Cluster instance management system

Patent metadata

Field	Value
Publication number	US-10361928-B2
Application number	US-201715682397-A
Country	US
Kind code	B2
Filing date	Aug 21, 2017
Priority date	Mar 20, 2015
Publication date	Jul 23, 2019
Grant date	Jul 23, 2019

Abstract

A system for cluster management comprises a status monitor and an instance replacement manager. The status monitor is for monitoring status of an instance of a set of instances on a cluster provider. The instance replacement manager is for determining a replacement strategy for the instance in the event the instance does not respond. The replacement strategy for the instance is based at least in part on a management criteria for on-demand instances and spot instances on the cluster provider.

First claim

What is claimed is:

1. A system for cluster management, comprising:
a processor; and
a memory coupled with the processor, wherein the memory is configured to provide the processor with instructions which when executed cause the processor to:
monitor status of an instance of a set of instances on a cluster provider; and
determine a replacement strategy for the instance in response to a determination the instance does not respond, wherein the replacement strategy for the instance is based at least in part on a management criteria for on-demand instances and spot instances on the cluster provider, and wherein the determining of the replacement strategy comprises to:
in response to a determination that the set of instances is to be stopped:
determine whether to maintain a master on-demand instance; and
in response to a determination to maintain the master on-demand instance, stop the set of instances except the master on-demand instance.

2. A system as in claim 1 , wherein the processor is further configured to:
receive the management criteria.

3. A system as in claim 2 , wherein the management criteria comprises one or more of the following: a maximum number of on-demand instances or a minimum number of on-demand instances.

4. A system as in claim 2 , wherein the management criteria comprises one or more of the following: a maximum number of total instances or a minimum number of total instances.

5. A system as in claim 2 , wherein the management criteria comprises a budget limit.

6. A system as in claim 2 , wherein the management criteria comprises a failure criterion.

7. A system as in claim 2 , wherein the management criteria comprises a reserve criterion.

8. A system as in claim 2 , wherein the management criteria comprises a reserve poolsize.

9. A system as in claim 1 , wherein the processor is further configured to:
determine a new set of instances to request based at least in part on the management criteria.

10. A system as in claim 9 , wherein the processor is further configured to:
determine availability of the new set of instances on the cluster provider; and
in response to a determination that there is availability, indicate to create the new set of instances.

11. A system as in claim 9 , wherein the determining of the replacement strategy comprises to:
determine a request strategy based at least in part on a cluster availability and on the management criteria.

12. A system as in claim 1 , wherein the monitoring of the status of the instance includes to monitor the status of the instance periodically.

13. A system as in claim 1 , wherein the monitoring of the status of the instance includes to monitor a status of each instance of the set of instances.

14. A system as in claim 1 , wherein the monitoring of the status of the instance includes to monitor a status of each spot instance of the set of instances.

15. A system as in claim 1 , wherein the replacement strategy comprises replacing the instance with a spot instance from a reserve pool.

16. A system as in claim 1 , wherein the replacement strategy comprises replacing the instance with an on-demand instance in response to a determination that this is allowed under the management criteria.

17. A system as in claim 1 , wherein the replacement strategy comprises monitoring for a replacement instance.

18. A system as in claim 1 , wherein the replacement strategy comprises stopping the set of instances.

19. A system as in claim 18 , wherein the replacement strategy comprises saving the state of the set of instances.

20. A system as in claim 18 , wherein the replacement strategy comprises maintaining one on-demand instance of the set of instances.

21. A method for cluster management, comprising:
monitoring status of an instance of a set of instances on a cluster provider; and
determining, using a processor, a replacement strategy for the instance in response to a determination the instance does not respond, wherein the replacement strategy for the instance is based at least in part on a management criteria for on-demand instances and spot instances on the cluster provider, and wherein the determining of the replacement strategy comprises:
in response to a determination that the set of instances is to be stopped:
determining whether to maintain a master on-demand instance; and
in response to a determination to maintain the master on-demand instance, stopping the set of instances except the master on-demand instance.

22. A computer program product for cluster management, the computer program product being embodied in a non-transitory computer readable storage medium and comprising computer instructions for:
monitoring status of an instance of a set of instances on a cluster provider; and
determining, a replacement strategy for the instance in response to a determination the instance does not respond, wherein the replacement strategy for the instance is based at least in part on a management criteria for on-demand instances and spot instances on the cluster provider, and wherein the determining of the replacement strategy comprises:
in response to a determination that the set of instances is to be stopped:
determining whether to maintain a master on-demand instance; and
in response to a determination to maintain the master on-demand instance, stopping the set of instances except the master on-demand instance.

Assignees

• Databricks Inc

Inventors

• Ghodsi Ali
• Stoica Ion
• Zaharia Matei

Classifications

• H04L41/5051 Primary

  Service on demand, e.g. definition and deployment of services in real time · CPC title

• G06F11/30

  Monitoring · CPC title

• H04L41/5096

  wherein the managed service relates to distributed or central networked applications · CPC title

• H04L43/0817

  by checking functioning · CPC title

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title