Re: separate log and structure from user data device?

[Nathan Scott <nathans@xxxxxxx>]

On Thu, Jun 08, 2006 at 03:15:39PM +0100, Peter Grandi wrote:
> 
> [ ... curious complaints about write performance ... ]
> 
> >> /* O_DIRECT for "realtime" */
> 
> nathans> You don't need O_DIRECT for realtime these days.
> 
> >> assert( (fid = open(pn,
> >>   O_WRONLY|O_CREAT|O_DIRECT|O_SYNC|O_LARGEFILE)) != -1 );
> 
> nathans> Thats a shocking use of assert. :)
> 
> Thats too kind :-).
> 
> However, I am puzzled as you comment on 'O_DIRECT' but not on
> 'O_SYNC'. IIRC XFS performance, and in particular the strategy

Heh, I completely overlooked the O_SYNC there.

> to used to allocate large extents, is based on hugely delayed
> flushing, and 'O_SYNC' would defeat that, resulting perhaps in

Indeed.

> the allocation of many small extents, and this might impact
> write speed too (more metadata writes). Is this plausible?

Indeed.

cheers.

-- 
Nathan