From: Al Boldi <a1426z@gawab.com>
To: Mike Galbraith <efault@gmx.de>
Subject: Re: konqueror deadlocks on 2.6.22
Date: Sun, 20 Jan 2008 08:41:56 +0300
User-Agent: KMail/1.5
References: <200801192114.41427.a1426z@gawab.com> <1200802290.4166.2.camel@homer.simson.net>
In-Reply-To: <1200802290.4166.2.camel@homer.simson.net>
MIME-Version: 1.0
Content-Disposition: inline
Message-Id: <200801200831.04964.a1426z@gawab.com>
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1246
Lines: 36

Mike Galbraith wrote:
> On Sat, 2008-01-19 at 21:14 +0300, Al Boldi wrote:
> > I was just attacked by some deadlock issue involving sqlite3 and
> > konqueror. While sqlite3 continues to slowly fill a 7M-record db in
> > transaction mode, konqueror hangs for a few minutes, then continues only
> > to hang again and again.
> >
> > Looks like an fs/blockIO issue involving fsync.
> >
> > As a workaround, is there a way to make fsync soft?
>
> Do you have the fs mounted data=writeback?  A while back, I ran into
> starvation on the order of minutes with my old/full ext2 fs until
> mounting data=writeback.

You are absolutely right.  With data=writeback the hangs completely 
disappear, and sqlite3 insert performance increases 10x fold.

Now data=writeback is known to be faster than data=ordered, but a 10x fold 
increase probably points to some sync contention within the data=ordered 
logic.  Any ideas how this could be fixed?


Thanks a lot!

BTW Mike: Your server bounces my messages.

--
Al


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/