Assembly and method for balancing processors in a partitioned server

• US 7,451,183 B2
• Filed: 03/21/2003
• Issued: 11/11/2008
• Est. Priority Date: 03/21/2003
• Status: Expired due to Fees

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Romero, Francisco J., Crawford, Isom
• Primary Examiner(s): Patel; Haresh N
• Application Number:
  US10/394,608
• Publication Number:
  US 20040199632A1
• Time in Patent Office: 2,062 Days
• Field of Search: 709212-226, 709200-204, 709206-209, 709245-250, 714 10- 12, 714/1, 714/13, 714/3, 714/7, 714/8, 719310-318, 710/104, 710/302, 712/13, 712/20, 712/23, 712/203, 712/227, 712/11, 712/15, 711/153, 718100-105
• US Class Current: 709/206
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system

Abstract

A system and method for automatically allocating computing resources in a partitioned server. The method includes determining that a partition of the partitioned server requires activation of a reserve processor, determining that another partition of the partitioned server has an active processor that may be deactivated, activating the reserve processor, and deactivating the active processor. An article of manufacture including a machine-readable medium having stored thereon instructions for automatically allocating computing resources in a partitioned server.

Claims

1. A method for automatically allocating computing resources in a partitioned server, the method comprising:

identifying in a partitioned server a first partition having at least one reserve processor available for activation to meat service level objectives associated with applications running on the first partition;

identifying in the partitioned server a second partition having at least one active processor;

wherein each of the first partition and the second partition is a hardware-based partition;

activating the at least one reserve processor in the first partition so that the first partition has an activated processor and the second partition does not have the activated processor; and

deactivating the at least one active processor in the second partition so that the at least one active processor becomes a second reserved processor in the second partition and the second partition has a deactivated processor and the first partition does not have the deactivated processor, and so that a total number of reserved processors in the partitioned server remains constant after activating the at least one reserved processor;

wherein a number of reserve processors activated in the partitioned server is equal to a number of active processors deactivated in the partitioned server;

wherein the number of reserve processors is at a same amount, before activating the at least one reserve processor and after activating the at least one reserve processor; and

wherein each reserve processor is available for operation as an active processor in the partitioned server.

2. The method of claim 1 additionally comprising collecting active processor requests in the first partition.

3. The method of claim 2 additionally comprising adjusting the active processor requests in accordance with a number of reserve processors in the first partition.

4. The method of claim 1 wherein a number of reserve processor activated in the first partition is equal to a number of active processors deactivated in the second partition.

5. The method of claim 1, wherein the activated processor does not cross a boundary of the first partition and the deactivated processor does not cross a boundary of the second partition.

6. A method for automatically allocating computing resources in a partitioned server, the method comprising:

collecting from a partitioned server an active-processor resource request at a first priority level;

determining that all active processor resources for all partitions in the partitioned server have already been allocated at the first priority level;

collecting from each partition in the partitioned server active-processor resource requests at a second priority level which is lower than the first priority level;

identifying in the partitioned server a first partition having at least one active processor which is not allocated at the second priority level;

identifying in the partitioned server a second partition having at least one reserve processor and which is requesting at least one active processor at the second priority level;

wherein each of the first partition and the second partition is a hardware-based partition;

activating the at least one reserve processor in the second partition so that the second partition has an activated processor and the first partition does not have the activated processor;

wherein the number of reserve processors is at a same amount, before activating the at least one reserve processor and after activating the at least one reserve processor; and

deactivating the at least one active processor in the first partition so that the first partition has a deactivated processor and the second partition does not have the deactivated processor;

wherein a number of reserve processors activated in the partitioned server is equal to a number of active processors deactivated in the partitioned server; and

wherein each reserve processor is available for operation as an active processor in the partitioned server.

7. The method of claim 6 additionally comprising adjusting the active-processor resource requests at the second priority level to take into account whether there are reserve processors available in the second partition that are capable of satisfying the active-processor resource requests at the second priority level.

8. The method of claim 6, wherein the activated processor does not cross a boundary of the second partition and the deactivated processor does not cross a boundary of the first partition.

9. A system for automatically allocating computing resources in a partitioned server comprising:

a workload manager; and

a partitioned server coupled to the workload manager and including a plurality of partitions having active processors and a number of reserve processors;

wherein the plurality of partitions comprises a first partition and a second partition and wherein each of the first partition and the second partition is a hardware-based partition;

wherein the workload manager activates a reserve processor in the first partition so that the first partition has an activated processor and the second partition does not have the activated processor; and

wherein the workload manager deactivates an active processor in the second partition so that the active processor becomes a second reserved processor in the second partition and the second partition has a deactivated processor and the first partition does not have the deactivated processor and so that a total number of reserved processors in the partitioned server remains constant after activating the reserved processor;

wherein a number of reserve processors activated in the partitioned server is equal to a number of active processors deactivated in the partitioned server;

wherein the number of reserve processors is at a same amount, before activating the reserve processor and after activating the reserve processor; and

wherein each reserve processor is available for operation as an active processor in the partitioned server.

10. The system of claim 9 wherein a number of reserve processors remains the same regardless of a number of active processors deactivated in the second partition and a number of reserve processors activated in the first partition.

11. The system of claim 9, wherein the activated processor does not cross a boundary of the first partition and the deactivated processor does not cross a boundary of the second partition.

12. A system for automatically allocating computing resources in a partitioned server comprising:

a partitioned server including a plurality of partitions including a first partition and second partition, wherein each of the first partition and the second partition is a hardware-based partition; and

means, coupled to the partitioned server, for determining that the first partition of the partitioned server requires the activation of at least one reserve processor in the first partition and for determining that a the second partition of the partitioned server has at least one active processor to be deactivated in the second partition;

wherein the determining means activates the at least one reserve processor in the first partition so that the first partition has an activated processor and the second partition does not have the activated processor; and

wherein the determining means deactivates the at least one active processor in the second partition so that the second partition has a deactivated processor and the first partition does not have the deactivated processor one;

wherein a number of reserve processors activated in the partitioned server is equal to a number of active processors deactivated in the partitioned server;

wherein the number of reserve processors is at a same amount, before activating the at least one reserve processor and after activating the at least one reserve processor; and

wherein each reserve processor is available for operation as an active processor in the partitioned server.

13. The system of claim 12, wherein the activated processor does not cross a boundary of the first partition and the deactivated processor does not cross a boundary of the second partition.