Re: Bug#605009: serious performance regression with ext4

[Bastian Blank <waldi@debian.org>]

On Mon, Nov 29, 2010 at 01:22:25PM +0000, Ian Jackson wrote:
> Olaf van der Spek writes ("Re: Bug#605009: serious performance regression with ext4"):
> > Are there any plans to provide an API for atomic (non-durable) file
> > updates, not involving fsync?
> Yes.  Such an API has already been defined by POSIX, SuSv3, et al.
> It's called "rename".

Please quote the appropriate standards. "rename" is atomic, aka there is
no time both the old and the new name exists.

> I'm told that the Linux fs maintainers have now accepted that 
>    open("file.new",O_CREAT);
>    write();
>    close();
>    rename("file.new","file");
> should not result, even after a crash, in "file" containing garbage.

It does not contain gargabe. It is empty.

> If this is the case then all the fsyncs can be taken out again.

Again: Please quote the standard instead of crying. Your view of things
disallows many of the recent improvements in filesystems, so you have to
show evidence. All the databases and other reliable data handing tools
uses fsync since a long time, because the writes may or may not hit the
disk otherwise.

Bastian

-- 
Even historians fail to learn from history -- they repeat the same mistakes.
		-- John Gill, "Patterns of Force", stardate 2534.7