Cluster arrangement

• US 8,301,599 B2
• Filed: 12/19/2005
• Issued: 10/30/2012
• Est. Priority Date: 06/18/2003
• Status: Expired due to Fees

• Current Assignee: Siemens It Solutions and Services GmbH (Atos SE), Fujitsu Siemens Computers GmbH (Fujitsu Limited)
• Original Assignee: Siemens It Solutions and Services GmbH (Atos SE), Fujitsu Siemens Computers GmbH (Fujitsu Limited)
• Inventors: Friess, Bernhard, Kraus, Rudi, Nawabi, Grana
• Primary Examiner(s): MENG, JAU SHYA
• Application Number:
  US11/312,209
• Publication Number:
  US 20060161637A1
• Time in Patent Office: 2,507 Days
• Field of Search: None
• US Class Current: 707/636
• CPC Class Codes: G06F 11/1482 by means of middleware or OS functionality
• CPC Class Codes: G06F 11/203 using migration
• CPC Class Codes: G06F 9/4856 resumption being on a different machine, e.g. task migration, virtual machine migration G06F9/5088 takes precedence
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/101 based on network conditions
• CPC Class Codes: H04L 67/1023 based on a hash applied to IP addresses or costs
• CPC Class Codes: H04L 67/1029 using data related to the state of servers by a load balancer
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]
• CPC Class Codes: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06

Abstract

A cluster arrangement having a first network, at least two data processing systems which each form an application node and which each have an operating system, and an application agent on each application node, which application agent is operable to monitor and identify entities executed on the application node; start a new entity or restart an entity which has been terminated prematurely on the application node; assess and decide whether it is possible to execute a new entity on the application node; send a request for execution of an entity to application agents at application nodes connected to the network; and report to application agents at application nodes connected to the network following an acceptance of the request for execution of an entity.

Claims

1. A cluster arrangement of a system, comprising:

a first network;

a plurality of data processing systems, each of said plural data processing systems forming an application node and having an operating system which is the operating system in each of said plural data processing systems, each said application node including an entity which is executed;

an inspection node connected to the first network, the inspection node including the operating system and an inspection agent configured to manage reports received from each application node; and

an application agent on the each application node, the application node comprising a service of the operating system and being started following initialization of the operating system on the each application node, the application agent being configured to;

monitor and identify entities executed on the each application node;

start a new entity or restart an executed entity which has been terminated prematurely on the each application node;

assess and decide whether the new or restarted entity can be executed on the each application node;

forward a request to execute the new or restarted entity to the application agents at all of the application nodes connected to the first network;

report to the application agents at all of the application nodes connected to the network following an acceptance of the request to execute the new or restarted entity; and

report to the application agents at all of the application nodes connected to the network if an application agent rejects the request to execute the new or restarted entity.

2. The cluster arrangement of the system as claimed in claim 1, wherein the application agent on each application node is monitored to compile a list which respectively contains the entities executed on the each application node and all data required to execute the new or restarted executed entity as list element.

3. The cluster arrangement of the system as claimed in claim 2, wherein the list includes information about the each application node.

4. The cluster arrangement of the system as claimed in claim 2, wherein the execution request comprises a compiled list element from the each application agent.

5. The cluster arrangement of the system as claimed in claim 1, wherein monitoring the application agent comprises identifying dependencies of the entities executed on the application node on at least one of other entities and parameters.

6. The cluster arrangement of the system as claimed in claim 1, wherein the application agent includes a function which is configured to identify an unsafe operating state of the entity executed on the application node.

7. The cluster arrangement of the system as claimed in claim 1, wherein the application agent includes a function which terminates the entity executed on the application node.

8. The cluster arrangement of the system as claimed in claim 1, wherein a memory device is connected to the first network and the entity which is executed on an application node.

9. The cluster arrangement of the system as claimed in claim 8, wherein the memory device stores an application node operating system.

10. The cluster arrangement of the system as claimed in claim 1, wherein the application agent at each application node comprises a service of the operating system running on the each application node.

11. The cluster arrangement of the system as claimed in claim 1, wherein the inspection agent is configured to:

check operability of the application nodes which are connected to the first network;

forward the request to execute the new or restarted entity to the application agents at the application nodes connected to the network; and

determine whether an application node is located within the cluster arrangement and place the request to execute the new or restarted entity at this application node if it is located within the cluster arrangement.

12. The cluster arrangement of the system as claimed in claim 11, wherein said inspection agent being configured to use lists compiled by the application agents at the application nodes to determine which of the application nodes to check.

13. The cluster arrangement of the system as claimed in claim 1, wherein at least one of a compiled list, the execution request and the report is stored as a file in a memory device.

14. The cluster arrangement of the system as claimed in claim 1, wherein each entity stored in a memory device includes a unique identifier which is allocated to the each application node where the entity is executed.

15. The cluster arrangement of the system as claimed in claim 1, wherein the first network is configured to communicate using Transmission Control Protocol Internet Protocol (TCP/IP) or Network File System (NFS) protocol.

16. The cluster arrangement of the system as claimed in claim 1, wherein the inspection node includes means configured to control an initialization operation at the application nodes, the initialization operation comprising a command to load an operating system from a memory device to the each application node.

17. The cluster arrangement of the system as claimed in claim 1, wherein a plurality of application nodes are connected to a second network which is connected to a client computer.

18. The cluster arrangement of the system as claimed in claim 1, wherein entities stored in a memory device include a database.

19. The cluster arrangement of the system as claimed in claim 1, wherein a data processing system of said plural data processing systems includes an associated identifier.

20. The cluster arrangement of the system as claimed in claim 19, wherein each of said plural data processing systems in the cluster arrangement of the system having the same identifier from a pool.

21. The cluster arrangement of the system as claimed in claim 19, wherein the application agent at an application node is configured to:

evaluate the identifier associated with the each application node;

forward the request to execute the new or restarted entity to the application agents on the application nodes having the same identifier; and

report to application agents on the application nodes having the same identifier following acceptance of the request to execute the new or restarted entity.

22. The cluster arrangement of the system as claimed in claim 19, wherein the identifier comprises an IP address or part of an IP address.

23. The cluster arrangement of the system as claimed in claim 19, wherein each of said plural data processing systems defines the inspection node which includes an associated identifier and the inspection agent executed on the inspection node is configured to check a function of the application nodes having the same identifier.

24. The cluster arrangement of the system as claimed in claim 19, wherein the entity executed on each application node includes an associated priority and the application agent is configured to evaluate the priority of the following functions:

assessment, decision, and the request to execute the new or restarted entity.

25. The cluster arrangement of the system as claimed in claim 19, wherein the inspection agent at the inspection node is configured to evaluate a priority of the entity executed on each application node upon assessing, deciding and forwarding the request to execute this entity.

26. The cluster arrangement of the system as claimed in claim 1, wherein the application agent is further configured to one of (i) identify the premature termination of a monitored entity and (ii) continue to maintain correct operation by terminating an erroneous entity, and terminate the entity being executed on the application node.

27. A method in a cluster arrangement comprising a plurality of application nodes and an inspection node including a device for controlling an initialization operation at each of said plurality of application nodes, each of said plurality of application nodes and the inspection node being connected to a network, the method comprising the steps of:

receiving at the inspection node from each of said plurality of application nodes a first list including all entities executed on the each of said plurality of application nodes and data required to execute the entities, the initialization operation including a command for loading the operating system of the application node to each of said plurality of application nodes;

checking with the inspection node, at regular intervals of time, to determine whether any of said plurality of application nodes has failed;

if any of said plurality of application nodes has failed, compiling with the inspection node a second list of entities executed on the failed application node and the data required to execute the entities executed on the failed application node, and forwarding the second list to all of said plurality of application nodes connected to the network with a request to execute the entities on the forwarded second list;

assessing and deciding, at each of said plurality of application nodes, whether to start a new entity or restart a failed entity upon receipt of the request to execute an entity on the forwarded second list; and

reporting to application agents at all of said plurality of application nodes if an application agent rejects the request to execute the entity on the forwarded second list.

28. The method as claimed in claim 27, further comprising:

forwarding the compiled second list of entities executed on the failed application node together with the execution request to at least one additional application node if any of said plural application nodes has failed.

29. The method as claimed in claim 27, wherein an application node compiles the first list including all the entities executed on the each of said plural application nodes and sends the execution request with the compiled first list to at least one additional application node.

30. The method as claimed in claim 27, wherein if any of said plural application nodes has failed, the inspection node ascertains an additional application node using the first lists and settable parameters, and the inspection node transmits the request to execute the entities executed on the failed application node to said additional application node.

31. The method as claimed in claim 27, wherein the inspection node ascertains which of said plural application nodes are to be checked by evaluating the first lists received from the plural application agents.

32. The method as claimed in claim 27, wherein the method is performed by an inspection agent executed on the inspection node and an application agent executed on each of the application nodes.

33. The method as claimed in claim 27, further comprising:

allocating an identifier to each of said plurality of application nodes and the inspection node;

wherein the inspection agent at the inspection node receives a report from an application agent of each of said plurality of application nodes having the same identifier.

34. The method as claimed in claim 27, further comprising:

allocating a priority to an executed entity; and

formatting, at an inspection agent, a priority associated with the entity executed on the failed application node and sending the execution request based on the priority if any of said plurality of application nodes has failed.