handover module composed of the cluster nodes of the management succession node receives the information of the backup storage cluster node composition list;
step b3-4 includes:
step b3-4-1, the container cluster management node deletes the exited management successor node from the cluster node composition list through the cluster node composition management module;
step b3-4-2, calculating to obtain a new management successor node of the cluster according to the following preset strategy management successor node selection calculation formula:
wherein N represents the number of working nodes in the container cluster; rand (1, N) represents that one of all working nodes forming a list from cluster nodes is randomly selected; p s Indicating the position number of the selected management successor node in all working nodes (starting with 1 and adding 1 one by one from front to back), P s 0 represents that the cluster (which is a stand-alone cluster) has no management successor node;
step b3-4-3, after the new management succession node of the cluster is determined: changing the role of the corresponding node in the cluster node composition list into a management succession node, simultaneously sending the information of the management node and the management succession node in the cluster node composition list to all working nodes (including the management succession node), receiving and updating the information of the management node and the management succession node recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management succession node changes the role of the working node into the role of the management succession node; the cluster node composition list is issued to a management take-over node, and a backup handover module is composed of the cluster nodes of the management take-over node and receives information of a backup storage cluster node composition list;
step b3-5 includes:
b3-5-1, changing the management succession node of the container cluster into a management node role, and forming a backup handover module through the cluster nodes to send a management node succession notice to all the non-management succession working nodes of the original cluster;
b3-5-2, after all non-management succession working nodes receive the management node succession notice, when confirming that the node status report is the original management succession node recorded by the node status report monitoring module, forming a backup handover module to reply the receipt confirmation to the cluster node of the current management node (original management succession node);
b3-5-3, after the management node receives the acknowledgement of all the working nodes in sequence, adding the corresponding nodes as the working node roles to the cluster node composition list and recording the IP addresses and role information of the nodes;
b3-5-4, calculating to obtain a new management successor node of the cluster according to the preset strategy management successor node selection calculation formula in the step b 3-4-2: changing the role of the corresponding node in the cluster node composition list into a management succession node, simultaneously sending the information of the management node and the management succession node in the cluster node composition list to all working nodes (including the management succession node), receiving and updating the information of the management node and the management succession node recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management succession node changes the role of the working node into the role of the management succession node; and then, issuing the cluster node composition list to a management take-over node, and receiving the information of the backup storage cluster node composition list by a backup handover module consisting of the cluster nodes of the management take-over node.
Has the advantages that: under the condition of high mobility of the edge environment, the invention can self-organize and create and adjust the container cluster server node composition according to the connection condition and real-time change of the network adjacent server, and correspondingly dynamically construct new container clusters in the form of single machine clusters and multi-machine clusters. The method not only supports the dynamic high-availability construction of the container cluster, but also realizes the scene application of the container cluster construction facing high-mobility environment, so that the method can be more practically applied to service information systems of motor vehicles, ships, aircrafts and the like, and simultaneously supports the organization and application of the container cloud platform environment applying the method to provide more services with guaranteed quality for various end equipment, thereby being beneficial to improving the stability of the system cluster service environment and strengthening the operation efficiency of the information system.
Drawings
The foregoing and/or other advantages of the invention will become more apparent from the following detailed description of the invention when taken in conjunction with the accompanying drawings.
Fig. 1 is a schematic view of an application scenario of the method of the present invention.
FIG. 2 is a schematic diagram of container cluster startup build according to an embodiment of the present invention.
Fig. 3 is a process diagram of initial creation of a container cluster according to an embodiment of the invention.
Fig. 4 is a schematic diagram of a process for initial aggregation of a container cluster according to an embodiment of the present invention.
Fig. 5 is a schematic diagram of container cluster operation adjustment according to an embodiment of the present invention.
Fig. 6 is a schematic process diagram of container cluster node expansion according to an embodiment of the present invention.
Fig. 7 is a process diagram of container cluster proximity merging according to an embodiment of the present invention.
Fig. 8 is a schematic diagram of a container cluster node exit process according to an embodiment of the present invention.
Detailed Description
With reference to fig. 1, according to an embodiment of the present invention, a method for dynamically constructing a container cluster oriented to a high-mobility environment includes the following steps:
the first step, starting and constructing a container cluster: as shown in fig. 2, when the container deployment server is started, a container cluster is dynamically constructed according to the connection condition of neighboring servers in the network, or a single-machine cluster is independently constructed, or a multi-machine cluster is added, and simultaneously, roles in the cluster are divided into management nodes, management replacement nodes and working nodes, and corresponding cluster management control module groups are loaded according to the roles. The method comprises the following steps:
(1-1) initial creation of a container cluster: when a first container deployment server in the network is started, the server is independently constructed into a stand-alone cluster, so that the initial construction of the container cluster is realized. The process is shown in fig. 3, and the detailed steps are as follows:
1) and (3) node network access application: detecting time range T in network adjacent environment smax In this network, the server is at time t, the first container in the network to be deployed s Upper timing (interval set to T) d The number of detection counts is n d ) Sending a node network access application to the adjacent (if any container cluster) in a network broadcast mode, namely detecting whether the container cluster which can be added exists in the network adjacent environment;
2) single cluster creation confirmation: time-out, i.e. t s ＝n d ×T d ≥T smax When the network access permission is not received yet, the situation that no container cluster can be added in the network adjacent environment exists at the moment is shown, namely the local machine is the first container deployment server started in the network, and the situation that the local machine server independently constructs a single machine cluster is confirmed;
3) and (3) management node configuration: the method is characterized in that a local machine is set to be a management node role, namely a cluster management control module required by loading and configuring the role comprises a cluster node composition management module (used for a main management container cluster node composition list comprising IP addresses and role information of all nodes), a cluster node state monitoring module (used for collecting working states of all nodes in a container cluster in real time), a cluster expansion authentication module (used for monitoring network access of new nodes of the authentication container cluster and grid connection of other container clusters) and a neighboring cluster searching module (used for searching neighboring other container clusters in the network in real time), and then the management node is taken as a first node of the composition list to be added to a cluster node composition list of the cluster node composition management module and the IP addresses and the role information of the node are recorded.
(1-2) initial polymerization of container clusters: when any other container deployment server in the network (if any) is started, all the other container deployment servers are added into the created container cluster, and all the servers in the adjacent network are initially aggregated to form the same container cluster. The process is shown in fig. 4, and the detailed steps are as follows:
1) and (3) node network access application: detecting time range T in network adjacent environment smax Internal, native (other container deployment in the network) servers at time sequence t s Upper timing (interval set to T) d The number of detection counts is n d ) In a network broadcast mode toSending a node network access application by adjacent (if any container cluster), namely detecting whether an addable container cluster exists in a network adjacent environment, wherein the application incidental information comprises a local IP address and a system verification code;
2) and (3) node network access permission: when a cluster expansion authentication module of a cluster management node of an adjacent container monitors a server node network access application outside a cluster through a network, whether a verification code submitted by the application and a uniform identification code of a system are consistent or not is verified, if the verification code and the uniform identification code are consistent, the network access application is passed, at the moment, an applicant node is used as a work node role and added to a cluster node composition list of a cluster node composition management module, the IP address and role information of the node are recorded, and then the node network access permission is replied to the applicant according to a destination IP address;
3) and (3) configuration of a working node: when receiving the network access permission, the applicant confirms that the local machine is used as a working node to join a corresponding container cluster, and sets the local machine as a working node role at the moment, namely, the cluster management control module group required by loading and configuring the role comprises a node state report monitoring module (used for reporting the working state of the local machine to a container cluster management node in real time and receiving a management node response and the like).
4) The management successor node confirms: the management node indicates that the container cluster has not confirmed the management successor node when finding that the management node is not the management successor node by checking the node at the first position in the cluster node composition list, and at this time, the management successor node selection calculation formula according to the preset strategy specifically comprises the following steps:
the compute knows the management successor node for the container cluster. Wherein N represents the number of working nodes in the container cluster; rand (1, N) represents that one of the working node lists is randomly selected; p s Indicating the position of the selected management successor node in the list of working nodes, P s And 0 represents that the cluster (which is a stand-alone cluster) has no management successor node. At the moment, the role of the corresponding position node in the cluster node composition list is changed into managementTaking over the node, adjusting the node (if not at the first position) to the first position of the list, and then sending all the working nodes (including management take-over nodes) according to the destination IP address by using the management take-over relationship of the container cluster, namely the management node and the management take-over node information in the cluster node composition list;
5) managing the succession relationship for standby: when the node state report monitoring modules of all the working nodes receive the management succession relationship of the management node, updating the IP address information of the management node and the management succession node, namely the management node and the management succession node, of the container cluster management succession relationship recorded by the node for standby;
6) managing the configuration of the succession node: finding that the IP address of the local machine is consistent with the management successor node of the management successor relationship, indicating that the local machine is confirmed to be the management successor node of the cluster, setting the local machine to be the role of the management successor node at the moment, namely, the local machine is used as a node state report monitoring module required by a working node, and a cluster management control module group required by the role is loaded and configured and comprises a cluster node composition backup handover module (used for receiving cluster composition all node information sent by a record container cluster management node and transiting and handing over the cluster management node);
7) the cluster nodes form backup: the management node sends a cluster node composition list to the management take-over node according to the destination IP address, and the cluster node composition backup handover module of the management take-over node receives and stores the list information of the backup.
Step two, adjusting the operation of the container cluster: as shown in fig. 5, after the container cluster deployment server starts to work, according to the connection change of the neighboring servers in the network and the cooperation of the management control module group among the roles in the cluster, the container cluster is dynamically adjusted, or a new (or rejoined) server node is added, or a combined multi-machine cluster is added, or a server node is withdrawn to reconstruct a new single-machine cluster or a multi-machine cluster. The method comprises the following steps:
(2-1) expanding the container cluster nodes: when the container cluster monitors, authenticates and discovers that a container deployment server does not access the network in the vicinity (generally under the conditions of server startup, restart and the like), the container cluster constituent nodes are expanded to join the server to reconstruct the cluster. The process is shown in FIG. 6 (same as FIG. 4), and the detailed steps are the same as those of (1-2).
(2-2) container cluster neighborhood merging: when the container cluster searching authentication finds that other container clusters exist in the vicinity of the network (generally under the condition that a bearing platform of one container cluster, such as a vehicle, and the like, is in a network communication range of the other container cluster in a maneuvering and approaching manner), one of the container clusters selects to join and merge to the other container cluster to form a combined reconstructed cluster according to a preset strategy. The process is shown in fig. 7, and the detailed steps are as follows:
1) neighborhood container cluster query: a side container cluster is timed (with interval set to T) by a neighbor cluster search module managing nodes f ) Sending a neighbor cluster inquiry to a neighbor (if any container cluster) in a network broadcast mode, namely inquiring whether other container clusters exist in the neighbor in the network, wherein the inquiry incidental information comprises a local (namely a cluster management node) IP address and a system verification code;
2) neighbor container cluster reply: when the cluster extension authentication module of the management node of the other adjacent container cluster monitors whether an inquiry exists in the adjacent cluster outside the cluster through a network, whether the verification code submitted by the inquiry is consistent with the uniform identification code of the system is verified, if the verification code is consistent with the uniform identification code of the system, the inquiry cluster passes the authentication, the IP address and the role information of the management node of the inquiry container cluster are recorded to the container cluster to be merged, then the adjacent cluster response is replied to the adjacent cluster search module of the inquiry party according to the destination IP address, and the response incidental information comprises a list formed by the cluster nodes of the container cluster of the answering party; (Note: the above 1, 2 steps are simultaneous mutual inquiry responses of both clusters)
3) Neighbor cluster list merging: when the neighbor cluster search module of any container cluster receives the neighbor cluster response of the other side, the cluster merging relationship calculation formula according to the preset strategy specifically comprises:
and calculating to obtain the cluster merging relation. Wherein N is a 、N b Representing clusters of containers separatelyThe number of nodes formed by A and cluster B; i is a 、I b Respectively representing IP address representation values of management nodes of the container cluster A and the cluster B (namely four numbers of the IP addresses are used as integers after combination conversion from high bytes to low bytes from front to back); g s 1 means this cluster is a merge, G s 0 means that the present cluster is the merging side. At this time, the cluster is waiting for the merging side to be merged; the local cluster party is a merging party and starts cluster merging, and all nodes of the merging party are used as working nodes and handed to a cluster node composition management module to complete sequential addition and merging of a cluster node list;
4) the management successor node confirms: the management node indicates that the container cluster does not confirm the management successor node temporarily when finding that the management node is not the management successor node by checking the node at the next position in the cluster node composition list, changes the role of the working node corresponding to the original management node of the combined party in the list into the management successor node, adjusts the node (if the node is not at the next position) to the next position in the list, and then sends all the working nodes (including the management successor node) according to the management successor relationship of the container cluster, namely the information of the management node and the management successor node in the cluster node composition list and the destination IP address;
5) managing the succession relationship for standby: when the node state report monitoring modules of all the working nodes receive the management succession relations of the management nodes, updating the IP address information of the container cluster management succession relations, namely the management nodes and the management succession nodes, recorded by the node for standby;
6) managing the configuration of the succession node: finding that the IP address of the local machine is consistent with the management successor node of the management successor relationship, indicating that the local machine is confirmed to be the management successor node of the cluster, setting the local machine to be the role of the management successor node at the moment, namely, the local machine is used as a node state report monitoring module required by a working node, and a cluster management control module group required by the role is loaded and configured and comprises a cluster node composition backup handover module (used for receiving cluster composition all node information sent by a record container cluster management node and transiting and handing over the cluster management node);
7) the cluster nodes form backup: the management node sends a cluster node composition list to the management take-over node according to the destination IP address, and the cluster node composition backup handover module of the management take-over node receives and stores the list information of the backup.
(2-3) the container cluster node exits: and (2) monitoring the searching state of the container cluster (generally under the conditions of shutdown or restart of a server in the cluster, failure disconnection or interruption, distance of a bearing platform from a network communication range and the like), reconstructing the cluster when finding that a node exits from the container cluster. The process is shown in fig. 8, and the detailed steps are as follows:
1) and (3) reporting the cluster node state: all the working nodes (including the management succession nodes) of the container cluster are timed (the interval is set as T) r ) Reporting the state information of the local working node to a cluster node state monitoring module of the management node according to the destination IP address, wherein the working state information comprises the local IP address and the local working state, and the working state is normal (caused by fault suspension, shutdown restart and the like), and then waiting for receiving the state report response of the management node;
2) cluster node exception assertion: the working state of the management node which receives the report of any working node is abnormal or overtime (equal to T) r ) When the working state of any working node can not be received, the corresponding working node (including the management successor node) exits the container cluster, otherwise, the state report response is replied to the corresponding working node according to the destination IP address, wherein the management successor node is overtime (equal to T) r ) When the response of the management node cannot be received, the management node is indicated to exit the container cluster;
3) the non-management succession working node exits the cluster reconstruction: and if the working node which is not managed and taken over in the container cluster exits, deleting the node to reconstruct the cluster. The method comprises the following specific steps:
1. and (4) deleting the exit node: the management node of the container cluster deletes the withdrawn non-management successor working node from the cluster node composition list through the cluster node composition management module;
2. the cluster nodes form backup: the management node sends a cluster node composition list to the management take-over node according to the destination IP address, and the cluster node composition backup handover module of the management take-over node receives and stores the list information of the backup.
4) And (3) the management succession node exits the cluster reconstruction: and if the management successor node in the container cluster exits, deleting the node and replacing the management successor node to reconstruct the cluster. The method comprises the following specific steps:
1. deleting the exit node: the container cluster management node deletes the withdrawn management successor node from the cluster node composition list through the cluster node composition management module;
2. the management successor node acknowledges: the selection of the calculation formula according to the preset strategy management successor node specifically comprises the following steps:
the compute learns the management successor nodes for the container cluster. Wherein N represents the number of working nodes in the container cluster; rand (1, N) represents that one of the working node lists is randomly selected; p s Indicating the position of the selected management successor node in the working node list, P s And 0 represents that the cluster (which is a stand-alone cluster) has no management successor node. At this time, if a management successor node is obtained, changing the role of the corresponding position node in the cluster node composition list into the management successor node, adjusting the node (if not at the secondary primary position) to the secondary primary position of the list, and then issuing all working nodes (including the management successor node) according to the destination IP address by using the management successor relationship of the container cluster, namely the management node and the management successor node information in the cluster node composition list;
3. managing the succession relationship for standby: when the node state report monitoring modules of all the working nodes receive the management succession relations of the management nodes, updating the IP address information of the container cluster management succession relations, namely the management nodes and the management succession nodes, recorded by the node for standby;
4. managing the configuration of the succession node: finding that the IP address of the local machine is consistent with the management successor node of the management successor relationship, indicating that the local machine is confirmed to be the management successor node of the cluster, setting the local machine to be the role of the management successor node at the moment, namely, the local machine is used as a node state report monitoring module required by a working node, and a cluster management control module group required by the role is loaded and configured and comprises a cluster node composition backup handover module (used for receiving cluster composition all node information sent by a record container cluster management node and transiting and handing over the cluster management node);
5. the cluster nodes form backup: the management node sends a cluster node composition list to the management take-over node according to the destination IP address, and the cluster node composition backup handover module of the management take-over node receives and stores the list information of the backup.
5) And (3) the management succession node exits the cluster reconstruction: and if the management node in the container cluster exits, replacing the management node to reconstruct the cluster. The method comprises the following specific steps:
1. the management node takes over: the management successor nodes of the container cluster change and set the roles of the management successor nodes as the management nodes, and then a backup handover module is formed by the cluster nodes to send management node successor notifications to all the non-management successor working nodes of the original cluster according to the destination IP addresses;
2. managing, taking over and handing over: after receiving the management node replacement notification, all the non-management replacement working nodes reply to a management replacement receipt confirmation to a backup handover module formed by cluster nodes of the existing replacement management node (the original management replacement node) according to the destination IP address when the node state report monitoring module confirms that the node state report monitoring module records the original management replacement node;
3. updating the management node list: after receiving the management receiving acknowledgments of all the working nodes in sequence, the present succession management node adds the corresponding node as the working node role to the cluster node composition list and records the IP address and role information of the node;
4. the management successor node confirms: the selection of the calculation formula according to the preset strategy management successor node specifically comprises the following steps:
the compute knows the management successor node for the container cluster. Wherein N represents a worker in a cluster of containersMaking the number of nodes; rand (1, N) represents that one of the working node lists is randomly selected; p s Indicating the position of the selected management successor node in the list of working nodes, P s And 0 represents that the cluster (which is a stand-alone cluster) has no management successor node. At this time, if a management successor node is obtained, changing the role of the corresponding position node in the cluster node composition list into the management successor node, adjusting the node (if not at the secondary primary position) to the secondary primary position of the list, and then issuing all working nodes (including the management successor node) according to the destination IP address by using the management successor relationship of the container cluster, namely the management node and the management successor node information in the cluster node composition list;
5. managing the succession relationship for standby: when the node state report monitoring modules of all the working nodes receive the management succession relations of the management nodes, updating the IP address information of the container cluster management succession relations, namely the management nodes and the management succession nodes, recorded by the node for standby;
6. managing the configuration of the succession node: finding that the IP address of the local machine is consistent with the management successor node of the management successor relationship, indicating that the local machine is confirmed to be the management successor node of the cluster, setting the local machine to be the role of the management successor node at the moment, namely, the local machine is used as a node state report monitoring module required by a working node, and a cluster management control module group required by the role is loaded and configured and comprises a cluster node composition backup handover module (used for receiving cluster composition all node information sent by a record container cluster management node and transiting and handing over the cluster management node);
7. the cluster nodes form backup: the management node sends a list of cluster nodes to the management take-over node according to the destination IP address, and a backup handover module of the cluster nodes of the management take-over node receives and stores the list information of the backup.
The invention provides a container cluster dynamic construction method oriented to a high-mobility environment, and a plurality of methods and ways for implementing the technical scheme, the above description is only a preferred embodiment of the invention, and it should be noted that, for those skilled in the art, a plurality of improvements and decorations can be made without departing from the principle of the invention, and these improvements and decorations should also be regarded as the protection scope of the invention. All the components not specified in the present embodiment can be realized by the prior art.
Claims (1)
1. A container cluster dynamic construction method oriented to a high-mobility environment is characterized by comprising the following steps:
when the container deployment server is started, dynamically constructing a container cluster according to the connection condition of adjacent servers in a network, or independently constructing a single-machine cluster or joining a multi-machine cluster, simultaneously dividing roles in the container cluster into a management node, a management succession node and a working node, and loading a corresponding cluster management control module group according to the roles;
after the container deployment server starts working, the dynamic adjustment of the container cluster is realized according to the connection change of adjacent servers in the network and the cooperation of the management control module groups among all the roles in the cluster, or a new server node is added, or a combined multi-computer cluster is added, or a server node is withdrawn to be reconstructed into a new single-computer cluster or a multi-computer cluster;
the container refers to a software entity which can be created, started, stopped, deleted and suspended when an image runs, wherein the image is a file system and comprises all contents required by the running of the container;
the network is a network formed by an internal network and an inter-network of a maneuvering platform to which the container deployment server belongs, wherein the internal network is connected in a wired mode, and the inter-network is connected in a wireless mode and conforms to an IP communication standard;
the working node is a service node which can only be assigned to execute tasks in the container cluster, the cluster management control module group which is correspondingly loaded by the working node comprises a node state report monitoring module, and the node state report monitoring module is used for reporting the working state of the local machine to the container cluster management node in real time and receiving the response of the management node;
the management node has the functions of a working node and also has the master control scheduling capability of a container cluster, the management node is provided with a correspondingly loaded cluster management control module which comprises a cluster node composition management module, a cluster node state monitoring module, a cluster expansion authentication module and an adjacent cluster searching module, wherein the cluster node composition management module is used for managing a cluster node composition list of the container cluster, and the cluster node composition list comprises IP addresses and role information of all nodes;
the cluster node state monitoring module is used for collecting the working states of all the working nodes in the container cluster in real time;
the cluster extension authentication module is used for monitoring the access of a new node of an authentication container cluster to the network and the synchronization of other container clusters;
the adjacent cluster searching module is used for searching other adjacent container clusters in the network in real time;
the management successor node is a working node, but can directly become a management node of a new cluster after the management node in the container cluster loses energy and loses contact, and the management successor node sets a corresponding loaded cluster management control module group, which comprises: the node status report monitoring module and the cluster node which are required by the working node form a backup handover module,
the cluster node composition backup handover module is used for receiving all node information of cluster composition issued by the cluster management node of the recording container and transitionally handing over the cluster management node;
when the container deployment server is started, dynamically constructing a container cluster according to the connection condition of adjacent servers in a network, or independently constructing a single-machine cluster or joining a multi-machine cluster, simultaneously dividing roles in the container cluster into a management node, a management succession node and a working node, and loading a corresponding cluster management control module group according to the roles, wherein the container deployment server specifically comprises:
step a1, detecting time range T of network adjacent environment smax In the local server at time sequence t s Sending a network access application to the adjacent sending node at an upper timing until receiving the network access permission or t of the adjacent container cluster management node cluster extended authentication module s ≥T smax : when receiving the network access permission, the local server node is authenticated and joins the cluster node of the adjacent cluster as the role of the working node to form a list to form a new multi-computer cluster, and the setting is carried out at the momentThe local server takes the role of a working node; otherwise, the local server is independently constructed into a single machine cluster and set as a management node role;
step a2, if the container cluster added by the local server is originally a stand-alone cluster: the management node of the newly constructed container cluster changes the role of the newly constructed network access node into a management take-over node through a cluster node composition management module, simultaneously sends the management node and the management take-over node information in a cluster node composition list to all working nodes, receives and updates the management node and the management take-over node information recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management take-over node changes the role of the working node into the role of the management take-over node;
step a3, the management node of the newly constructed container cluster sends the cluster node composition list to the management take-over node, and the backup handover module composed of the cluster nodes of the management take-over node receives the information of the backup storage cluster node composition list;
after the container deployment server starts working, according to connection change of adjacent servers in a network and cooperation of management control module groups among all roles in the cluster, the container cluster is dynamically adjusted, or a new server node is added, or a combined multi-computer cluster is added, or a server node is withdrawn to be reconstructed into a new single-computer cluster or a multi-computer cluster, and the method specifically comprises the following steps:
step b1, when monitoring and authenticating that there is a non-networked server node in the network, adding a new node into the container cluster;
b2, when searching and authenticating that other container clusters exist nearby in the network, according to a preset strategy, one container cluster is selectively added and merged to the other container cluster, and a new cluster is constructed in a combined manner;
step b3, reconstructing the container cluster as a new cluster when the state monitoring finds that the nodes in the container cluster exit;
step b1 includes:
b1-1, the cluster extension authentication module of the container cluster management node monitors the external server node network access application through the network, and verifies the system verification code attached to the application when the node network access application is monitored;
step b1-2, after the system verification code passes the authentication, adding the networking application node as a work node role to a cluster node composition list, recording the IP address and role information of the networking application node, simultaneously replying networking permission to the networking application node, and setting the networking application node as a work node role when receiving the networking permission;
step b1-3, if the container cluster before adding the node is originally a stand-alone cluster: the management node of the newly constructed container cluster changes the role of the newly constructed network access node into a management take-over node through a cluster node composition management module, simultaneously sends the management node and the management take-over node information in a cluster node composition list to all working nodes, receives and updates the management node and the management take-over node information recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management take-over node changes the role of the working node into the role of the management take-over node;
b1-4, the management node of the newly constructed container cluster sends the cluster node composition list to the management take-over node, and the backup handover module composed of the cluster nodes of the management take-over node receives the information of the backup storage cluster node composition list;
step b2 includes:
b2-1, the adjacent cluster searching module of the container cluster management node inquires whether other container clusters exist in the network at regular time, and submits the response to the cluster expansion authentication module to verify the system verification code attached to the response of other container clusters;
step b2-2, after the system verification code passes the authentication, starting container cluster merging, and calculating the cluster merging relationship according to the following formula:
wherein N is a 、N b Respectively representing the number of nodes formed by the container cluster A and the cluster B groupThe number of nodes; I.C. A a 、I b Respectively representing an IP address representation value of a management node of a container cluster A and an IP address representation value of a management node of a cluster B; g s 1 indicates that this cluster is a merge side, G s 0 means that the present cluster is the merging party;
b2-3, after the cluster merging relation is determined, the merge side container cluster management node obtains the cluster node composition list from the cluster node composition management module, and sends the cluster node composition list information to the adjacent cluster searching module of the merge side container cluster management node;
b2-4, when the container cluster management node of merger receives the cluster node composition list of container cluster of merger, all the nodes of merger are used as working nodes to be exchanged to the cluster node composition management module to merge the cluster node composition list, if the container cluster before merging is single cluster, the original management node of merger is changed into management substitution node, and simultaneously the management nodes and the management substitution node information in the cluster node composition list are issued to all the working nodes, the node state report monitoring module of all the working nodes receives and updates the management nodes and the management substitution node information recorded by the node, wherein the working node matched with the IP address of the management substitution node changes its own role to be set as the role of management substitution node;
b2-5, the management node of the newly constructed container cluster sends a cluster node composition list to a management take-over node, and a backup handover module composed of the cluster nodes of the management take-over node receives information of a backup storage cluster node composition list;
step b3 includes:
b3-1, all the working nodes of the container cluster report the working state of the machine to the cluster node state monitoring module of the management node at regular time, and correspondingly receive the response of the management node;
step b3-2, when the management node receives the working state of any working node is abnormal or the working state of any working node can not be received overtime, the management node indicates that the corresponding working node exits the container cluster; when the management successor node does not receive the response of the management node after overtime, the management node is indicated to exit the container cluster;
step b3-3, when the non-management successor working node in the container cluster exits, deleting the exited non-management successor working node to reconstruct the non-management successor working node as a new cluster;
step b3-4, when the management successor node in the container cluster exits, deleting the exited management successor node and replacing the management successor node to reconstruct the management successor node into a new cluster;
step b3-5, when the management node in the container cluster exits, replacing the management node to reconstruct as a new cluster;
wherein, step b3-3 includes:
b3-3-1, the container cluster management node deletes the quitted non-management successor work node from the cluster node composition list through the cluster node composition management module;
b3-3-2, the cluster node composition management module sends the modified cluster node composition list to the management succession node, and the backup handover module composed of the cluster nodes of the management succession node receives the information of the backup storage cluster node composition list;
step b3-4 includes:
step b3-4-1, the container cluster management node deletes the exited management successor node from the cluster node composition list through the cluster node composition management module;
step b3-4-2, calculating to obtain a new management successor node of the cluster according to the following preset strategy management successor node selection calculation formula:
wherein N represents the number of working nodes in the container cluster; rand (1, N) represents that one of all working nodes forming a list from cluster nodes is randomly selected; p s Indicating the position number, P, of the selected management successor node in all working nodes s 0 represents the node without management of the cluster;
step b3-4-3, after the new management succession node of the cluster is determined: changing the role of the corresponding node in the cluster node composition list into a management succession node, simultaneously sending the information of the management node and the management succession node in the cluster node composition list to all working nodes, receiving and updating the information of the management node and the management succession node recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management succession node changes the role of the working node into the role of the management succession node; the cluster node composition list is issued to a management take-over node, and a backup handover module is composed of the cluster nodes of the management take-over node and receives information of a backup storage cluster node composition list;
step b3-5 includes:
b3-5-1, changing the management succession node of the container cluster into a management node role, and forming a backup handover module through the cluster nodes to send a management node succession notice to all the non-management succession working nodes of the original cluster;
b3-5-2, after all non-management-successor working nodes receive the management node successor notification, when the original management successor node recorded by the node state report monitoring module is confirmed, the collection acknowledgement is confirmed by a backup handover module formed by cluster nodes of the current management node;
b3-5-3, after the management node receives the acknowledgement of all the working nodes in sequence, adding the corresponding nodes as the working node roles to the cluster node composition list and recording the IP addresses and role information of the nodes;
step b3-5-4, according to the preset strategy management successor node selection calculation formula in the step b3-4-2, calculating to obtain a new management successor node of the cluster: changing the role of the corresponding node in the cluster node composition list into a management succession node, simultaneously sending the information of the management node and the management succession node in the cluster node composition list to all working nodes, receiving and updating the information of the management node and the management succession node recorded by the node by a node state report monitoring module of all the working nodes, wherein the working node matched with the IP address of the management succession node changes the role of the working node into the role of the management succession node; and then, issuing the cluster node composition list to a management successor node, and receiving the information of the backup storage cluster node composition list by a backup handover module formed by the cluster nodes of the management successor node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110677750.2A CN113364633B (en) | 2021-06-18 | 2021-06-18 | Container cluster dynamic construction method oriented to high-mobility environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110677750.2A CN113364633B (en) | 2021-06-18 | 2021-06-18 | Container cluster dynamic construction method oriented to high-mobility environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113364633A CN113364633A (en) | 2021-09-07 |
| CN113364633B true CN113364633B (en) | 2022-09-06 |
Family
ID=77535067
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110677750.2A Active CN113364633B (en) | 2021-06-18 | 2021-06-18 | Container cluster dynamic construction method oriented to high-mobility environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113364633B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11953972B2 (en) | 2022-04-06 | 2024-04-09 | International Business Machines Corporation | Selective privileged container augmentation |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105635199A (en) * | 2014-10-28 | 2016-06-01 | 广州睿哲网络科技有限公司 | Method and device for implementation of self-organization cluster server supporting load balancing |
| CN110647580A (en) * | 2019-09-05 | 2020-01-03 | 南京邮电大学 | Distributed container cluster mirror image management main node, slave node, system and method |
-
2021
- 2021-06-18 CN CN202110677750.2A patent/CN113364633B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105635199A (en) * | 2014-10-28 | 2016-06-01 | 广州睿哲网络科技有限公司 | Method and device for implementation of self-organization cluster server supporting load balancing |
| CN110647580A (en) * | 2019-09-05 | 2020-01-03 | 南京邮电大学 | Distributed container cluster mirror image management main node, slave node, system and method |
Non-Patent Citations (1)
| Title |
|---|
| 容器化部署调度策略的分析与改进;陈雪门;《电信快报》;20200810(第08期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113364633A (en) | 2021-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111290834B (en) | Method, device and equipment for realizing high service availability based on cloud management platform | |
| CN109151045B (en) | Distributed cloud system and monitoring method | |
| CN108632067B (en) | Disaster recovery deployment method, device and system | |
| CN116599972B (en) | Edge cloud aggregation and dispersion method oriented to high-mobility environment | |
| WO2012071920A1 (en) | Method, system, token conreoller and memory database for implementing distribute-type main memory database system | |
| CN104935672A (en) | High available realizing method and equipment of load balancing service | |
| WO2016177231A1 (en) | Dual-control-based active-backup switching method and device | |
| CN102394914A (en) | Cluster brain-split processing method and device | |
| CN110445697B (en) | Video big data cloud platform equipment access service method | |
| CN105721582A (en) | Multi-node file backup system | |
| WO2011082581A1 (en) | Method and system for managing configuration of network management data | |
| CN103036934A (en) | Wide area network clustered deployment system and method based on mirror image | |
| CN113364633B (en) | Container cluster dynamic construction method oriented to high-mobility environment | |
| CN110391940A (en) | Response method, device, system, equipment and the storage medium of address of service | |
| CN115576655B (en) | Container data protection system, method, device, equipment and readable storage medium | |
| CN113515316A (en) | Novel edge cloud operating system | |
| US20240097979A1 (en) | Fabric availability and synchronization | |
| CN105323271B (en) | Cloud computing system and processing method and device thereof | |
| CN109274532B (en) | Method, device and system for issuing policy, centralized control equipment and readable storage medium | |
| CN111767166A (en) | Data backup method and device | |
| CN107483257B (en) | Application system deployment method and architecture based on X86 and ARM mixed environment | |
| CN116723077A (en) | Distributed IT automatic operation and maintenance system | |
| CN114840495A (en) | Database cluster split-brain prevention method, storage medium and device | |
| CN112948177A (en) | Disaster recovery backup method and device, electronic equipment and storage medium | |
| CN102281159A (en) | Recovery method for cluster system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |