Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 Feb 2006 18:13:29 -0600 (CST)
From:      Greg Rivers <gcr+freebsd-stable@tharned.org>
To:        Kris Kennaway <kris@obsecurity.org>
Cc:        stable@freebsd.org, "Michael R. Wayne" <freebsd@wayne47.com>
Subject:   Re: Disk I/O system hang on 5.4-RELEASE-p8 i386
Message-ID:  <20060223175345.U12100@w10.sac.fedex.com>
In-Reply-To: <20060223235055.GA93873@xor.obsecurity.org>
References:  <200602231753.k1NHr8c1079056@manor.msen.com> <20060223163849.I12100@w10.sac.fedex.com> <20060223235055.GA93873@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 23 Feb 2006, Kris Kennaway wrote:

>> I believe this issue has made it onto the show-stopper list for
>> 6.1-RELEASE and is being actively worked on.
>
> It's on the todo list, but I don't think it's being worked on yet.
> The main problem is that we need a way to reproduce it on command.
> I'd forgotten that snapshots are involved, so maybe it's just a matter
> of running lots of mksnap_ffs while I/O is in progress.
>
> kris
>

It happens with or without snapshots, but snapshots are a lot more likely 
to make it happen.  In my case, approximately 1 in 3 snapshots will do it. 
Without snapshots, I get a deadlock about every ten days in a population 
of three hosts.

Tor Egge and Don Lewis were kind enough to work with me off-list for a bit 
last December.  They analyzed several of the core files I produced and I 
think they have a fair understanding of what the problems are.  But I 
wouldn't presume to put words in their mouths; perhaps they'll give us an 
update.  I see from the todo list that Tor may already be working on the 
deadlock for amd64.

I'm at the disposal of anyone who's willing to look into this further.

-- 
Greg



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060223175345.U12100>