IN-QUEUE JOBS INFORMATION MONITORING AND FILTERING

• US 20060259349A1
• Filed: 07/25/2006
• Published: 11/16/2006
• Est. Priority Date: 11/06/2001
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Lam, Thanh V., Lauria, Giampaolo
• Application Number:
  US11/459,709
• Publication Number:
  US 20060259349A1
• Time in Patent Office: Days
• US Class Current: 705/10
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/0631 Resource planning, allocation, distributing or scheduling for enterprises or organisations
• CPC Class Codes: G06Q 10/06314 Calendaring for a resource

Abstract

A remote monitor technique based on a client/server model is provided for a distributed computing system having a plurality of computing nodes. The remote monitor includes a server application for collecting jobs information on jobs in a queue of the system and a client application for presenting or employing the jobs information. The client application can display the jobs information in a single display window comprising multiple subwindows for various characteristics of the jobs information. Jobs information displayed by the client application can be periodically refreshed. The remote monitor enables a user to monitor in detail jobs of a state defined by the user. The remote monitor can also track usage of preselected filesystems and automatically trims filesystems if usage is higher than a defined threshold.

Claims

1. A computer-implemented method of monitoring jobs in a queue of a distributed system, said method comprising:

responsive to a request by a client monitor application for jobs information on computer jobs in a queue having a user selected state of execution, wherein computer jobs in the queue are in any one of a plurality of possible states of execution, collecting by a monitoring server application jobs information on a set of computer jobs in the queue of a distributed system having the user selected state, said jobs information comprising JobTracker information for the set of computer jobs, said JobTracker information including jobs data on each computer job of the set of computer jobs having the user selected, tracked job state and, for each job having said tracked job state, a time of entering the tracked job state; and

employing or displaying, by the client monitor application, said jobs information for the set of computer jobs in the queue having the user selected state.

2. The method of claim 1, further comprising user selecting said tracked job state from a predefined list of job states, said predefined list comprising at least some states of “

waiting”

, “

running”

, “

pending” and

“

held”

.

3. The method of claim 1, wherein said jobs information comprises multiple characteristics, said multiple characteristics including said JobTracker information, and at least one of jobs data, filesystem information, cluster information and core files information, said jobs data comprising for each job in a queue at least some of job id, job owner, job starting time, job state, job priority, job class, and name of node where the job is running, said filesystem information including filesystem usage, said cluster information including for each node of said system, a name of the node, status of a scheduler, number of in-queue jobs, number of active jobs, average load, idle time, type of node architecture and operating system running on the node, and said core files information including a list of core files.

4. The method of claim 1, wherein said collecting jobs information further comprises:

reading a set of nodes of said system;

querying nodes of said set of nodes on said jobs information; and

retrieving from said system results of said querying.

5. The method of claim 4, wherein said reading comprises:

retrieving from said system a full set of nodes of said system; and

removing down nodes from said full set of nodes, wherein said down nodes include nodes that fail to respond to operating system commands within a predefined period of time.

6. The method of claim 4, wherein said querying comprises at least some of:

querying nodes in said set of nodes on jobs information;

querying nodes in said set of nodes on JobTracker information;

querying nodes in said set of nodes on filesystem information;

querying nodes in said set of nodes on cluster information; and

querying nodes in said set of nodes on core files information.

7. The method of claim 4, wherein said querying nodes on JobTracker information comprises:

selecting said tracked job state from a predefined list of job states; and

producing a list of jobs in a queue of said system having said tracked job state.

8. The method of claim 7, wherein said producing a list of jobs comprises:

retrieving jobs information on at least one job in a queue and storing said jobs information into a list, wherein said jobs information includes a current timestamp and job state on said at least one job;

removing from said list jobs having a state different from said tracked job state; and

periodically updating said list by adding jobs that acquire said tracked job state after a most recent update of said list, and by removing from said list jobs that change job state after the most recent update of said list.

9. The method of claim 1, wherein said employing or displaying said jobs information comprises at least one of storing said jobs information, using said jobs information as an input of a computer program or a script, or presenting said jobs information in a single display window for viewing thereof.

10. A computer-implemented monitor system for monitoring jobs in a queue of a distributed system, said monitor system comprising:

responsive to a request by a client monitor application for jobs information on computer jobs in a queue having a user selected state of execution, wherein computer jobs in the queue are in any one of a plurality of possible states of execution, means for collecting by a monitoring server application jobs information on a set of computer jobs in the queue of a distributed system having the user selected state, said jobs information comprising JobTracker information for the set of computer jobs, said JobTracker information including jobs data on each computer job of the set of computer jobs having the user selected, tracked job state and, for each job having said tracked job state, a time of entering the tracked job state; and

means for employing or displaying, by the client monitor application, said jobs information for the set of computer jobs in the queue having the user selected state.

11. The monitor system of claim 10, further comprising means for user selecting said tracked job state from a predefined list, said predefined list comprising at least some states of “

waiting”

, “

running”

, “

pending” and

“

held”

.

12. The monitor system of claim 10, wherein said jobs information comprises multiple characteristics, said multiple characteristics including said JobTracker information, and at least one of jobs data, filesystem information, cluster information and core files information, said jobs data comprising for each job in a queue at least some of job id, job owner, job starting time, job state, job priority, job class, and name of node where the job is running, said filesystem information including filesystem usage, said cluster information including for each node of said system, a name of the node, status of a scheduler, number of in-queue jobs, number of active jobs, average load, idle time, type of node architecture and operating system running on the node, and said core files information including a list of core files.

13. The monitor system of claim 10, wherein said means for collecting jobs information further comprises:

means for reading a set of nodes of said system;

means for querying nodes of said set of nodes on said jobs information; and

means for retrieving from said system results of said means for querying.

14. The monitor system of claim 13, wherein said means for reading comprises:

means for retrieving from said system a full set of nodes of said system; and

means for removing down nodes from said full set of nodes, wherein said down nodes include nodes that fail to respond to operating system commands within a predefined period of time.

15. The monitor system of claim 13, wherein said means for querying comprises at least some of:

means for querying nodes in said set of nodes on jobs information;

means for querying nodes in said set of nodes on JobTracker information;

means for querying nodes in said set of nodes on filesystem information;

means for querying nodes in said set of nodes on cluster information; and

means for querying nodes in said set of nodes on core files information.

16. The monitor system of claim 13, wherein said means for querying comprises:

means for selecting said tracked job state from a predefined list of job states; and

means for producing a list of jobs in a queue of said system having said tracked job state.

17. The monitor system of claim 16, wherein said means for producing a list of jobs comprises:

means for retrieving jobs information on at least one job in a queue and means for storing said jobs information into a list, wherein said jobs information includes a current timestamp and job state on said at least one job;

means for removing from said list jobs having a state different from said tracked job state; and

means for periodically updating said list by adding jobs that acquire said tracked job state after a most recent update of said list, and by removing from said list jobs that change job state after the most recent update of said list.

18. The monitor system of claim 10, wherein said means for employing or displaying said jobs information comprises at least one of means for storing said jobs information, means for using said jobs information as an input of a computer program or a script, or means for presenting said jobs information in a single display window for viewing thereof.

19. At least one program storage device readable by a computer, tangibly embodying at least one program of instructions executable by the computer to perform a method of monitoring jobs in a queue of a distributed system, comprising:

responsive to a request by a client monitor application for jobs information on computer jobs in a queue having a user selected state of execution, wherein computer jobs in the queue are in any one of a plurality of possible states of execution, collecting by a monitoring server application jobs information on a set of computer jobs in the queue of a distributed system having the user selected state, said jobs information comprising JobTracker information for the set of computer jobs, said JobTracker information including jobs data on each computer job of the set of computer jobs having the user selected, tracked job state and, for each job having said tracked job state, a time of entering the tracked job state; and

employing or displaying, by the client monitor application, said jobs information for the set of computer jobs in the queue having the user selected state.