Ext2 and Ext3 file systems

Ext2: is a standard file system in GNU/Linux systems. It features excellent file access performance and advantages for small and medium-sized files. This is mainly due to the excellent design of its cluster cache layer. The single file size and the file system's own capacity limit are related to the file system's own cluster size. In the common x86 computer system, the maximum cluster size is 4KB, the single file size is limited to 2048GB, and the file system capacity is limited to 16384GB. . However, since the current single partition of core 2.4 can only use a maximum of 2048GB, the actual file system capacity can be used up to 2048GB.

Ext3: As its name implies, it is the next generation of ext2, which is the ability to add logging under the current ext2 format. It is still a long way from the practical stage, and maybe it will be on the core of the next version. Ext3 is a journaled file system. The advantage of the journaled file system is that since the file system has a cache layer to participate in the operation, if the file system is not used, the file system must be removed to write the data of the cache layer back to the disk. Therefore, whenever the system is to be shut down, all of its file systems must be removed before it can be shut down.

If the file system is shut down (such as a power outage) before it is removed, the file system data will be inconsistent after the next reboot. Therefore, the file system must be reworked, which will be inconsistent. The wrong place to fix. However, this reorganization is quite time consuming, especially for file systems with large capacity, and it is not 100% guaranteed that all data will not be lost. This can cause problems on large servers.

In order to overcome this problem, the industry has developed for a long time and completed the so-called "Journal File System" & rsquo;. The biggest feature of this type of file system is that it will completely record the entire disk write action on an area of ​​the disk so that it can be traced back when needed. Because the data write action contains many details, such as changing the file header data, searching the disk writeable space, writing data sections one by one, etc., if each detail is half-cut, it will cause the file. The system is inconsistent and therefore needs to be reorganized. However, in the log file system, since each detail is recorded in detail, when interrupted in a certain process, the system can directly review and reconstruct the interrupted portion based on these records without having to take the time to check. The other part, so the reorganization work is quite fast, and it takes almost no time.

In addition, there is a swap file system specifically for swap partitions in Linux. Linux uses the entire partition as the swap space, instead of using swap files like Windows. Generally, the swap partition of this SWAP format is twice the main memory.