Dynamically scaling out pods using a recursive way
In an approach for pod scheduling and recursion, a processor schedules a master pod and one or more worker pods for applications to be deployed on a cluster. A processor builds a topology between the master pod and the one or more worker pods. A processor monitors a workload in the one or more worker pods. A processor determines whether any of the one or more worker pods needs to scale out. In response to determining that one of the one or more worker pods needs to scale out, a processor schedules a next layer of the one or more worker pods according to the workload.
1. A computer-implemented method comprising:
scheduling, by one or more processors, a first master pod and one or more worker pods for applications be deployed on a cluster;
building, by one or more processors, a topology between the first master pod and the one or more worker pods;
monitoring, by one or more processors, a workload in the one or more worker pods; and
dynamically scaling out, by one or more processors, the first master pod and the one or more worker pods using a recursive way for multiple layers of the first master pod and the one or more worker pods, wherein dynamically scaling out the first master pod and the one or more worker pods using the recursive way comprises:
determining, based on the workload exceeding a pre-defined threshold, that one of the one or more worker pods needs to scale out,
converting any one of the one or more worker pods into a second master pod,
scheduling a next layer of the one or more worker pods according to the workload, and
scheduling one or more new worker pods associated with the second master pod.
2. The computer-implemented method of claim 1 , further comprising:
in response to determining that none of the one or more worker pods needs to scale out, processing, by one or more processors, the workload; and
returning, by one or more processors, a processing result to the first master pod.
3. The computer-implemented method of claim 1 , wherein the topology is selected from the group consisting of: tree, star, net, and mesh topology.
4. A computer program product comprising:
one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising:
program instructions to schedule a first master pod and one or more worker pods for applications be deployed on a cluster;
program instructions to build a topology between the first master pod and the one or more worker pods;
program instructions to monitor a workload in the one or more worker pods; and
program instructions to dynamically scale out the first master pod and the one or more worker pods using a recursive way for multiple layers of the first master pod and the one or more worker pods, wherein program instructions to dynamically scale out the first master pod and the one or more worker pods using the recursive way comprise:
program instructions to determine, based on the workload exceeding a pre-defined threshold, that one of the one or more worker pods needs to scale out,
program instructions to convert any one of the one or more worker pods into a second master pod,
program instructions to schedule a next layer of the one or more worker pods according to the workload, and
program instructions to schedule one or more new worker pods associated with the second master pod.
5. The computer program product of claim 4 , further comprising:
program instructions, stored on the one or more computer-readable storage media, in response to determining that none of the one or more worker pods needs to scale out, to process the workload; and
program instructions, stored on the one or more computer-readable storage media, to return a processing result to the first master pod.
6. The computer program product of claim 4 , wherein the topology is selected from the group consisting of: tree, star, net, and mesh topology.
7. A computer system comprising:
one or more computer processors, one or more computer readable storage media, and program instructions stored on the one or more computer readable storage media for execution by at least one of the one or more computer processors, the program instructions comprising:
program instructions to schedule a first master pod and one or more worker pods for applications be deployed on a cluster;
program instructions to build a topology between the first master pod and the one or more worker pods;
program instructions to monitor a workload in the one or more worker pods; and
program instructions to dynamically scale out the first master pod and the one or more worker pods using a recursive way for multiple layers of the first master pod and the one or more worker pods, wherein program instructions to dynamically scale out the first master pod and the one or more worker pods using the recursive way comprise:
program instructions to determine, based on the workload exceeding a pre-defined threshold, that one of the one or more worker pods needs to scale out,
program instructions to convert any one of the one or more worker pods into a second master pod,
program instructions to schedule a next layer of the one or more worker pods according to the workload, and
program instructions to schedule one or more new worker pods associated with the second master pod.
8. The computer system of claim 7 , further comprising:
program instructions, stored on the one or more computer-readable storage media, in response to determining that none of the one or more worker pods needs to scale out, to process the workload; and
program instructions, stored on the one or more computer-readable storage media, to return a processing result to the first master pod.
9. The computer system of claim 7 , wherein the topology is selected from the group consisting of: tree, star, net, and mesh topology.