Cluster arrangement

• US 20060161637A1
• Filed: 12/19/2005
• Published: 07/20/2006
• Est. Priority Date: 06/18/2003
• Status: Active Grant

• Current Assignee: Siemens It Solutions and Services GmbH (Atos SE), Fujitsu Siemens Computers GmbH (Fujitsu Limited)
• Original Assignee: Fujitsu Siemens Computers GmbH (Fujitsu Limited), Siemens AG
• Inventors: Nawabi, Grana, Friess, Bernhard, Kraus, Rudi
• Granted Patent: US 8,301,599 B2
• Time in Patent Office: Days
• US Class Current: 709/218
• 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 (NW1), at least two data processing systems which each form an application node (AK) and which each have an operating system (OS), and an application agent (AA) on each application node (AK), which has the following functions: monitoring (UB) and identifying entities (I1, I2, I3) executed on the application node (AK); starting (ST) a new entity (I3) or restarting an entity which has been terminated prematurely on the application node; assessing (BE) and deciding whether it is possible to execute a new entity (I3) on the application node (AK); putting a request (AF) for execution of an entity to application agents (AA) at application nodes connected to the network (NW1); and reporting (ME) to application agents (AA) at application nodes (AK) connected to the network (NW1) following an acceptance of the request (AF) for execution of an entity (I3).

Claims

1. A cluster arrangement comprising:

a first network (NW1);

at least two data processing systems which each form an application node (AK) and which each have an operating system (OS), where an application node (AK) has an entity (I1, I2, I3, I4) which is executed; and

an application agent (AA) on each application node (AK), which has the following functions;

monitoring (UB) and identifying entities (I1, I2, I3) executed on the application node (AK);

starting (ST) a new entity (I3) or restarting an entity which has been terminated prematurely on the application node;

assessing (BE) and deciding whether it is possible to execute a new entity (I3) on the application node (AK);

putting a request (AF) for execution of an entity to application agents (AA) at application nodes connected to the network (NW1); and

reporting (ME) to application agents (AA) at application nodes (AK) connected to the network (NW1) following an acceptance of the request (AF) for execution of an entity (I3).

2. The cluster arrangement as claimed in claim 1,

wherein

the function of monitoring (UB) the application agent comprises the compilation (L) of a list (T) which respectively contains the entities (I1, I2) executed on the application node (AK) and all the data (D) required for executing the executed entity as list elements (TI1, TI2).

3. The cluster arrangement as claimed in claim 1,

wherein

the function of monitoring (UB) the application agent (AA) is designed to identify dependencies of entities executed on the application node on other entities and/or parameters.

4. The cluster arrangement as claimed in claim 1,

wherein

the application agent (AA) has a function which is designed to identify an unsafe operating state of an entity.

5. The cluster arrangement as claimed in claim 1,

wherein

the application agent (AA) has a function for terminating an entity executed on the application node.

6. The cluster arrangement as claimed in claim 2,

wherein

the list (T) also has information about the application node.

7. The cluster arrangement as claimed in claim 1,

wherein

a memory device SP is connected to the first network and has the at least one entity (I1) which can be executed on an application node (AK).

8. The cluster arrangement as claimed in claim 1,

wherein

the application agent at the application node is a service of the operating system (OS) running on the application node.

9. The cluster arrangement as claimed in claim 7,

wherein

the memory device (SP) stores the operating system (OS) for the application nodes (AK).

10. The cluster arrangement as claimed in claim 1,

wherein

an application node (AK′

) is provided on which it is possible to execute a new entity (I3).

11. The cluster arrangement as claimed in claim 1,

wherein

the cluster arrangement has at least one data processing system which is in the form of an inspection node (KK), which is connected to the first network and in which the inspection node (KK) has an operating system and an inspection agent (KA) which has the following functions:

checking (UP) the operability of the application nodes (AK) which are connected to the first network (NW1);

putting a request (AF) for execution of an entity to application agents (AA) at application nodes (AK) connected to the network (NW1); and

determining (BS) an application node (AK) and putting a request for execution of a new entity to this application node.

12. The cluster arrangement as claimed in claim 11,

wherein

when checking (UP) the application nodes (AK), it is possible to ascertain the application nodes (AK) to be checked using the lists (T) compiled by the application agents (AA) at the application nodes (AK).

13. The cluster arrangement as claimed in claim 2,

wherein

the request (AF) for execution comprises a compiled list element (TI1, TI2) from an application agent (AA).

14. The cluster arrangement as claimed in claim 1,

wherein

the compiled list (T) and/or the request (AF) for execution and/or the report (ME) is/are stored as at least one file in the memory device (SP).

15. The cluster arrangement as claimed in claim 1,

wherein

each entity stored in the memory device (SP) has a unique identifier which is allocated to the application node (AK) on which the entity is executed.

16. The cluster arrangement as claimed in claim 1,

wherein

the first network (NW1) is designed for communication with the TCP/IP or with the NFS protocol.

17. The cluster arrangement as claimed in claim 1,

wherein

the at least one inspection node (KK) has means for controlling an initialization operation at the application nodes (AK), with the initialization operation comprising a command for loading the operating system (OS) of the application node (AK).

18. The cluster arrangement as claimed in claim 1,

wherein

the at least two application nodes (AK, AK′

) are connected to a second network (NW2) which is connected to a client computer (CL).

19. The cluster arrangement as claimed in claim 1,

wherein

the entities stored in the memory device (SP) have a database.

20. The cluster arrangement as claimed in claim 1,

wherein

at least one data processing system has an associated identifier (IP1, IP2).

21. The cluster arrangement as claimed in claim 20,

wherein

the data processing systems (DV) in the cluster arrangement with the same identifier (IP1, IP2) form a pool (VC1, VC2, VC3).

22. The cluster arrangement as claimed in claim 20,

wherein

the application agent (AA) at an application node (AK) has the following functions:

evaluation of the identifier (IP1, IP2) associated with the application node (AK);

putting a request for execution of an entity to application agents (AA) on application nodes (AK) with the same identifier (IP1, IP2); and

reporting to application agents (AA) on application nodes (AK) with the same identifier following acceptance of a request for execution of an entity.

23. The cluster arrangement as claimed in claim 20,

wherein

the identifier (IP1, IP2) comprises of an IP address or part of an IP address.

24. The cluster arrangement as claimed in claim 20,

wherein

a data processing system (DV) in the form of an inspection node (KK1) has an associated identifier (IP2) and the inspection agent (KA) executed on the inspection node (KK1) is designed to check the function of the application nodes (AK) with the same identifier.

25. The cluster arrangement as claimed in claim 20,

wherein

an entity (I1, I2, L1, L2, L3) executed on an application node (AK) has an associated priority (PS1, PS2) and the application agent (AA) is designed to evaluate the priority (PS1, PS2) for the following functions:

assessment, decision, and the request for execution of this entity.

26. The cluster arrangement as claimed in claim 20,

wherein

the inspection agent (KA) at an inspection node (KK1) is designed to evaluate the priority (PS1, PS2) of an entity executed on an application node (AK) when assessing, deciding and putting a request for execution of this entity.

27. A method in a cluster arrangement comprising at least two application nodes (AK, AK′

) and an inspection node (KK), which are connected to a network (NW1), comprising the steps of;

receiving at the inspection node (KK) from the application nodes (AK, AK′

) a list (T) with all the entities (I1, I2, I3) executed on the nodes and with the data required for executing the entities;

checking with the inspection node (KK) the application nodes (AK) for failure at regular intervals of time; and

if an application node (AK, AK′

) has failed, then compiling with the inspection node (KK) a list (TI3) with entities executed on the failed application node and with the data required for execution and forwards it to application nodes (AK, AK′

) connected to the network (NW1) with a request for execution.

28. The method as claimed in claim 27,

wherein

if an application node has failed then the inspection node (KK) forwards the list (TI3) compiled by the application node together with a request for execution to the at least one further application node (AK).

29. The method as claimed in claim 27,

wherein

an application node compiles a list (T) for execution of an entity and sends a request with this list (T) to the at least one further application node.

30. The method as claimed in claim 27,

wherein

if an application node has failed then the inspection node (KK) ascertains a further application node using the list (T) compiled by the application nodes and using settable parameters and transmits a request for execution of the entities executed on the failed application node to said further application node.

31. The method as claimed in claim 27,

wherein

the inspection node (KK) ascertains the application nodes which are to be checked by evaluating the lists (T) received from the application agents.

32. The method as claimed in claim 27,

wherein

the method is performed by an inspection agent (KA) executed on the inspection node and by an application agent (AA) executed on the application nodes.

33. The method as claimed in claim 27,

wherein

the at least one application node (AK) and the inspection node (KK) are allocated an identifier (IP2);

and

the inspection agent (KA) at the inspection node (KK1) receives a testament from the application agent (AA) at the at least one application node (AK) with the same identifier (IP2).

34. The method as claimed in claim 27,

wherein

an executed entity (I1, I2, L1, L2, L3) is allocated a priority (PS1, PS2);

and

if an application node (AK) has failed then the inspection agent (KA) evaluates the priority (PS1, PS2) associated with the entity executed on the failed application node (AK) and sends a request for execution on the basis of the priority (PS1, PS2).