xfs
[Top] [All Lists]

Re: [PATCH 15/22] xfsdocs: document the operation of the realtime device

To: "Darrick J. Wong" <darrick.wong@xxxxxxxxxx>
Subject: Re: [PATCH 15/22] xfsdocs: document the operation of the realtime device and inodes
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Tue, 5 Jan 2016 13:36:03 +1100
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20160105001450.GF28330@xxxxxxxxxxxxxxxx>
References: <20151219091208.15558.65013.stgit@xxxxxxxxxxxxxxxx> <20151219091351.15558.51417.stgit@xxxxxxxxxxxxxxxx> <20160103134151.GB9144@xxxxxxxxxxxxx> <20160105001450.GF28330@xxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Mon, Jan 04, 2016 at 04:14:50PM -0800, Darrick J. Wong wrote:
> On Sun, Jan 03, 2016 at 05:41:51AM -0800, Christoph Hellwig wrote:
> > Looks generally good, although I wonder the take on the realtime device
> > is a bit too positive given that we generally advise people not to use
> > it.
> 
> Last time I tried formatting with a rt device (a month ago) the kernel 
> crashed,
> as soon as I created an inode, so I guess we ought to warn people that this
> feature is severely bitrotted.
> 
> What /is/ the official status of XFS RT?  Most sane distros (looking right at
> you, Ubuntu) disable kernel support, so I guess:
> 
> ------------------------------------------------------------------------
> WARNING: The realtime feature is not well maintained.  Use with caution.
> ------------------------------------------------------------------------

It gets quite a long way through xfstests here before there's
an assert failure in the rt summary locking code. I'm planning on
fixing that when I get a chance and then running xfstests on it
regularly.

I also have patches for xfstests that allow the config section
code to use external devices, which makes it somewhat easier to
have a "forget+test" config....

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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