Asynchronous Garbage Collection in Database Redo Log Replay

Patent metadata

Field	Value
Publication number	US-2018239676-A1
Application number	US-201715436535-A
Country	US
Kind code	A1
Filing date	Feb 17, 2017
Priority date	Feb 17, 2017
Publication date	Aug 23, 2018
Grant date	—

Abstract

Recovery of a database system by taking the database system offline is initiated. Thereafter, recovery operations specified by a redo log of the database system are replayed. A cleanup log is generated that identifies cleanup operations occurring during the replay of the recovery operations for garbage collection. Concurrent with the startup of the database, garbage collection of the cleanup operations as specified in the database savepoint is initiated. In addition, concurrent with the replay of the recovery operations, garbage collection of the cleanup operations specified by the cleanup log is initiated. The database system is later brought online after all of the recovery operations are replayed.

First claim

What is claimed is:

1 . A computer-implemented method comprising:
initiating recovery of a database system by taking the database system offline; replaying recovery operations specified by a redo log of the database system; generating a cleanup log identifying cleanup operations occurring during the replay of the recovery operations for garbage collection; initiating, concurrent with the startup of the database, garbage collection of the cleanup operations as specified in the database savepoint; initiating, concurrent with the replay of the recovery operations, garbage collection of the cleanup operations specified by the cleanup log; and bringing the database system online after all of the recovery operations are replayed.

2 . The method of claim 1 , wherein the cleanup log is passed to a history manager after it is generated and the history manager subsequently initiates the garbage collection of the cleanup operations specified by the cleanup log.

3 . The method of claim 1 , wherein the garbage collection of the cleanup operations continues subsequent to the database system being brought online.

4 . The method of claim 1 , wherein the replay operations are executed by recovery threads and the cleanup operations are executed by garbage collection threads.

5 . The method of claim 4 , wherein the database comprises a plurality of programmable processor cores and the threads are striped such that each core executes a maximum of one recovery thread and one garbage collection thread.

6 . The method of claim 5 , wherein at least one core prioritizes execution of the corresponding recovery thread over the corresponding garbage collection thread.

7 . The method of claim 1 , wherein the database system comprises an in-memory database storing data in main memory.

8 . A computer-implemented method comprising:
initiating replication of a primary database system by having an associated secondary database system mirroring data stored in the primary database system and sending the most recent savepoint of the primary system to the secondary; replaying, by the secondary database system, recovery operations specified by a redo log of the primary database system and sent to the secondary system; initiating, concurrent to subsequent recovery operations, garbage collection of cleanup operations which are part of the savepoint sent from the primary; generating, by the secondary database system, a cleanup log identifying cleanup operations occurring during the replay of the recovery operations for garbage collection; initiating, by the secondary database system concurrent with the replay of the recovery operations, garbage collection of the cleanup operations; and redirecting readonly transactions for execution by the primary database system to the secondary database system which are able to block the asynchronous garbage collection in the secondary database system to get a stable view of data replicated on the secondary as required by certain transaction isolation levels.

9 . The method of claim 8 , wherein the cleanup log is passed to a history manager after it is generated and the history manager subsequently initiates the garbage collection of the cleanup operations specified by the cleanup log.

10 . The method of claim 9 further comprising:
determining whether garbage collection can be initiated based on a snapshot timestamp of an oldest reader of the primary database system relative to a commit timestamp of a transaction that generated the cleanup log.

12 . The method of claim 10 , wherein the garbage collection is initiated if the commit timestamp of the cleanup log is older than the snapshot timestamp of the oldest reader.

13 . The method of claim 10 further comprising:
delaying initiation of the garbage collection of the cleanup operations if the commit timestamp of the cleanup log is more recent than the snapshot timestamp of the oldest reader.

14 . The method of claim 8 , wherein the garbage collection of the cleanup operations continues subsequent to the primary database system being brought online.

15 . The method of claim 8 , wherein the garbage collection of the cleanup operations continues subsequent to the secondary database system executing the redirected transactions.

16 . The method of claim 8 , wherein the replay operations are executed by recovery threads and the cleanup operations are executed by garbage collection threads.

17 . The method of claim 16 , wherein the database comprises a plurality of programmable processor cores and the threads are striped such that each core executes a maximum of one recovery thread and one garbage collection thread.

18 . The method of claim 17 , wherein at least one core prioritizes execution of the corresponding recovery thread over the corresponding garbage collection thread.

19 . The method of claim 8 , wherein the database system comprises an in-memory database storing data in main memory.

20 . A database system comprising:
at least one programmable data processor; and memory storing instructions which, when executed by the at least one programmable data processor, result in operations comprising:
initiating recovery of the database system by taking the database system offline;
replaying recovery operations specified by a redo log of the database system;
generating a cleanup log identifying cleanup operations occurring during the replay of the recovery operations for garbage collection;
initiating, concurrent with the replay of the recovery operations, garbage collection of the cleanup operations specified by the cleanup log or specified in the savepoint; and
bringing the database system online after all of the recovery operations are replayed.

Assignees

• Sap Se

Inventors

• Glebe Thorsten
• Thesing Werner
• Roterring Christoph

Classifications

• G06F11/1469 Primary

  Backup restoration techniques · CPC title

• G06F12/0253

  Garbage collection, i.e. reclamation of unreferenced memory · CPC title

• G06F2212/1041

  Resource optimization · CPC title

• G06F2201/84

  Using snapshots, i.e. a logical point-in-time copy of the data · CPC title

• G06F2201/80

  Database-specific techniques · CPC title