Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 17 Jan 2006 22:26:05 -0500
From:      Kris Kennaway <kris@obsecurity.org>
To:        Michael Barnett <mbarnett@measuremap.com>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: PAE causing system crashes
Message-ID:  <20060118032604.GA97030@xor.obsecurity.org>
In-Reply-To: <0F01FAC2-09A7-47EC-AB05-B75564FC2433@measuremap.com>
References:  <47A794C0-FE22-46FC-9D81-BCF3359CBD61@measuremap.com> <20060118010418.GA95598@xor.obsecurity.org> <0F01FAC2-09A7-47EC-AB05-B75564FC2433@measuremap.com>

next in thread | previous in thread | raw e-mail | index | archive | help

--GvXjxJ+pjyke8COw
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Jan 17, 2006 at 08:38:08PM -0600, Michael Barnett wrote:
> It ran fine when we were only throwing a few dozen qps at it... when =20
> we tried to throw about 350qps its way, it was much much slower... =20
> slower than our existing server with significantly fewer resources =20
> available.
>=20
> I never could figure out what the issue was... tried recompiling with =20
> different options, allocating more and less memory to mysql (really =20
> to innodb as all our tables are innodb), and a number of system =20
> tweaks to no avail.
>=20
> Are you running it on an opteron, or on a xeon?  I have heard things =20
> go much smoother on AMD hardware.

I've used it for mysql benchmarking (using supersmack) on amd hardware
and got quite good performance from it.  I've not used it with a
'real' database on that machine.

> Anyways.. the damage is done.  We are back on i386 and I just need to =20
> make it go!    Right now i am just running with the stable non-PAE =20
> kernel that limits access to 4G of memory while i work this out.

It could be a driver issue, although since you're not adding drivers
to the PAE config I don't know why this would be.  Still, you could
post your dmesg so that someone might be able to recognize a problem.

Try looking for an updated BIOS.  It could be that your HW has a buggy
implementation.

Kris
--GvXjxJ+pjyke8COw
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (FreeBSD)

iD8DBQFDzbVMWry0BWjoQKURAqioAJ45gI8QyE/SVImVVCFky36NTCN4ugCbBrmG
kGmAfb732LVB6vMdK7TQXZU=
=3kcb
-----END PGP SIGNATURE-----

--GvXjxJ+pjyke8COw--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060118032604.GA97030>