METHOD AND MECHANISM FOR EFFICIENT RE-DISTRIBUTION OF IN-MEMORY COLUMNAR UNITS IN A CLUSTERED RDBMS ON TOPOLOGY CHANGE

• US 20200133950A1
• Filed: 12/23/2019
• Published: 04/30/2020
• Est. Priority Date: 01/27/2016
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: MUKHERJEE, NILOY, KULKARNI, KARTIK, LAHIRI, TIRTHANKAR, MARWAH, VINEET, LOAIZA, JUAN
• Granted Patent: US 11,586,641 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/2477 Temporal data queries
• CPC Class Codes: H04L 41/0668 by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 43/0817 by checking functioning

Abstract

Techniques are described herein for executing queries on distinct portions of a database object that has been separate into chunks and distributed across the volatile memories of a plurality of nodes in a clustered database system. The techniques involve redistributing the in-memory database object portions on changes to the clustered database system. Each node may maintain a mapping indicating which nodes in the clustered database system store which chunks, and timestamps indicating when each mapping entry was created or updated. A query coordinator may use the timestamps to select a database server instance with local in memory access to data required by a portion of a query to process that portion of the query.

Claims

1. A method comprising:

for each map-storing node of a plurality of map-storing nodes in a database system, maintaining a respective mapping that indicates;

a plurality of chunks of data;

for each chunk of the plurality of chunks of data;

one or more host nodes assigned to the chunk, and

for each host node of the one or more host nodes assigned to the chunk, a registration timestamp corresponding to a time when the map-storing node established that the host node was a host of the chunk;

in response to a re-distribution event, determining a first set of one or more chunks, of the plurality of chunks, affected by the re-distribution event;

for each chunk of the first set of chunks;

determining at least one additional host node for the chunk;

each map-storing node, of the plurality of map-storing nodes, updating the respective mapping to indicate the at least one additional host node and an additional registration timestamp for the chunk for the at least one additional host node;

wherein the method is performed by one or more computing devices.

2. The method of claim 1 further comprising detecting when a node, that hosts one or more chunks, has been removed, wherein the re-distribution event is in response to detecting removal of the node.

3. The method of claim 1 further comprising detecting when a node has been added, wherein the re-distribution event is in response to detecting addition of the node.

4. The method of claim 1 further comprising:

for each map-storing node of the plurality of map-storing nodes, storing a global completion timestamp that indicates when all host nodes completed loading one or more assigned chunks, of the plurality of chunks, into their respective volatile memories;

in response to the redistribution event, for each map-storing node of the plurality of map-storing nodes, storing an updated global completion timestamp that indicates when the at least one additional host node completed loading one or more assigned chunks, of the first set of chunks, into their respective volatile memories.

5. The method of claim 1 further comprising:

for each map-storing node of the plurality of map-storing nodes, storing, for each chunk of the plurality of chunks, a respective global completion timestamp that indicates when all host nodes assigned to said each chunk completed loading said each chunk into their respective volatile memories;

in response to the redistribution event, for each map-storing node of the plurality of map-storing nodes, storing, for each chunk of the first set of chunks, an updated global completion timestamp that indicates when one or more additional host nodes assigned to said each chunk completed loading said each chunk into their respective volatile memories.

6. The method of claim 1 further comprising:

determining a second set of one or more chunks, of the plurality of chunks, that are not affected by the re-distribution event;

wherein the second set of chunks does not include chunks of the first set of chunks;

not determining an additional host node for any chunks of the second set of chunks.

7. The method of claim 4 further comprising:

receiving, at a particular map-storing node, a query requiring access to a particular chunk of the first set of chunks;

in response to receiving the query;

selecting a particular host node, of one or more host nodes assigned to the particular chunk, based at least in part on the respective mapping maintained at the particular map-storing node; and

causing the particular host node to perform work, for the query, that involves accessing data items in the particular chunk.

8. The method of claim 7, wherein the one or more host nodes assigned to the particular chunk include a first host node and a second host node, wherein the re-distribution event is in response to detecting removal of the first host node, and wherein the step of selecting the particular host node involves selecting the second host node over the first host node.

9. The method of claim 7, wherein the one or more host nodes assigned to the particular chunk include a first host node where a registration time of the first host node is greater than the global completion time and a second host node where a registration time of the second host is less than the global completion time, and wherein the step of selecting a particular host node involves selecting the second host node over the first host node.

10. The method of claim 7, wherein the one or more host nodes assigned to the particular chunk include a first host node where a startup time of the first host node is greater than a registration time corresponding to the first host node and a second host node where a startup time of the second host is less than a registration time corresponding to the second host node, and wherein the step of selecting the particular host node involves selecting the second host node over the first host node.

11. The method of claim 7, wherein the one or more host nodes assigned to the particular chunk include a first host node where a startup time of the first host node is greater than a registration time corresponding to the first host node and a second host node where a startup time of the second host is lesser than a registration time corresponding to the second host, and wherein the step of selecting a particular host node involves selecting the second host node over the first host node.

12. One or more non-transitory computer-readable media storing instructions, which when executed by one or more hardware processors, cause:

for each map-storing node of a plurality of map-storing nodes in a database system, maintaining a respective mapping that indicates;

a plurality of chunks of data;

for each chunk of the plurality of chunks of data;

one or more host nodes assigned to the chunk, and

for each host node of the one or more host nodes assigned to the chunk, a registration timestamp corresponding to a time when the map-storing node established that the host node was a host of the chunk;

in response to a re-distribution event, determining a first set of one or more chunks of the plurality of chunks affected by the re-distribution event;

for each chunk of the first set of chunks;

determining at least one additional host node for the chunk;

each map-storing node, of the plurality of map-storing nodes, updating the respective mapping to indicate the at least one additional host node and an additional registration timestamp for the chunk for the at least one additional host node.

13. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause detecting when a node, that hosts one or more chunks, has been removed, wherein the re-distribution event is in response to detecting removal of the node.

14. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause detecting when a node, that hosts one or more chunks, has been removed, wherein the re-distribution event is in response to detecting addition of the node.

15. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause:

for each map-storing node of the plurality of map-storing nodes, storing a global completion timestamp that indicates when all host nodes completed loading one or more assigned chunks, of the plurality of chunks, into their respective volatile memories;

in response to the redistribution event, for each map-storing node of the plurality of map-storing nodes, storing an updated global completion timestamp that indicates when the at least one additional host node completed loading one or more assigned chunks, of the first set of chunks, into their respective volatile memories.

16. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause:

for each map-storing node of the plurality of map-storing nodes, storing, for each chunk of the plurality of chunks, a respective global completion timestamp that indicates when all host nodes assigned to said each chunk completed loading said each chunk into their respective volatile memories;

in response to the redistribution event, for each map-storing node of the plurality of map-storing nodes, storing, for each chunk of the first set of chunks, an updated global completion timestamp that indicates when one or more additional host nodes assigned to said each chunk completed loading said each chunk into their respective volatile memories.

17. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause:

determining a second set of one or more chunks, of the plurality of chunks, that are not affected by the re-distribution event;

wherein the second set of chunks does not include chunks of the first set of chunks;

not determining an additional host node for any chunks of the second set of chunks.

18. The one or more non-transitory computer-readable media of claim 12, the instructions further including instructions which, when executed by the one or more hardware processors, cause:

receiving, at a particular map-storing node, a query requiring access to a particular chunk of the first set of chunks;

in response to receiving the query;

selecting a particular host node, of one or more host nodes assigned to the particular chunk, based at least in part on the respective mapping maintained at the particular map-storing node; and

causing the particular host node to perform work, for the query, that involves accessing data items in the particular chunk.

19. The one or more non-transitory computer-readable media of claim 18, wherein the one or more host nodes assigned to the particular chunk include a first host node and a second host node, wherein the re-distribution event is in response to detecting removal of the first host node, and wherein the step of selecting the particular host node involves selecting the second host node over the first host node.

20. The one or more non-transitory computer-readable media of claim 18, wherein the one or more host nodes assigned to the particular chunk include a first host node where a registration time of the first host node is greater than the global completion time and a second host node where a registration time of the second host is less than the global completion time, and wherein the step of selecting a particular host node involves selecting the second host node over the first host node.

21. The one or more non-transitory computer-readable media of claim 18, wherein the one or more host nodes assigned to the particular chunk include a first host node where a startup time of the first host node is greater than a registration time corresponding to the first host node and a second host node where a startup time of the second host is less than a registration time corresponding to the second host node, and wherein the step of selecting the particular host node involves selecting the second host node over the first host node.

22. The one or more non-transitory computer-readable media of claim 18, wherein the one or more host nodes assigned to the particular chunk include a first host node where a startup time of the first host node is greater than a registration time corresponding to the first host node and a second host node where a startup time of the second host is lesser than a registration time corresponding to the second host, and wherein the step of selecting a particular host node involves selecting the second host node over the first host node.