CN-122019274-A - Container fault self-healing method for underwater acoustic signal processor containerization platform

CN122019274ACN 122019274 ACN122019274 ACN 122019274ACN-122019274-A

Abstract

The invention provides a container fault self-healing method of a container platform of an underwater sound signal processor, which comprises the steps of obtaining hardware index parameters and service index parameters of any first node in the process of completing an underwater sound signal processing task, determining abnormal probability of the first node based on the hardware index parameters and the service index parameters, determining a second node based on a pre-built node label registry when the abnormal probability is higher than an early warning threshold, preferentially inquiring the node which is the same as the first node cluster attribution label as the second node, configuring the same cluster attribution label by the nodes in the same physical domain, synchronously forwarding an underwater sound signal flow input into the first node to the second node, loading a Pod mirror image and a configuration file at the second node, and processing the underwater sound signal flow in a time-sequence cache queue based on the Pod mirror image and the configuration file at the second node when the first node is detected to be faulty. By implementing the invention, the fault self-healing capacity can be improved, and the instantaneity of the underwater sound signal can be ensured.

Inventors

Chen Shuaike
SU SHUAI
XU JIALIN
ZHAO SHUAI

Assignees

中国舰船研究设计中心

Dates

Publication Date: 20260512
Application Date: 20260408

Claims (10)

1. The container fault self-healing method of the underwater acoustic signal processor containerization platform is characterized by comprising the following steps of: Acquiring a hardware index parameter and a service index parameter of any first node in the process of completing the underwater acoustic signal processing task; Determining the abnormal probability of the first node based on the hardware index parameter and the service index parameter; When the abnormal probability is higher than the early warning threshold value, determining a second node based on a pre-constructed node label registry, preferentially inquiring the node which is the same as the cluster attribution label of the first node and is healthy and idle as the second node, and configuring the same cluster attribution label by the nodes in the same physical domain; Synchronously forwarding the underwater sound signal stream input into the first node to a time-sequence buffer queue of the second node, and loading the underwater sound signal stream of the first node at the second node to process the Pod mirror image and the configuration file; And when the first node is detected to be faulty, processing the Pod mirror image and the configuration file based on the preloaded underwater sound signal flow at the second node, and processing the underwater sound signal flow in the time-sequence buffer queue.
2. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor of claim 1, wherein synchronously forwarding an underwater acoustic signal stream input to a first node to a clocked cache queue of a second node comprises: Segmenting an input underwater sound signal stream according to a processing time sequence, and marking a sequence label to generate a data segment with the sequence label; recording check point information of the processing completion degree of the first node on each data segment while storing the data segments into a second node cache queue, and synchronizing the check point information to the second node; When the first node is detected to be faulty, processing the Pod mirror image and the configuration file based on the preloaded underwater sound signal flow at the second node, and processing the underwater sound signal flow in the time-sequence buffer queue, wherein the processing comprises the following steps: when the first node is detected to be faulty, the second node starts processing from the data segment corresponding to the sequence tag in the time-sequence cache queue according to the received latest check point information, and discards the data segment which is completely processed by the first node.
3. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor of claim 1, wherein determining the second node based on a pre-built node tag registry comprises: determining whether a target node which is the same as the first node hardware type label, the same as the cluster attribution label and healthy and idle exists in a node label registry; when a plurality of target nodes exist, evaluating the plurality of target nodes based on a pre-constructed evaluation system to obtain evaluation values of the target nodes; Selecting a preset number of candidate target nodes based on the evaluation value; Transmitting a reference test task to each candidate target node, wherein the reference test task is a key operation for simulating the first node to process the underwater sound signal flow; Task completion data for each candidate target node is received and a second node is determined among each candidate target node based on the task completion data.
4. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor according to claim 1, wherein determining the anomaly probability of the first node based on the hardware index parameter and the traffic index parameter comprises: Based on the hardware index parameter and the service index parameter, determining the self-abnormality probability of the first node by adopting a pre-trained abnormality detection model; Acquiring a node propagation map in a physical domain corresponding to a first node; determining an upstream node of the first node based on the node propagation map; Acquiring the current state of each upstream node; determining propagation anomaly probabilities of the first node based on the current states of the upstream nodes; The anomaly probability of the first node is determined based on the self anomaly probability of the first node and the propagation anomaly probability.
5. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor of claim 1, further comprising: acquiring an output result of the underwater sound signal flow in the second node processing time-sequence buffer queue and a historical output result of the underwater sound signal flow before the first node processing fault occurs; Extracting first characteristic information representing a continuous acoustic event from an output result of the second node; extracting second characteristic information representing continuous acoustic events from the historical output result of the first node; Performing time sequence correlation analysis on the first characteristic information and the second characteristic signal to obtain a continuity analysis result; When the continuity analysis result does not meet the preset requirement, backtracking and locking an original data segment interval with time sequence fracture, and selecting a third node to process the original data segment interval according to a preset rule to obtain a processing result of the third node; And carrying out association fusion on the third node processing result and the underwater sound signal stream output result in the second node processing time-sequence buffer queue to obtain the underwater sound signal stream processing result.
6. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor of claim 1, wherein synchronously forwarding an underwater acoustic signal stream input to a first node to a clocked cache queue of a second node comprises: when the abnormal probability is lower than the target threshold but higher than the early warning threshold, the underwater sound signal stream forwarded to the second node pair time-sequence buffer queue is a signal stream coded by adopting a lossy coding mode with a high compression ratio; When the abnormal probability is higher than the target threshold, the second node encodes the underwater sound signal stream forwarded by the time-sequence buffer queue in a lossless encoding mode.
7. The method for self-healing a container failure of a containerized platform of an underwater acoustic signal processor of any of claims 1-6, further comprising: analyzing a hardware log, a container running log and a signal processing service log of a first node, and positioning a fault source; if the fault source is a hardware fault, marking the first node as a to-be-maintained state, and sending a hardware maintenance alarm to the target port; If the fault source is a software fault, the Pod image and the configuration file are redeployed at the first node.
8. The container fault self-healing device of the underwater acoustic signal processor containerization platform is characterized by comprising: the parameter acquisition module is used for acquiring hardware index parameters and service index parameters of any first node in the process of completing the underwater acoustic signal processing task; the probability determining module is used for determining the abnormal probability of the first node based on the hardware index parameter and the service index parameter; The node determining module is used for determining a second node based on a pre-constructed node tag registry when the abnormal probability is higher than the early warning threshold value, preferentially inquiring the node which is the same as the first node cluster attribution tag and is healthy and idle as the second node, and configuring the same cluster attribution tag by the nodes in the same physical domain; the forwarding module is used for synchronously forwarding the underwater sound signal flow input into the first node to the time-sequence buffer queue of the second node, and loading the underwater sound signal flow of the first node into the second node to process the Pod mirror image and the configuration file; And the fault processing module is used for processing the Pod mirror image and the configuration file based on the preloaded underwater sound signal flow at the second node and processing the underwater sound signal flow in the time-sequence buffer queue when the first node is detected to be faulty.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor performs the steps of a container failure self-healing method of a hydroacoustic signal processor containerization platform according to any of the preceding claims 1-7.
10. A computer storage medium having stored thereon computer instructions which, when executed by a processor, perform the steps of a container fault self-healing method of a hydroacoustic signal processor containerisation platform according to any of claims 1 to 7.

Description

Container fault self-healing method for underwater acoustic signal processor containerization platform Technical Field The invention belongs to the technical field of fault treatment, and particularly relates to a container fault self-healing method of a container platform of an underwater acoustic signal processor. Background With the technical development of the fields of ocean exploration, underwater communication and the like, the underwater acoustic signal processor is used as core equipment and is widely deployed on mobile platforms such as ships, underwater Unmanned Vehicles (UUV) and the like to bear key tasks such as acquisition, filtering, spectrum analysis, target identification and the like of underwater acoustic signals. The mobile platform has extremely strict requirements on real-time performance and reliability of underwater acoustic signal processing, on one hand, the original underwater acoustic signal has large data volume and strong timeliness, the primary pretreatment is required to be completed within microsecond delay, signal attenuation and delay superposition caused by transmission across hardware modules are avoided, otherwise, the synchronization of target recognition precision and underwater communication is influenced, on the other hand, the mobile platform faces complicated interference of marine environment, equipment vibration, power supply fluctuation and the like in the navigation process, the container or hardware node carrying signal processing tasks is easy to fail, and service is required to be quickly restored to avoid service interruption. The existing container fault self-healing technology is mainly developed towards a general IT scene, and has a gap with a local scheduling scene of a container platform of the underwater acoustic signal processor, and because the underwater acoustic signal generally requires continuous processing, the existing technology is difficult to meet the real-time requirement of recovery after the underwater acoustic signal processing fails. Disclosure of Invention Therefore, the invention aims to provide a container fault self-healing method of a container platform of an underwater acoustic signal processor, so as to meet the real-time requirement of recovery after the underwater acoustic signal processing is failed. In order to achieve the above purpose, the present invention provides the following technical solutions: according to a first aspect, the invention provides a container fault self-healing method of a container platform of an underwater sound signal processor, which comprises the steps of obtaining hardware index parameters and business index parameters of any first node in the process of completing an underwater sound signal processing task, determining abnormal probability of the first node based on the hardware index parameters and the business index parameters, determining a second node based on a pre-built node label registry when the abnormal probability is higher than an early warning threshold, preferentially inquiring nodes which are the same as the first node cluster attribution label and are healthy and idle as the second node, configuring the same cluster attribution label by the nodes in the same physical domain, synchronously forwarding the underwater sound signal flow input into the first node to a time sequence buffer queue of the second node, loading the underwater sound signal flow of the first node to process Pod mirror images and configuration files at the second node, and processing the underwater sound signal flow in the time sequence buffer queue based on the preloaded underwater sound signal flow when the first node is detected to fail. According to a second aspect, the embodiment provides a container fault self-healing device of a container platform of an underwater sound signal processor, which comprises a parameter acquisition module, a probability determination module, a node determination module, a forwarding module and a fault processing module, wherein the parameter acquisition module is used for acquiring hardware index parameters and service index parameters of any first node in the process of completing an underwater sound signal processing task, the probability determination module is used for determining the abnormal probability of the first node based on the hardware index parameters and the service index parameters, the node determination module is used for determining a second node based on a pre-built node label registry when the abnormal probability is higher than an early warning threshold, preferentially querying nodes which are the same as the cluster attribution labels of the first node and are healthy and idle as the second node, the nodes in the same physical domain are configured with the same cluster attribution labels, the forwarding module is used for synchronously forwarding the underwater sound signal flow input into a time-ordered buffer queue of the second node, loading the