Shared memory based monitoring for application servers

• US 20060143525A1
• Filed: 12/28/2004
• Published: 06/29/2006
• Est. Priority Date: 12/28/2004
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Kilian, Frank
• Granted Patent: US 7,562,138 B2
• Time in Patent Office: Days
• US Class Current: 714/31
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 2201/86 Event-based monitoring
• CPC Class Codes: G06F 2201/875 Monitoring of systems including the internet
• CPC Class Codes: G06F 9/505 considering the load

Abstract

A system and method to monitor application servers using shared monitoring memory. An application server includes worker nodes having Java virtual machines to process work requests received by the application server. Status information is generated for each of the worker nodes while processing the work requests. The status information for each of the worker nodes is then stored into the shared monitoring memory while the worker nodes operate.

Claims

1. A method, comprising:

operating at least two worker nodes within an application server (“

AS”

) instance, each of the worker nodes including a Java virtual machine (“

JVM”

) to process work requests received by the AS instance;

generating status information for each of the worker nodes during processing of the work requests; and

storing the status information for each of the worker nodes into a shared monitoring memory during operation of the worker nodes.

2. The method of claim 1, wherein storing the status information for each of the worker nodes into the shared monitoring memory comprises storing the status information from each of the worker nodes into topic buffers of the shared monitoring memory to organize the status information into logical topics.

3. The method of claim 2, wherein each topic buffer includes at least one slot corresponding to each of the worker nodes to store topically related portions of the status information from a corresponding one of the worker nodes.

4. The method of claim 3, wherein the topic buffers include a session buffer to store session information related to sessions active on each of the worker nodes, each slot within the session buffer corresponding to a different one of the sessions.

5. The method of claim 3, wherein the topic buffers include a thread buffer to store thread information related to threads active on each of the worker nodes, each slot within the thread buffer corresponding to a different one of the threads.

6. The method of claim 3, wherein the topic buffers include a virtual machine (“

VM”

) buffer to store general information related to internal operation of each of the JVMs within each of the worker nodes, each slot within the VM buffer corresponding to a different one of the worker nodes.

7. The method of claim 6, wherein the general information related to internal operation of each of the JVMs includes at least one of a scarce memory indication and a work request overload indication.

8. The method of claim 1, further comprising:

terminating one of the worker nodes;

copying the status information corresponding to the terminated one of the worker nodes into a log file; and

reclaiming portions of the shared monitoring memory occupied by the status information corresponding to the terminated one of the worker nodes.

9. The method of claim 1, further comprising:

retrieving the status information from the shared monitoring memory; and

transmitting the status information to a management console to display the status information.

10. A machine-accessible medium that provides instructions that, if executed by a machine, will cause the machine to perform operations comprising:

receiving work requests at an application server (“

AS”

) instance;

servicing the work requests with at least two worker nodes each operating a Java virtual machine (“

JVM”

);

generating status information for each of the worker nodes in response to servicing the work requests; and

storing the status information for each of the worker nodes into a shared monitoring memory during operation of the worker nodes.

11. The machine-accessible medium of claim 10, wherein storing the status information for each of the worker nodes into the shared monitoring memory comprises storing the status information from each of the worker nodes into topic buffers of the shared monitoring memory to organize the status information into logical topics.

12. The machine-accessible medium of claim 11, wherein each topic buffer includes at least one slot corresponding to each of the worker nodes to store topically related portions of the status information from a corresponding one of the worker nodes.

13. The machine-accessible medium of claim 12, wherein the topic buffers include at least one of:

a session buffer to store session information related to sessions active on each of the worker nodes, each slot within the session buffer corresponding to a different one of the sessions;

a thread buffer to store thread information related to threads active on each of the worker nodes, each slot within the thread buffer corresponding to a different one of the threads; and

a virtual machine (“

VM”

) buffer to store VM information related to operation of each of the JVMs within each of the worker nodes, each slot within the VM buffer corresponding to a different one of the worker nodes.

14. The machine-accessible medium of claim 10, wherein the status information includes at least one of a scarce memory indication and work request overload indication.

15. The machine-accessible medium of claim 10, wherein storing the status information into the shared monitoring memory is event driven.

16. A system, comprising:

a server node to execute an application server (“

AS”

) instance, the AS instance including logic executable by a processor of the server node to;

operate at least two worker nodes within an application server (“

AS”

) instance, each of the worker nodes including a Java virtual machine (“

JVM”

) to process work requests received by the AS instance;

generate status information for each of the worker nodes during processing of the work requests; and

store the status information for each of the worker nodes into a shared monitoring memory during operation of the worker nodes.

17. The system of claim 16, wherein the logic to store the status information into the shared monitoring memory is event driven to store the status information into the shared monitoring memory in response to generating the status information.

18. The system of claim 17, wherein the shared monitoring memory includes topic buffers to organize the status information from each of the worker nodes into logical topics and wherein the topic buffers each include slots configured to store a portion of the status information from a corresponding one of the worker nodes and wherein the slots are grouped into the logical topics.

19. The system of claim 18, wherein the logic to generate the status information includes logic to one of:

generate session information related to active work sessions on each of the worker nodes during execution of the work requests;

generate thread information related to active threads on each of the worker nodes during execution of the work requests; and

generate virtual machine information related to operation of the JVMs during execution of the work requests.

20. The system of claim 16, further comprising a client node to execute a management console, the management console including logic executable by a processor of the client node to:

send a status query to the server node;

receive the status information from the server node; and

display the status information.