From owner-freebsd-stable@FreeBSD.ORG Sun May 17 07:30:38 2015 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 1EAD3C45 for ; Sun, 17 May 2015 07:30:38 +0000 (UTC) Received: from hermes.heuristicsystems.com.au (hermes.heuristicsystems.com.au [203.41.22.115]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "hermes.heuristicsystems.com.au", Issuer "Heuristic Systems Type 4 Host CA" (not verified)) by mx1.freebsd.org (Postfix) with ESMTPS id 8F0361DEC for ; Sun, 17 May 2015 07:30:36 +0000 (UTC) Received: from [10.0.5.3] (ewsw01.hs [10.0.5.3]) (authenticated bits=0) by hermes.heuristicsystems.com.au (8.14.6/8.13.6) with ESMTP id t4H7SwDe078079 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT) for ; Sun, 17 May 2015 17:29:03 +1000 (EST) (envelope-from dewayne.geraghty@heuristicsystems.com.au) Message-ID: <55584366.7080708@heuristicsystems.com.au> Date: Sun, 17 May 2015 17:29:42 +1000 From: Dewayne Geraghty User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:31.0) Gecko/20100101 Thunderbird/31.6.0 MIME-Version: 1.0 To: FreeBSD Stable Mailing List Subject: System load greater than logical cores and idprio 30 getting time. Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 17 May 2015 07:30:38 -0000 I don't understand why processes running at idpro 30 are gettting as much CPU time as stated below when the load average is greater than the number of logical cores. I don't think its just a measuring error as I have other short lived processes that also complete at idprio 31. The scheduler for this server is kern.sched.name: ULE kern.sched.interact: 0 This top extract was taken from a Xeon E3-1230L with 8 logical cores, running 10.1Stable built/installed May 6,2015. -- top extract -- last pid: 83135; load averages: 11.32, 8.98, 6.24 up 1+01:33:51 14:42:19 81 processes: 13 running, 68 sleeping CPU: 95.9% user, 0.0% nice, 4.1% system, 0.0% interrupt, 0.0% idle Mem: 299M Active, 5507M Inact, 1351M Wired, 5580K Cache, 899M Buf, 568M Free Swap: 3881M Total, 16M Used, 3865M Free PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 40867 root 1 154 i30 50316K 4720K RUN 2 17:59 42.97% ssh-keygen 40869 root 1 154 i30 50316K 4728K RUN 6 17:49 42.48% ssh-keygen 83099 root 1 76 0 86468K 70696K RUN 4 0:02 16.46% c++ And couple of hours later (note not WCPU) last pid: 21436; load averages: 20.13, 14.73, 12.00 up 1+04:15:45 17:24:13 101 processes: 23 running, 78 sleeping CPU: 95.6% user, 0.0% nice, 4.4% system, 0.0% interrupt, 0.0% idle Mem: 724M Active, 5292M Inact, 1241M Wired, 32M Cache, 847M Buf, 443M Free Swap: 3881M Total, 21M Used, 3861M Free PID USERNAME THR PRI NICE SIZE RES STATE C TIME CPU COMMAND 20141 root 1 155 i31 83084K 37688K RUN 7 0:09 25.98% ssh-keygen 21380 root 1 79 0 110M 101M CPU3 3 0:03 24.46% c++ 21340 root 1 78 0 74492K 66468K CPU4 4 0:03 20.46% c++ I recall the BSD scheduler behaving as expected, ie when the load > cores, then idprio 31 processes barely received any time. Regards, Dewayne.