Disaster resilient federated kubernetes operator

Patent metadata

Field	Value
Publication number	US-11656960-B2
Application number	US-202117205245-A
Country	US
Kind code	B2
Filing date	Mar 18, 2021
Priority date	Mar 18, 2021
Publication date	May 23, 2023
Grant date	May 23, 2023

Abstract

Disclosed herein are system, method, and computer program product embodiments for disaster resilience of applications managed by Kubernetes operators. An embodiment operates by creating an orchestration and worker cluster, where the worker cluster is coupled to the orchestration cluster by a proxy server. Custom resources are deployed to the orchestration cluster and custom resource controllers are deployed to the worker cluster. The proxy server federates these custom resources between the orchestration cluster and the worker cluster. During disasters, the worker cluster is recreated and reconciled to prevent loss of the federated cluster.

First claim

What is claimed is:

1. A computer implemented method for cloud processing, comprising:
creating, by a cloud system, an orchestration cluster, wherein the orchestration cluster includes at least a first application programming interface (API) server;
deploying, by the cloud system, custom resources to the orchestration cluster;
creating, by the cloud system, a worker cluster, wherein the worker cluster includes at least a proxy server for watching the orchestration cluster, wherein the orchestration cluster does not know about the worker cluster, and wherein the proxy server is configured to detect, through the watching, a change to metadata associated with the worker cluster;
deploying, by the cloud system, resource controllers and a second API server to the worker cluster;
executing, by the proxy server, federation of the custom resources between the orchestration cluster and the worker cluster;
determining that the worker cluster has become unavailable and the orchestration cluster remains operational, wherein the orchestration cluster includes Kubernetes resources defining the unavailable worker cluster;
recreating the worker cluster responsive to determining that the worker cluster has become unavailable;
initiating port forwarding to the first API server to provide communications between the recreated worker cluster and the orchestration cluster; and
registering the resource controllers with the proxy server to reconcile the custom resources;
wherein at least one of the creating, deploying, and executing are performed by one or more computers.

2. The computer implemented method of claim 1 , wherein the first API server comprises a Kubernetes API server.

3. The computer implemented method of claim 1 , further comprising:
establishing port-forwarding from the proxy server to the first API server.

4. The computer implemented method of claim 1 , wherein the proxy server comprises a Kubernetes API proxy server.

5. The computer implemented method of claim 1 , wherein the resource controllers comprise Kubernetes resource operators.

6. The computer implemented method of claim 5 , further comprising:
registering the Kubernetes resource operators with the first API server.

7. A system, comprising:
a memory; and
at least one processor coupled to the memory and configured to:
create an orchestration cluster;
deploy custom resources to the orchestration cluster;
create a worker cluster coupled to the orchestration cluster, the worker cluster including watching the orchestration cluster, wherein the orchestration cluster does not know about the worker cluster, and wherein the proxy server is configured to detect, through the watching, a change to metadata associated with the worker cluster;
deploy resource operators to the worker cluster; and
federate the custom resources between the orchestration cluster and the worker cluster;
determine that the worker cluster has become unavailable and the orchestration cluster remains operational, wherein the orchestration cluster includes Kubernetes resources defining the unavailable worker cluster;
recreate the worker cluster responsive to determining that the worker cluster has become unavailable;
initiate port forwarding to the first API server to provide communications between the recreated worker cluster and the orchestration cluster; and
register the resource controllers with the proxy server to reconcile the custom resources.

8. The system of claim 7 , the at least one processor further configured to:
deploy a first application programming interface (API) server to the orchestration cluster.

9. The system of claim 8 , wherein the first API server comprises a Kubernetes API server.

10. The system of claim 8 , wherein the at least one processor is further configured to:
deploy a second API server to the worker cluster.

11. The system of claim 7 , wherein the at least one processor is further configured to:
deploy a proxy server to the worker cluster, wherein the proxy server provides the communications with the orchestration cluster.

12. The system of claim 11 , wherein the proxy server comprises a Kubernetes application programming interface (API) proxy server.

13. The system of claim 11 , wherein the at least one processor is further configured to: register the resource operators with the proxy server to reconcile the custom resources.

14. A non-transitory computer-readable device having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:
creating an orchestration cluster;
deploying custom resources to the orchestration cluster;
creating a worker cluster, wherein the worker cluster includes at least a proxy server for watching the orchestration cluster, wherein the orchestration cluster does not know about the worker cluster, and wherein the proxy server is configured to detect, through the watching, a change to metadata associated with the worker cluster;
deploying custom resource controllers to the worker cluster;
executing federation of the custom resources between the orchestration cluster and the worker cluster;
determining that the worker cluster has become unavailable and the orchestration cluster remains operational, wherein the orchestration cluster includes Kubernetes resources defining the unavailable worker cluster;
recreating the worker cluster responsive to determining that the worker cluster has become unavailable;
initiating port forwarding to the first API server to provide communications between the recreated worker cluster and the orchestration cluster; and
registering the resource controllers with the proxy server to reconcile the custom resources.

15. The non-transitory computer-readable device of claim 14 , the operations further comprising, in response to a disaster related loss of the worker cluster:
recreating the worker cluster;
initiating communications between the orchestration cluster and the worker cluster; and
registering the custom resource controllers with the worker cluster to reconcile the custom resources.

16. The non-transitory computer-readable device of claim 14 , the operations further comprising:
for the communications, establishing a port-forwarding from the worker cluster to the orchestration cluster.

17. The non-transitory computer-readable device of claim 14 , the operations further comprising:
for the communications, establishing a port-forwarding from a proxy server of the worker cluster to the orchestration cluster.

18. The non-transitory computer-readable device of claim 14 , the operations further comprising:
registering the custom resource controllers with the orchestration cluster.

Assignees

• Sap Se

Inventors

• Auch Nathan
• Bowman Ivan
• Kirmse Daniel
• Zahn Henning

Classifications

• H04L67/10 Primary

  in which an application is distributed across nodes in the network (software deployment G06F8/60; multiprogramming arrangements G06F9/46) · CPC title

• H04L67/56

  Provisioning of proxy services (store-and-forward switching systems in data switching networks H04L12/54) · CPC title

• G06F11/2025 Primary

  using centralised failover control functionality · CPC title

• G06F9/5072

  Grid computing · CPC title

• G06F2201/805

  Real-time · CPC title