Abstract: A method for replaying a log on a data node, a data node, and a database system are described. The method is used to improve a log replay speed, and includes: the data node obtaining a plurality of logs, wherein the plurality of logs includes at least one transaction commit log and at least one page operation log; and a first thread of a plurality of threads replaying the at least one page operation log, and a same second thread replaying all page operation logs including an operation on a same page. As such, log replay progress does not need to be synchronized between threads. A process in which the first thread replays the transaction commit log is independent of a process in which at least one second thread replays the at least one page operation log.