Re: Journalled File System (long)

[Arun Sharma <adsharma@xxxxxxxx>]

BGanesh <BGa@xxxxxxxxxxxxx> writes:

Ganesh,

Thanks for the nice technical post.

> 
> Traditionally, UNIX file systems such as the Berkeley fast file system
> have to write everything through the disk. When you're updating,
> creating your file, they have to write all the way through the disk and
> wait for that write to complete. This is a synchronous write.
> 

If a filesystem does write through, it'd be unusable. Berkeley FFS
does synchronous writes of only metadata (directory entries, file
distribution etc). This is only the default behaviour. You can make it
async through a mount option.

Also, there is some work layered on top of FFS called
"softupdates". It tries to reorder changes to metadata and thus
optimize for performance.

There was some work by Rosenblum et al, on LFS (log structure
filesystem) which does logging for data also.

Again, thanks for a good post.

	-Arun

- --------------------------------------------------------------------
For more information on Linux in India visit http://www.linux-india.org/

------------------------------