Introduction to Linux File system logs

We refer to inode blocks and block blocks as data storage areas or blocks of data, such as superblock, block blocks, inode bitmaps, etc., which are called intermediate or meta-data, because metadata changes frequently and there is a case of inconsistency with the actual data. So there is the file system log function.

Ext3 File System log: First put a copy of the block to be written into the log, when the IO data sent to the log to the log (the data submitted to the log), the block will be written to the file system, when sent to the file system IO data is completed, (the data submitted to the file system), the log block copy is discarded.

In fact, log records of metadata can restore the consistency of the file system structure. Can not log the actual data of the file, if not recorded can not recover data, ext3 let the system administrator decide what to log in, there are 3 modes:

1. Log (Journal)

   All data and metadata for the file system are recorded in the log. The safest and slowest Ext3 file system.

2. Reservation (Ordered)

   Only changes to the file system metadata are recorded in the log. The default ext3 log mode.

3. Write Back (Writeback)

   Only changes to the file system metadata are recorded in the log. The fastest mode.

Introduction to Linux File system logs