Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 04 Dec 2008 15:20:20 -0800
From:      "Kevin Oberman" <oberman@es.net>
To:        josh.carroll@gmail.com
Cc:        Alexandr Pakhomov <pahom.rt@gmail.com>, Mike Tancsa <freebsd-stable@freebsd.org>, Steve Polyack <korvus@comcast.net>
Subject:   Re: repeatable crash on RELENG7 
Message-ID:  <20081204232020.7289A45019@ptavv.es.net>
In-Reply-To: Your message of "Thu, 04 Dec 2008 17:11:46 EST." <8cb6106e0812041411t13ca9ffcjc4cc66a33ef43133@mail.gmail.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
--==_Exmh_1228432820_12904P
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

> Date: Thu, 4 Dec 2008 17:11:46 -0500
> From: "Josh Carroll" <josh.carroll@gmail.com>
> 
> > It seems that the term "swap-backed" is misleading for some people.  It does
> > NOT mean your md(4) device will be constantly swapping to disk (and the man
> > page does an alright job of relaying this).  It simply means that generally
> > available memory will be used, and so will swap iff available memory happens
> > to drop low enough.
> >
> > The bottom line in my experience with md(4) devices greater than ~100MB is
> > that "swap-backed" is always reliable, while malloc'd md(4) devices will
> > cause unpredictable kernel panics.
> 
> Using -t swap instead of -t malloc will prevent a panic, but creating
> an md greater than the size of the VM available and filling it will
> cause resource exhaustion and OOM will kick in and start killing
> processes, right? So while it won't panic the box, I still wouldn't
> consider it safe to use unless the size of the md is chosen carefully.
> Someone please correct me if I'm wrong, but I think that'd be the
> behavior if an md with -t swap is used, right?

Yes, but the VM available is just a bit larger than the amount of
KVM. you still need at least a bit of sanity when you create (and fill)
an MD device. 

For performance reasons, I suggest keeping the size of an MD so that it
will fit in RAM. Paging in and out is rather painfully slow.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman@es.net			Phone: +1 510 486-8634
Key fingerprint:059B 2DDF 031C 9BA3 14A4  EADA 927D EBB3 987B 3751

--==_Exmh_1228432820_12904P
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (FreeBSD)
Comment: Exmh version 2.5 06/03/2002

iD8DBQFJOGW0kn3rs5h7N1ERAh9jAJoDr0hX/7rTzLodd6wF8gm0MsfdLACfebHl
FSfNU/zkkOoIxPN91Owe/V0=
=hyxw
-----END PGP SIGNATURE-----

--==_Exmh_1228432820_12904P--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20081204232020.7289A45019>