Database replication

• US 8,924,347 B1
• Filed: 10/05/2012
• Issued: 12/30/2014
• Est. Priority Date: 10/05/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Kasera, Vishal, Leon, Jean-Michel, Ashcraft, Kenneth M, Agarwal, Amit
• Primary Examiner(s): LE, DEBBIE M
• Application Number:
  US13/646,019
• Time in Patent Office: 816 Days
• US Class Current: 707/615
• CPC Class Codes: G06F 12/0806 Multiuser, multiprocessor or multiprocessing cache systems
• CPC Class Codes: G06F 16/215 Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: H04L 49/9005 using dynamic buffer space allocation
• CPC Class Codes: H04L 67/01 Protocols
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]

Abstract

A replication server receives a first write request from a database server for writing first data to persistent data storage. The replication server determines that the first write request is designated as asynchronous. In response to determining that the first write request is asynchronous, the first data is saved to local cache at the replication server and an acknowledgement is sent to the database server. The local cache stores second data to be written to persistent data storage when the first data is saved to the local cache. The first data and the second data are written from the local cache to the persistent data storage in a batch operation.

Claims

1. A method comprising:

receiving, at a database server from a client application, a first client write request for writing first data to persistent data storage and an indication that the first write request is designated as asynchronous;

receiving, at a replication server, a first write request from the database server for writing first data to persistent data storage;

determining, by the replication server, that the first write request is designated as asynchronous;

in response to determining that the first write request is asynchronous, saving the first data to local cache at the replication server and sending an acknowledgement to the database server, wherein the local cache stores second data to be written to persistent data storage when the first data is saved to the local cache; and

writing the first data and the second data from the local cache to the persistent data storage in a batch operation;

receiving, at the database server from the client application, a second client write request for writing second data to persistent data storage and an indication that the second write request is designated as synchronous;

receiving, at the replication server, a second write request from the database server for writing third data to persistent data storage;

determining, by the replication server, that the second write request is designated as synchronous; and

in response to determining that the second write request is synchronous, writing the third data to persistent data storage and sending an acknowledgement to the database server after writing the third data to persistent storage.

2. The method of claim 1, wherein the write request includes information on a lease on the persistent data storage by the database server for servicing the first write request, the method further comprising:

determining, by the replication server, that the information on the lease matches lease information locally stored at the replication server; and

saving the first data to the local cache comprises saving the first data to the local cache in response to determining that the information on the lease matches lease information locally stored at the replication server.

3. The method of claim 2, wherein writing the first data and the second data from the local cache to the persistent data storage in a batch operation comprises renewing the lease on the persistent data storage for the database server.

4. The method of claim 1, wherein determining that the first write request is designated as asynchronous comprises determining that the first write request is asynchronous based on information received with the first write request.

5. The method of claim 4, wherein the information is provided by one of the database server or a client application that originated the first write request.

6. The method of claim 1, wherein writing the first data and the second data from the local cache to the persistent data storage in a batch operation comprises:

determining whether a threshold is reached for writing the locally cached data to the persistent data storage; and

based on determining that the threshold is reached, writing the first data and the second data from the local cache to the persistent data storage in the batch operation.

7. The method of claim 6, wherein the threshold is based on at least one of an amount of data in the local cache or a time period that is user configurable.

8. The method of claim 1, wherein a replication server is configured to perform read and write operations directly on the persistent data storage.

9. The method of claim 1, wherein writing the first data and the second data to the persistent data storage in a batch operation comprises:

determining that sequence numbers associated with the first data and the second data are contiguous; and

writing the first data and the second data to the persistent data storage in the batch operation in response to determining that the sequence numbers are contiguous.

10. The method of claim 1, wherein writing the first data and the second data from the local cache to the persistent data storage in a batch operation comprises:

determining that the first data and the second data have not been written to persistent data storage by another replication server; and

writing the first data and the second data from the local cache to the persistent data storage in the batch operation in response to determining that the first data and the second data have not been written to persistent data storage by another replication server.

11. The method of claim 1, wherein writing the first data and the second data from the local cache to the persistent data storage in the batch operation is performed independently of a state of operation of the database server.

12. The method of claim 1, wherein the replication server is included in a group of replication servers that are geographically distributed.

13. A computer program product, suitably embodied in a non-transitory machine-readable medium and including instructions executable by a processor, the instructions configured to cause the processor to perform operations comprising:

receiving, at a database server from a client application, a first client write request for writing first data to persistent data storage and an indication that the first write request is designated as asynchronous;

receiving, at a replication server, a first write request from the database server for writing first data to persistent data storage;

determining, by the replication server, that the first write request is designated as asynchronous;

in response to determining that the first write request is asynchronous, saving the first data to local cache at the replication server and sending an acknowledgement to the database server, wherein the local cache stores second data to be written to persistent data storage when the first data is saved to the local cache; and

writing the first data and the second data from the local cache to the persistent data storage in a batch operation;

receiving, at the database server from the client application, a second client write request for writing second data to persistent data storage and an indication that the second write request is designated as synchronous;

receiving, at the replication server, a second write request from the database server for writing third data to persistent data storage;

determining, by the replication server, that the second write request is designated as synchronous; and

in response to determining that the second write request is synchronous, writing the third data to persistent data storage and sending an acknowledgement to the database server after writing the third data to persistent storage.

14. The computer program product of claim 13, wherein the write request includes information on a lease on the persistent data storage by the database server for servicing the first write request, the computer program product further including instructions that are configured to cause the processor to perform operations comprising:

determining, by the replication server, that the information on the lease matches lease information locally stored at the replication server; and

saving the first data to the local cache comprises saving the first data to the local cache in response to determining that the information on the lease matches lease information locally stored at the replication server.

15. The computer program product of claim 14, wherein the instructions that are configured to cause the processor to perform operations comprising writing the first data and the second data from the local cache to the persistent data storage in a batch operation includes instructions that are configured to cause the processor to perform operations comprising renewing the lease on the persistent data storage for the database server.

16. The computer program product of claim 13, wherein the instructions that are configured to cause the processor to perform operations comprising determining that the first write request is designated as asynchronous includes instructions that are configured to cause the processor to perform operations comprising determining that the first write request is asynchronous based on information received with the first write request.

17. The computer program product of claim 16, wherein the information is provided by one of the database server or a client application that originated the first write request.

18. The computer program product of claim 13, wherein the instructions that are configured to cause the processor to perform operations comprising writing the first data and the second data from the local cache to the persistent data storage in a batch operation includes instructions that are configured to cause the processor to perform operations comprising:

determining whether a threshold is reached for writing the locally cached data to the persistent data storage; and

based on determining that the threshold is reached, writing the first data and the second data from the local cache to the persistent data storage in the batch operation.

19. The computer program product of claim 18, wherein the threshold is based on at least one of an amount of data in the local cache or a time period that is user configurable.

20. The computer program product of claim 13, wherein a replication server is configured to perform read and write operations directly on the persistent data storage.

21. The computer program product of claim 13, wherein the instructions that are configured to cause the processor to perform operations comprising writing the first data and the second data to the persistent data storage in a batch operation includes instructions that are configured to cause the processor to perform operations comprising:

determining that sequence numbers associated with the first data and the second data are contiguous; and

writing the first data and the second data to the persistent data storage in the batch operation in response to determining that the sequence numbers are contiguous.

22. The computer program product of claim 13, wherein the instructions that are configured to cause the processor to perform operations comprising writing the first data and the second data to the persistent data storage in a batch operation includes instructions that are configured to cause the processor to perform operations comprising:

determining that the first data and the second data have not been written to persistent data storage by another replication server; and

writing the first data and the second data from the local cache to the persistent data storage in the batch operation in response to determining that the first data and the second data have not been written to persistent data storage by another replication server.

23. The computer program product of claim 13, wherein writing the first data and the second data from the local cache to the persistent data storage in the batch operation is performed independently of a state of operation of the database server.

24. The computer program product of claim 13, wherein the replication server is included in a group of replication servers that are geographically distributed.

25. A method comprising:

receiving, at a database server from a client application, a first write request for writing first data to persistent data storage and an indication that the first write request is designated as asynchronous;

in response to receiving the first write request, selecting, by the database server, a first set of replication servers;

determining, by the database server, that the first write request is designated as asynchronous;

sending the first data from the database server to the first set of replication servers for writing to the persistent data storage, along with an indication that the first write request is asynchronous, wherein replication servers included in the first set are configured to save the data to respective local caches in response to receiving the indication that the first write request is asynchronous, send a confirmation to the database server when the data is saved to the respective local caches, and write the first data to persistent storage in a batch operation with other data stored in the respective local caches;

receiving, at the database server, confirmation from the replication servers in the first set;

in response to receiving confirmation from the replication servers in the first set, sending information indicating success of the first write request from the database server to the client application;

receiving, at the database server from the client application, a second write request for writing second data to persistent data storage and an indication that the second write request is designated as synchronous;

in response to receiving the second write request, selecting, by the database server, a second set of replication servers;

determining, by the database server, that the second write request is designated as synchronous;

sending the second data from the database server to the second set of replication servers for writing to the persistent data storage, along with an indication that the second write request is synchronous, wherein replication servers included in the second set are configured to write the second data to persistent data storage in response to receiving the indication that the second write request is synchronous and send confirmation to the database server after writing the second data to persistent storage;

receiving, at the database server, confirmation from the replication servers in the second set; and

in response to receiving confirmation from the replication servers in the second set, sending information indicating success of the second write request from the database server to the client application.

26. The method of claim 25, wherein the replication servers included in at least one of the first set or the second set are geographically distributed.

27. The method of claim 25, wherein a replication server is configured for performing read and write operations directly on the persistent data storage.

28. The method of claim 25, wherein receiving the first write request from the client application for writing the first data to persistent data storage comprises:

determining, by the database server, whether a mastership is acquired by the database server for servicing the first write request; and

selecting, by the database server, the first set of replication servers in response determining that mastership is acquired for servicing the first write request.

29. The method of claim 28, wherein the mastership indicates a lease on persistent data storage by an instance of the database server that is configured for servicing the first write request from the client application, and wherein determining whether a mastership is acquired for servicing the first write request comprises:

based on determining that mastership is not acquired for servicing the first write request, generating, by the database server, a mastership identifier associated with servicing the first write request; and

sending an instruction from the database server to a replication server for writing a lease to an underlying replication mechanism that is configured for writing data to the persistent data storage, the instruction including the mastership identifier.

30. The method of claim 25, wherein sending the first data from the database server to the first set of replication servers for writing to the persistent data storage comprises:

sending information on a lease on the persistent data storage by the database server for servicing the first write request;

determining, by a replication server included in the first set, whether the information on the lease matches lease information locally stored at the replication server; and

saving the first data to the local cache at the replication server based on determining that the information on the lease matches lease information locally stored at the replication server.

31. The method of claim 30, wherein the replication server is configured for renewing the lease on the persistent data storage for the database server upon writing the locally cached data to the persistent data storage.

32. A computer program product, suitably embodied in a non-transitory machine-readable medium and including instructions executable by a processor, the instructions configured to cause the processor to perform operations comprising:

receiving, at a database server from a client application, a first write request for writing first data to persistent data storage and an indication that the first write request is designated as asynchronous;

in response to receiving the first write request, selecting, by the database server, a first set of replication servers;

determining, by the database server, that the first write request is designated as asynchronous;

sending the first data from the database server to the first set of replication servers for writing to the persistent data storage, along with an indication that the first write request is asynchronous, wherein replication servers included in the first set are configured to save the data to respective local caches in response to receiving the indication that the first write request is asynchronous, send a confirmation to the database server when the data is saved to the respective local caches, and write the first data to persistent storage in a batch operation with other data stored in the respective local caches;

receiving, at the database server, confirmation from the replication servers in the first set; and

in response to receiving confirmation from the replication servers in the first set, sending information indicating success of the first write request from the database server to the client application;

receiving, at the database server from the client application, a second write request for writing second data to persistent data storage and an indication that the second write request is designated as synchronous;

in response to receiving the second write request, selecting, by the database server, a second set of replication servers;

determining, by the database server, that the second write request is designated as synchronous;

sending the second data from the database server to the second set of replication servers for writing to the persistent data storage, along with an indication that the second write request is synchronous, wherein replication servers included in the second set are configured to write the second data to persistent data storage in response to receiving the indication that the second write request is synchronous and send confirmation to the database server after writing the second data to persistent storage;

receiving, at the database server, confirmation from the replication servers in the second set; and

in response to receiving confirmation from the replication servers in the second set, sending information indicating success of the second write request from the database server to the client application.

33. The computer program product of claim 32, wherein the replication servers included in at least one of the first set or the second set are geographically distributed.

34. The computer program product of claim 32, wherein a replication server is configured for performing read and write operations directly on the persistent data storage.

35. The computer program product of claim 32, wherein the instructions that are configured to cause the processor to perform operations comprising receiving the first write request from the client application for writing the first data to persistent data storage includes instructions that are configured to cause the processor to perform operations comprising:

determining, by the database server, whether a mastership is acquired by the database server for servicing the first write request; and

selecting, by the database server, the first set of replication servers in response determining that mastership is acquired for servicing the first write request.

36. The computer program product of claim 35, wherein the mastership indicates a lease on persistent data storage by an instance of the database server that is configured for servicing the first write request from the client application, and wherein the instructions that are configured to cause the processor to perform operations comprising determining whether a mastership is acquired for servicing the first write request includes instructions that are configured to cause the processor to perform operations comprising:

based on determining that mastership is not acquired for servicing the first write request, generating, by the database server, a mastership identifier associated with servicing the first write request; and

sending an instruction from the database server to a replication server for writing a lease to an underlying replication mechanism that is configured for writing data to the persistent data storage, the instruction including the mastership identifier.

37. The computer program product of claim 32, wherein the instructions that are configured to cause the processor to perform operations comprising sending the first data from the database server to the first set of replication servers for writing to the persistent data storage includes instructions that are configured to cause the processor to perform operations comprising:

sending information on a lease on the persistent data storage by the database server for servicing the first write request;

determining, by a replication server included in the first set, whether the information on the lease matches lease information locally stored at the replication server; and

saving the first data to the local cache at the replication server based on determining that the information on the lease matches lease information locally stored at the replication server.

38. The computer program product of claim 37, wherein the replication server is configured for renewing the lease on the persistent data storage for the database server upon writing the locally cached data to the persistent data storage.