xfs
[Top] [All Lists]

Re: realtime ( was RE: unexpected high fragmentation, any ideas? )

To: Scott Fagg <scott.fagg@xxxxxxxxxxx>, Steve Lord <lord@xxxxxxx>
Subject: Re: realtime ( was RE: unexpected high fragmentation, any ideas? )
From: Nathan Scott <nathans@xxxxxxx>
Date: Thu, 7 Apr 2005 11:58:37 +1000
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <425491C2.50108@xxxxxxx>
References: <2B6FD80A38020542803BA5DC9B3C9754224C7E@xxxxxxxxxxxxxxxxxxxxxxxx> <425491C2.50108@xxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.5.3i
On Wed, Apr 06, 2005 at 08:49:54PM -0500, Steve Lord wrote:
>  o In order to get a file to live on the realtime subvolume, you have
>    to use a special ioctl call on the file descriptor after it is
>    opened and before any space is allocated.

Actually, theres an easy way to do this now which doesn't require
applications to know about realtime volumes - a directory can have
the "realtime-inherit" flag set on it (e.g. via xfs_io) and files
subsequently created in that directory will automatically be marked
as realtime (at creation time).

I would be surprised if there aren't one or two bugs remaining in
the realtime port for Linux, as Steve pointed out, but there is a
decent foundation with years of testing on IRIX for the core code,
so it shouldn't be too hard to knock out any remaining issues (if
someone wanted to give it a go..)

cheers.

-- 
Nathan


<Prev in Thread] Current Thread [Next in Thread>