Date: Thu, 5 Apr 2007 13:28:38 -0400 From: Kris Kennaway <kris@obsecurity.org> To: Don O'Neil <lists@lizardhill.com> Cc: freebsd-questions@freebsd.org Subject: Re: Problems with SMP on 6.1-STABLE-200608 Message-ID: <20070405172838.GA7857@xor.obsecurity.org> In-Reply-To: <00de01c777a1$c9bf83c0$0600020a@mickey> References: <447it12z00.fsf@be-well.ilk.org> <Pine.LNX.4.43.0703280847370.32025@hymn05.u.washington.edu> <00de01c777a1$c9bf83c0$0600020a@mickey>
next in thread | previous in thread | raw e-mail | index | archive | help
--J2SCkAp4GZ/dPZZf Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Apr 05, 2007 at 09:45:15AM -0700, Don O'Neil wrote: > More info on my problem..... >=20 > I swapped out the MB, CPU's, RAM, Power Supply and I still have the probl= em > with the kernel panicing when running on SMP. >=20 > When I re-build the kernel for NO SMP, the machine is rock solid, even un= der > VERY high loads. >=20 > I setup the old MB, CPU's, RAM & Power Supply on the bench, with a new > 6.1-STABLE-200608 AND 6.2-RELEASE install and run dozens of copies of the > stress port. Even with it bringing loads up to >250, and eating up all > available RAM and SWAP I could not get the kernel to panic. >=20 > The ONLY difference between the bench setup and the production setup is a > 3-Ware Escalade RAID card. I am going to setup another array on the bench > with a spare card I have and see if I can get it to panic under that setup > (which will be identical hardware wise to the production box). The only > thing I can think of right now is one of the following: >=20 > 1) Bad RAID card or cables <- unlikely since it should show up even in > uniprocessor mode > 2) Problem with the TWE driver in SMP mode <- more likely >=20 > I'm leaning towards #2, especially with the other recent reports of someo= ne > else getting kernel panics with 3ware products.=20 >=20 > Anyone else have any thoughts as to what scenarios/tools I should try to > isolate the problem? Update to 6.2-RELEASE or RELENG_6. It's really very easy, and you will quickly be able to evaluate whether someone has already fixed this bug. If it persists, then we have a known data point to proceed with fixing it. FYI I run twe on an extremely heavily loaded SMP system (master build server for the package builds) and it has had no relevant driver issues for at least the past 2 or 3 years that I can recall. Kris --J2SCkAp4GZ/dPZZf Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (FreeBSD) iD8DBQFGFTHGWry0BWjoQKURAvRZAKCfJMQFp9JZSlr8GLmYs4JQSdjnvQCfU0tm uRYl7+ETFvlTMjhD1HGmDoE= =J9b0 -----END PGP SIGNATURE----- --J2SCkAp4GZ/dPZZf--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070405172838.GA7857>