Method and system for data backup and restoration in cluster system
Embodiments of the present disclosure provide a method, a system and a computer readable storage medium for data backup and restoration in a cluster system. The method comprises obtaining performance data of the nodes, the performance data being collected periodically by the backup and restore proxy modules; predicting workloads of the nodes based on the performance data; selecting, based on the predicted workloads, a backup node from the nodes, to take over a data backup and restore task from a working node which is currently in charge of the data backup and restore task.
1. A method of data backup and restoration at a manager processing node in a cluster system, the manager processing node being embodied as a computerized device, the method comprising:
configuring a plurality of processing nodes in the cluster system with a plurality of backup and restore proxy processing modules, respectively, each of the plurality of processing nodes being embodied as a respective computerized device, the plurality of backup and restore proxy processing modules executing on the plurality of processing nodes, respectively;
periodically collecting performance data of the plurality of processing nodes by the plurality of backup and restore proxy processing modules, respectively, the performance data including information about data that has been backed up and restored currently and information about data that is to be backed up and restored;
predicting, by the manager processing node, workloads of the plurality of processing nodes based on the performance data collected by the plurality of backup and restore proxy processing modules, respectively, the plurality of processing nodes including a working node that is currently in charge of a data backup and restore task;
selecting, by the manager processing node based on the predicted workloads, a backup node from among the plurality of processing nodes to take over the data backup and restore task from the working node,
periodically creating, by the manager processing node, a checkpoint record for recording the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored;
in response to detecting a request for node switching between the working node and the backup node, reading the checkpoint record by the manager processing node; and
obtaining, by the manager processing node, the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored.
2. The method according to claim 1 , wherein predicting workloads of the plurality of processing nodes comprises:
predicting the workloads based on an exponential average algorithm (EMA).
3. The method according to claim 1 , wherein the backup node is a local node with respect to a processing node whose data is backed up and restored.
4. The method according to claim 1 , wherein obtaining the performance data comprises obtaining data associated with at least one of the following:
usage of processing units of the respective processing nodes,
usage of memories of the respective processing nodes, and
I/O conditions of the respective processing nodes.
5. A system, comprising:
a cluster system including a plurality of processing nodes, wherein each processing node from among the plurality of processing nodes in the cluster system is embodied as a respective computerized device and configured with a backup and restore proxy processing module executing on the processing node;
wherein each respective processing node from among the plurality of processing nodes is configured to include computer-executable program logic encoded in memory of one or more computers for execution of:
periodically collecting performance data of the respective processing node by the backup and restore proxy processing module configured within the respective processing node, the performance data including information about data that has been backed up and restored currently and information about data that is to be backed up and restored, and
wherein at least a manager processing node from among the plurality of processing nodes is configured to include computer-executable program logic encoded in memory of one or more computers for the execution of:
predicting a workload of each respective processing node from among the plurality of processing nodes based on the performance data collected by the backup and restore proxy processing module configured within the respective processing node, the plurality of processing nodes including a working node that is currently in charge of a data backup and restore task;
selecting, based on the predicted workload of each of the plurality of processing nodes, a backup node from among the plurality of nodes to take over the data backup and restore task from the working node,
periodically creating a checkpoint record for recording the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored;
in response to detecting a request for node switching between the working node and the backup node, reading the checkpoint record; and
obtaining the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored.
6. The system according to claim 5 , wherein predicting workloads of the respective processing nodes comprises:
predicting the workloads based on an exponential average algorithm (EMA).
7. The system according to claim 5 , wherein the backup node is a local node with respect to a processing node whose data is backed up and restored.
8. The system according to claim 5 , wherein obtaining the performance data comprises obtaining data associated with at least one of the following:
usage of processing units of the respective processing nodes,
usage of memories of the respective processing nodes, and
I/O conditions of the respective processing nodes.
9. A computer program product for data backup and restoration at a manager processing node in a cluster system, the manager processing node being embodied as a computerized device, the computer program product comprising:
a non-transitory computer readable medium encoded with computer-executable code, the code configured to enable the execution of:
configuring a plurality of processing nodes in the cluster system with a plurality of backup and restore proxy processing modules, respectively, each of the plurality of processing nodes being embodied as a respective computerized device, the plurality of backup and restore proxy processing modules executing on the plurality of processing nodes, respectively;
periodically collecting performance data of the plurality of processing nodes by the backup and restore proxy processing modules, respectively, the performance data including information about data that has been backed up and restored currently and information about data that is to be backed up and restored;
predicting, by the manager processing node, workloads of the plurality of processing nodes based on the performance data collected by the plurality of backup and restore proxy processing modules, respectively, the plurality of processing nodes including a working node that is currently in charge of a data backup and restore task;
selecting, by the manager processing node based on the predicted workloads, a backup node from among the plurality of processing nodes to take over the data backup and restore task from the working node,
periodically creating, by the manager processing node, a checkpoint record for recording the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored;
in response to detecting a request for node switching between the working node and the backup node, reading the checkpoint record by the manager processing node; and
obtaining, by the manager processing node, the information about the data that has been backed up and restored currently and the information about the data that is to be backed up and restored.
10. The computer program product according to claim 9 , wherein predicting workloads of the plurality of processing nodes comprises:
predicting the workloads based on an exponential average algorithm (EMA).
11. The computer program product according to claim 9 , wherein the backup node is a local node with respect to a processing node whose data is backed up and restored.
12. The computer program product according to claim 9 , wherein obtaining the performance data comprises obtaining data associated with at least one of the following:
usage of processing units of the respective processing nodes,
usage of memories of the respective processing nodes, and
I/O conditions of the respective processing nodes.