MongoDB 2.0 Journaling

Wow haha, journaling is not really a new function, always in the previous version has appeared, but the official saying is still in the test, can not guarantee stability, but in 2.0 finally be correct, although said the server is not so easy to collapse, but better remedy is always exciting ah, hey.

Simply put, all MONGO data operations will be placed in memory first, and then one by one execution, once the machine is hung, those still in memory too late to execute the data will be lost, and journaling recorded every operation into the memory of the action, For example, the process of inserting a piece of data is: an insert operation into the memory-> in memory waiting for the completion of the-> in the database to actually complete, the journaling will be "an insert operation into memory" when it recorded, journaling like a real log, It produces a file that does exist, so once sent a request to the server, it will immediately be recorded in this file, even when the machine, the memory of the things are lost, this document is still clearly remembered, MONGO again according to this log to restore on the line.

Note that when the journaling feature is turned on, the operation is scheduled to be committed by the journaling to the database, and the length of the timer can be defined by the user, ranging from 2 to 300 milliseconds, and if not set, the default is 100 milliseconds.

There is no Linux on hand now, back to have a complete record of it: P

Original link: http://www.cnblogs.com/sasafly/archive/2011/10/07/2200639.html