From owner-freebsd-performance@FreeBSD.ORG Thu Jan 31 19:58:19 2008 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 0189316A41A; Thu, 31 Jan 2008 19:58:18 +0000 (UTC) (envelope-from mike@sentex.net) Received: from smarthost2.sentex.ca (smarthost2.sentex.ca [205.211.164.50]) by mx1.freebsd.org (Postfix) with ESMTP id 9D22913C43E; Thu, 31 Jan 2008 19:58:17 +0000 (UTC) (envelope-from mike@sentex.net) Received: from lava.sentex.ca (pyroxene.sentex.ca [199.212.134.18]) by smarthost2.sentex.ca (8.14.2/8.14.2) with ESMTP id m0VJwGAn039689; Thu, 31 Jan 2008 14:58:17 -0500 (EST) (envelope-from mike@sentex.net) Received: from mdt-xp.sentex.net (simeon.sentex.ca [192.168.43.27]) by lava.sentex.ca (8.13.8/8.13.3) with ESMTP id m0VJwF0f090920 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 31 Jan 2008 14:58:16 -0500 (EST) (envelope-from mike@sentex.net) Message-Id: <200801311958.m0VJwF0f090920@lava.sentex.ca> X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9 Date: Thu, 31 Jan 2008 15:00:20 -0500 To: "Steven Hartland" From: Mike Tancsa In-Reply-To: <015901c863a8$064a5750$b6db87d4@multiplay.co.uk> References: <200801302307.XAA22476@sopwith.solgatos.com> <015901c863a8$064a5750$b6db87d4@multiplay.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Cc: freebsd-performance@freebsd.org, freebsd-stable@freebsd.org Subject: Re: newfs + gstat locks entire machine for 20seconds X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 31 Jan 2008 19:58:19 -0000 At 08:24 PM 1/30/2008, Steven Hartland wrote: >The plot thickens.... This stall is not just related to newfs you have to >have gstat running as well. If I do the newfs without gstat running then >no stall occurs. As soon as Im running gstat while doing the newfs then >everything locks as described. Strange, I see the same thing sometimes. While running while true do date sleep .5 done Thu Jan 31 14:55:42 EST 2008 Thu Jan 31 14:55:42 EST 2008 Thu Jan 31 14:55:43 EST 2008 Thu Jan 31 14:55:43 EST 2008 Thu Jan 31 14:55:44 EST 2008 Thu Jan 31 14:55:44 EST 2008 Thu Jan 31 14:55:50 EST 2008 Thu Jan 31 14:55:50 EST 2008 Thu Jan 31 14:55:51 EST 2008 Thu Jan 31 14:55:51 EST 2008 Thu Jan 31 14:55:52 EST 2008 Thu Jan 31 14:55:52 EST 2008 Thu Jan 31 14:55:53 EST 2008 Thu Jan 31 14:55:53 EST 2008 Thu Jan 31 14:55:54 EST 2008 You can see it from 44 to 50 seconds, This is AMD64 da0 at arcmsr0 bus 0 target 0 lun 0 da0: Fixed Direct Access SCSI-5 device da0: 166.666MB/s transfers (83.333MHz DT, offset 32, 16bit) da0: 305175MB (624999424 512 byte sectors: 255H 63S/T 38904C) arcmsr0: mem 0xe8600000-0xe8600fff,0xe8000000-0xe83fffff irq 18 at device 14.0 on pci2 ARECA RAID ADAPTER0: Driver Version 1.20.00.15 2007-10-07 ARECA RAID ADAPTER0: FIRMWARE VERSION V1.43 2007-4-17 arcmsr0: [ITHREAD] >Running truss on gstat shows the issue / cause I believe but I dont >know what it means:- >[truss -o t.txt -p 61629 -d] >9.008933817 nanosleep({1.000000000}) = 0 (0x0) >9.008969017 gettimeofday({1201742426.147393},0x0) = 0 (0x0) >9.009009804 poll({0/POLLIN},1,0) = 0 (0x0) >9.009040534 gettimeofday({1201742426.147465},0x0) = 0 (0x0) >9.009076852 clock_gettime(0,{1201742426.147501706}) = 0 (0x0) >9.009294477 sigaction(SIGTSTP,{ SIG_IGN SA_RESTART ss_t },{ >0x800cb2470 SA_RESTART ss_t }) = 0 (0x0) >9.009335823 poll({0/POLLIN},1,0) = 0 (0x0) >9.009387785 poll({0/POLLIN},1,0) = 0 (0x0) >9.009457626 write(1,"\^[[4;11H 5\^[[6C2 32 467.8"...,213) = 213 (0xd5) >9.009488636 sigaction(SIGTSTP,{ 0x800cb2470 SA_RESTART ss_t },0x0) = 0 (0x0) >10.009930312 nanosleep({1.000000000}) = 0 (0x0) >10.009963836 gettimeofday({1201742427.148388},0x0) = 0 (0x0) >10.010005182 poll({0/POLLIN},1,0) = 0 (0x0) >10.010036192 gettimeofday({1201742427.148461},0x0) = 0 (0x0) >10.010073068 clock_gettime(0,{1201742427.148497922}) = 0 (0x0) >10.010292369 >mmap(0x801000000,1048576,PROT_READ|PROT_WRITE,MAP_PRIVATE|MAP_ANON,-1,0x0) >= 34376515584 (0x801000000) >10.010327569 >__sysctl(0x7fffffffe6c0,0x2,0x7fffffffe650,0x7fffffffe6b8,0x800844970,0x11) >= 0 (0x0) >25.052947791 >__sysctl(0x7fffffffe650,0x3,0x801000000,0x7fffffffe720,0x0,0x0) = 0 (0x0) >25.054030610 munmap(0x801000000,1048576) = 0 (0x0) >25.055022356 sigaction(SIGTSTP,{ SIG_IGN SA_RESTART ss_t },{ >0x800cb2470 SA_RESTART ss_t }) = 0 (0x0) >25.055067892 poll({0/POLLIN},1,0) = 0 (0x0) >25.055130470 poll({0/POLLIN},1,0) = 0 (0x0) >25.055230203 write(1,"\^[[4;11H1\^[[7C4 64 203.4"...,203) = 203 (0xcb) >25.055263448 sigaction(SIGTSTP,{ 0x800cb2470 SA_RESTART ss_t },0x0) = 0 (0x0) >26.055866597 nanosleep({1.000000000}) = 0 (0x0) >26.055900400 gettimeofday({1201742443.194324},0x0) = 0 (0x0) >26.055940070 poll({0/POLLIN},1,0) = 0 (0x0) >26.055969962 gettimeofday({1201742443.194394},0x0) = 0 (0x0) >26.056009073 clock_gettime(0,{1201742443.194433649}) = 0 (0x0) >26.056240388 sigaction(SIGTSTP,{ SIG_IGN SA_RESTART ss_t },{ >0x800cb2470 SA_RESTART ss_t }) = 0 (0x0) >26.056280896 poll({0/POLLIN},1,0) = 0 (0x0) >26.056334534 poll({0/POLLIN},1,0) = 0 (0x0) >26.056420299 poll({0/POLLIN},1,0) = 0 (0x0) >26.056485112 write(1,"\^[[1;6H6.046s w: 1.000s\^[[4;5"...,305) = 305 (0x131) >26.056516121 sigaction(SIGTSTP,{ 0x800cb2470 SA_RESTART ss_t },0x0) = 0 (0x0) >27.056863372 nanosleep({1.000000000}) = 0 (0x0) >[/truss -o t.txt -p 61629 -d] > >================================================ >This e.mail is private and confidential between Multiplay (UK) Ltd. >and the person or entity to whom it is addressed. In the event of >misdirection, the recipient is prohibited from using, copying, >printing or otherwise disseminating it or any information contained in it. >In the event of misdirection, illegible or incomplete transmission >please telephone +44 845 868 1337 >or return the E.mail to postmaster@multiplay.co.uk. > >_______________________________________________ >freebsd-performance@freebsd.org mailing list >http://lists.freebsd.org/mailman/listinfo/freebsd-performance >To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org"