MANAGING DATABASE NODES

• US 20150052383A1
• Filed: 08/15/2013
• Published: 02/19/2015
• Est. Priority Date: 08/15/2013
• Status: Active Grant

• Current Assignee: Micro Focus LLC (Open Text Corporation)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Finnerty, James L., Dave, Jaimin M., Walkauskas, Stephen G.
• Granted Patent: US 10,303,567 B2
• Time in Patent Office: Days
• US Class Current: 714/4.11
• CPC Class Codes: G06F 11/1461 Backup scheduling policy
• CPC Class Codes: G06F 11/2025 using centralised failover control functionality
• CPC Class Codes: G06F 11/2035 without idle spare hardware
• CPC Class Codes: G06F 11/2048 where the redundant components share neither address space nor persistent storage

Abstract

A method for managing database nodes includes determining that a data segment is on a failed node. The data segment is referenced by an operation of a query plan. The method includes selecting a victim node based on a segmentation ring, a buddy node for the data segment, a plurality of remaining operational nodes, and a predetermined selection parameter. The method includes generating a query plan such that the victim node performs double duty for operations accessing the data segment from a buddy projection on the victim node, and operations accessing a data segment for a primary projection of the victim node.

Claims

1. A method for managing database nodes, the method comprising:

determining that a data segment is on a failed node, the data segment being referenced by an operation of a query plan;

selecting a victim node based on a segmentation ring, a buddy node for the data segment, a plurality of remaining operational nodes, and a predetermined selection parameter; and

generating a query plan such that the victim node performs double duty for operations accessing the data segment from a buddy projection on the victim node, and operations accessing a data segment for a primary projection of the victim node.

2. The method of claim 1, comprising:

determining that a segment group is on the failed node, the segment group comprising a plurality of segments;

selecting a plurality of victim nodes based on the segmentation ring, buddy nodes hosting copies of the segments belonging to buddy projections, the plurality of remaining operational nodes, and the predetermined selection parameter; and

generating a query plan such that the victim nodes perform double duty for operations accessing the segments from buddy projections on the victim nodes, and operations accessing data segments for primary projections of the victim nodes.

3. The method of claim 2, the segments comprising a segment group being hosted on a failed node, and the predetermined selection parameter comprising a workload balance for the victim nodes.

4. The method of claim 3, wherein the workload balance comprises work being performed at a time of failure for the failed node, and work being performed at the time of failure by the failed node for the segment group.

5. The method of claim 1, each of the database nodes comprising at least one segment of a projection.

6. The method of claim 1, each of the database nodes comprising at least one segment group of a projection.

7. The method of claim 1, the predetermined selection parameter being a round-robin selection.

8. The method of claim 1, the predetermined selection parameter being a random selection.

9. The method of claim 1, the predetermined selection parameter being based on resource parameters of the remaining operational nodes.

10. A system, comprising:

a plurality of database nodes; and

a database node comprising;

a processor that is adapted to execute stored instructions; and

a memory device that stores instructions, the memory device comprising;

computer-implemented instructions to determine that a plurality of segment groups are on a plurality of failed nodes, the segment groups comprising pluralities of segments;

computer-implemented instructions to select a plurality of victim nodes based on a segmentation ring, buddy nodes hosting copies of the segments belonging to buddy projections, a plurality of remaining operational nodes, and a predetermined selection parameter; and

computer-implemented instructions to generate a query plan such that the victim nodes perform double duty for operations accessing the segments from buddy projections on the victim nodes, and operations accessing data segments for primary projections of the victim nodes.

11. The system of claim 10, the predetermined selection parameter comprising a workload balance for the victim nodes.

12. The system of claim 11, wherein the workload balance comprises work being performed by the victim nodes, at a time of failure for one of the failed nodes, and work being performed, at the time of failure for the one, by the failed nodes, for the segment groups.

13. The system of claim 10, each of the database nodes comprising at least one segment of a projection.

14. The system of claim 10, each of the database nodes comprising at least one segment group of a projection.

15. The system of claim 10, the predetermined selection parameter being a selected one of a round-robin selection, and a random selection, or being based on resource parameters of the remaining operational nodes.

16. A tangible, non-transitory, machine-readable medium that stores machine-readable instructions executable by a processor to manage database nodes, the tangible, non-transitory, machine-readable medium comprising:

machine-readable instructions that, when executed by the processor, determine that a plurality of segment groups are on a plurality of failed nodes, the segment groups comprising pluralities of segments;

machine-readable instructions that, when executed by the processor, select a plurality of victim nodes based on a segmentation ring, buddy nodes hosting copies of the segments belonging to buddy projections, a plurality of remaining operational nodes, and a predetermined selection parameter; and

machine-readable instructions that, when executed by the processor, generate a query plan such that the victim nodes perform double duty for operations accessing the segments from buddy projections on the victim nodes, and operations accessing data segments for primary projections of the victim nodes.

17. The medium of claim 16, the predetermined selection parameter comprising a workload balance for the victim nodes.

18. The medium of claim 17, wherein the workload balance comprises work being performed by the victim nodes, at a time of failure for one of the failed nodes, and work being performed, at the time of failure for the one, by the failed nodes, for the segment groups.

19. The medium of claim 16, each of the database nodes comprising at least one segment of a projection.

20. The medium of claim 16, each of the database nodes comprising at least one segment group of a projection.