Automatic online system upgrade

• US 9,792,110 B2
• Filed: 12/21/2015
• Issued: 10/17/2017
• Est. Priority Date: 12/21/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Baset, Salman A., Dai, Bao Hua, Tao, Shu, Tao, Tao
• Primary Examiner(s): Kang, Insun
• Application Number:
  US14/976,725
• Publication Number:
  US 20170177323A1
• Time in Patent Office: 666 Days
• Field of Search: None
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 8/656 while running

Abstract

Automatically upgrading a computing environment system may include automatically identifying a set of timeframes and nodes running user applications on physical machines, containers, or virtual machines (VMs) whose disruption during the identified timeframes minimally impact the user applications. The timeframes may be intelligently determined by leveraging the monitoring data obtained automatically and/or the hints supplied by the user.

Claims

1. A computer-implemented method of automatically upgrading a computing environment system, comprising:

determining, by one or more hardware processors, a candidate timeframe that requires a least number of compute nodes for supporting an overall service performance, the determining the candidate timeframe comprising determining which applications, virtual machines and containers are active and which are passive in a high availability cluster at different times;

for the candidate timeframe, determining by one or more of the hardware processors, number of requests the compute nodes are processing during the candidate timeframe;

storing by one or more of the hardware processors, identifiers of the compute nodes in a queue on a memory device in an order of least number of requests the compute nodes are processing during the candidate timeframe, wherein the compute nodes are ordered by the number of requests a compute node has for processing during the candidate timeframe and wherein multiple compute nodes occupy an entry within the queue;

selecting by one or more of the hardware processors, the compute nodes from the queue that are available to upgrade in the order of the least number of requests, wherein the multiple compute nodes occupying the entry selected from the queue are upgraded in parallel; and

upgrading by one or more of the hardware processors, the compute nodes until the candidate timeframe expires,

wherein the least number of nodes is determined based on user supplied hints comprising affinity of the entities, anti-affinity of the entities, and whether the entities are co-located or spread across the compute nodes.

2. The method of claim 1, further comprising:

if more of the compute nodes remain in the queue that have not been upgraded in the candidate time frame, using a next candidate timeframe to upgrade the compute nodes that remain.

3. The method of claim 1, wherein the least number of compute nodes is determined by automatically correlating data from one or more monitoring systems.

4. The method of claim 3, wherein automatically correlating data comprises determining whether one or more compute nodes are running one or more entities comprising one or more of an application, virtual machine or container with no network traffic, whether one or more compute nodes are running the entities with a threshold amount of network traffic, one or more compute nodes or entities running on compute nodes with a threshold level of central processing unit processing identified as low, one or more compute nodes or entities with a threshold level of memory processing identified as low, one or more compute nodes with a threshold level of network processing identified as low, one or more compute nodes with a threshold level of physical machine disk utilization, one or more compute nodes with a threshold level of virtual machine disk utilization, one or more compute nodes with a threshold level of container disk utilization, and compute nodes running entities that are communicating among one another.

5. The method of claim 1, wherein the user supplied hints further comprises scheduling of entities comprising one or more of an application, virtual machine, or container on the compute nodes, one or more active entities running on compute nodes in a high availability cluster, and one or more passive entities running on compute nodes in a high availability cluster.

6. A system for automatically upgrading a computing environment, comprising:

one or more hardware processors; and

a memory device coupled to one or more of the hardware processors,

one or more of the hardware processors configured to determine a candidate timeframe that requires a least number of compute nodes for supporting an overall service performance, by at least determining which applications, virtual machines and containers are active and which are passive in a high availability cluster at different times,

for the candidate timeframe, one or more of the hardware processors further configured to determine number of requests the compute nodes are processing during the candidate timeframe,

one or more of the hardware processors further configured to insert identifiers of the compute nodes in a queue on the memory device in an order of least number of requests the compute nodes are processing during the candidate timeframe, wherein the compute nodes are ordered by the number of requests a compute node has for processing during the candidate timeframe and wherein multiple compute nodes occupy an entry within the queue,

one or more of the hardware processors further configured to select the compute nodes from the queue that are available to upgrade in the order of the least number of request, wherein the multiple compute nodes occupying the entry selected from the queue are upgraded in parallel,

one or more of the hardware processors further configured to upgrade the compute nodes until the candidate timeframe expires,

wherein the least number of nodes is determined based on user supplied hints comprising affinity of the entities, anti-affinity of the entities, and whether the entities are co-located or spread across the compute nodes.

7. The system of claim 6, wherein if more of the compute nodes remain in the queue that have not been upgraded in the candidate time frame, one or more of the hardware processors are further configured to upgrade the compute nodes that remain in a next candidate timeframe.

8. The system of claim 6, wherein the least number of compute nodes is determined by automatically correlating data from one or more monitoring systems.

9. The system of claim 8, wherein automatically correlating data comprises determining whether one or more compute nodes are running one or more entities comprising one or more of an application, virtual machine or container with no network traffic, whether one or more compute nodes are running entities with a threshold amount of network traffic, one or more compute nodes or entities running on compute nodes with a threshold level of central processing unit processing identified as low, one or more compute nodes or entities with a threshold level of memory processing identified as low, one or more compute nodes with a threshold level of network processing identified as low, one or more compute nodes with a threshold level of physical machine disk utilization, one or more compute nodes with a threshold level of virtual machine disk utilization, one or more compute nodes with a threshold level of container disk utilization, compute nodes running entities that are communicating among one another, or combinations thereof.

10. The system of claim 6, wherein the user supplied hints further comprises scheduling of entities comprising one or more of an application, virtual machine, or container on the compute nodes, one or more active entities running on compute nodes in a high availability cluster, and one or more passive entities running on compute nodes in a high availability cluster.

11. A non-transitory computer readable storage medium storing a program of instructions executable by a machine to perform a method of automatically upgrading a computing environment system, the method comprising:

determining a candidate timeframe that requires a least number of compute nodes for supporting an overall service performance, the determining the candidate timeframe comprising determining which applications, virtual machines and containers are active and which are passive in a high availability cluster at different times;

for the candidate timeframe, determining number of requests the compute nodes are processing during the candidate timeframe;

storing identifiers of the compute nodes in a queue on a memory device in an order of least number of requests the compute nodes are processing during the candidate timeframe wherein the compute nodes are ordered by the number of requests a compute node has for processing during the candidate timeframe and wherein multiple compute nodes occupy an entry within the queue;

selecting the compute nodes from the queue that are available to upgrade in the order of the least number of requests, wherein the multiple compute nodes occupying the entry selected from the queue are upgraded in parallel; and

upgrading the compute nodes until the candidate timeframe expires,

wherein the least number of nodes is determined based on user supplied hints comprising affinity of the entities, anti-affinity of the entities, and whether the entities are co-located or spread across the compute nodes.

12. The computer readable storage medium of claim 11, further comprising:

if more of the compute nodes remain in the queue that have not been upgraded in the candidate time frame, using a next candidate timeframe to upgrade the compute nodes that remain.

13. The computer readable storage medium of claim 11, wherein the least number of compute nodes is determined by one or more of, automatically correlating data from one or more monitoring systems and user supplied hints,

wherein automatically correlating data comprises determining whether one or more compute nodes are running one or more entities comprising one or more of an application, virtual machine or container with no network traffic, whether one or more compute nodes are running one or more entities comprising one or more of an application, VM, or container with a threshold amount of network traffic, one or more compute nodes or entities running on compute nodes with a threshold level of central processing unit processing identified as low, one or more compute nodes or entities with a threshold level of memory processing identified as low, one or more compute nodes with a threshold level of network processing identified as low, one or more compute nodes with a threshold level of physical machine disk utilization, one or more compute nodes with a threshold level of virtual machine disk utilization, one or more compute nodes with a threshold level of container disk utilization, compute nodes running entities that are communicating among one another, or combinations thereof,

wherein the user supplied hints comprises one or more of scheduling of entities on the compute nodes, affinity of the entities, anti-affinity of the entities, whether the entities are co-located or spread across the compute nodes, one or more active entities running on compute nodes in a high availability cluster, or one or more passive entities running on compute nodes in a high availability cluster, or combinations thereof.