From owner-freebsd-questions@FreeBSD.ORG Wed Jan 13 17:31:47 2010 Return-Path: Delivered-To: questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 23073106566B for ; Wed, 13 Jan 2010 17:31:47 +0000 (UTC) (envelope-from doug@polands.org) Received: from hrndva-omtalb.mail.rr.com (hrndva-omtalb.mail.rr.com [71.74.56.124]) by mx1.freebsd.org (Postfix) with ESMTP id D3A728FC19 for ; Wed, 13 Jan 2010 17:31:46 +0000 (UTC) X-Authority-Analysis: v=1.0 c=1 a=n1bSx-6ldjUA:10 a=tH3TCQxNhl/vWwTfLVWeHw==:17 a=tFhaULID24oPdkNXNgkA:9 a=y5t2unBEo7gTqwY3gNRhfR3mIAsA:4 X-Cloudmark-Score: 0 X-Originating-IP: 75.87.219.217 Received: from [75.87.219.217] ([75.87.219.217:54409] helo=haran.polands.org) by hrndva-oedge03.mail.rr.com (envelope-from ) (ecelerity 2.2.2.39 r()) with ESMTP id 0F/42-05182-1830E4B4; Wed, 13 Jan 2010 17:31:46 +0000 Received: from email.polands.org (ammon.polands.org [172.16.1.7]) by haran.polands.org (8.14.3/8.14.3) with ESMTP id o0DHVjmw087551 for ; Wed, 13 Jan 2010 11:31:45 -0600 (CST) (envelope-from doug@polands.org) Received: from 209.103.214.35 (SquirrelMail authenticated user djp) by email.polands.org with HTTP; Wed, 13 Jan 2010 11:31:45 -0600 Message-ID: <8418112cdfada93d83ca0cb5307c1d21.squirrel@email.polands.org> Date: Wed, 13 Jan 2010 11:31:45 -0600 From: "Doug Poland" To: questions@freebsd.org User-Agent: SquirrelMail/1.4.20-RC2 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal Cc: Subject: 8.0-R-p2 ZFS: unixbench causing kmem exhaustion panic X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 13 Jan 2010 17:31:47 -0000 Hello, I'm trying to get an 8.0-RELEASE-p2 amd64 box to not crash when running benchmarks/unixbench. The box in question has 4GB RAM running 6 SCSI disks in a RAID1Z. dmesg | grep memory real memory = 4294967296 (4096 MB) avail memory = 3961372672 (3777 MB) zpool status pool: bethesda state: ONLINE scrub: none requested config: NAME STATE READ WRITE CKSUM bethesda ONLINE 0 0 0 raidz1 ONLINE 0 0 0 gpt/disk0 ONLINE 0 0 0 gpt/disk1 ONLINE 0 0 0 gpt/disk2 ONLINE 0 0 0 gpt/disk3 ONLINE 0 0 0 gpt/disk4 ONLINE 0 0 0 gpt/disk5 ONLINE 0 0 0 It appears unixbench causes the mem exhaustion when running the fstime / fsbuffer / fsdisk programs, depending on what I've got in /boot/loader.conf I began with a system with no tunables in /boot/loader.conf (vm.kmem_size and vm.kmem_size_max). Then I tried increasing vm.kmem_size and vm.kmem_size_max a GB at a time, until I was at 4GB. At every increase, the system panicked with a kmem exhaustion, until I used the 4GB settings. At that time, the system system became unresponsive and had to be reset. So the question is, can ZFS be tuned to not panic or hang no matter what I throw at it? In this case, it's the ancient and innocuous unixbench utility. This is a test box right now and I'm more than willing to try various tests or tweaks to get 8.x FreeBSD/ZFS into a stable state. -- Regards, Doug