From owner-freebsd-afs@FreeBSD.ORG Tue May 18 15:59:31 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 084541065675 for ; Tue, 18 May 2010 15:59:31 +0000 (UTC) (envelope-from me@janh.de) Received: from mailhost.uni-hamburg.de (mailhost.uni-hamburg.de [134.100.32.155]) by mx1.freebsd.org (Postfix) with ESMTP id 7DD1C8FC20 for ; Tue, 18 May 2010 15:59:29 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by mailhost.uni-hamburg.de (Postfix) with ESMTP id 9F0BB902C8; Tue, 18 May 2010 17:59:28 +0200 (CEST) X-Virus-Scanned: by University of Hamburg (RRZ/mailhost) Received: from mailhost.uni-hamburg.de ([127.0.0.1]) by localhost (mailhost.uni-hamburg.de [127.0.0.1]) (amavisd-new, port 10024) with LMTP id IR5fjVgDbOIJ; Tue, 18 May 2010 17:59:28 +0200 (CEST) Received: from pc861.math.uni-hamburg.de (pc861.math.uni-hamburg.de [134.100.222.11]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) (Authenticated sender: fmjv004) by mailhost.uni-hamburg.de (Postfix) with ESMTPSA id 811449009C; Tue, 18 May 2010 17:59:27 +0200 (CEST) Message-ID: <4BF2B962.50408@janh.de> Date: Tue, 18 May 2010 17:59:30 +0200 From: Jan Henrik Sylvester User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.1.9) Gecko/20100514 Thunderbird/3.0.4 MIME-Version: 1.0 To: Benjamin Kaduk References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: afs-list freebsd Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 15:59:31 -0000 On 05/04/10 18:45, Benjamin Kaduk wrote: > On Tue, 4 May 2010, Jan Henrik Sylvester wrote: > >> On 04/14/2010 18:52, Benjamin Kaduk wrote: >>> Could you send a bit of detail about your machine, though? Is it >>> single-processor, 32-bit, etc.? >> >> I just saw that I never answered this question, sorry. >> >> The machine is a Core2Quad (four cores) currently running on a GENERIC >> FreeBSD-8.0-RELEASE i386. I plan to test amd64 in a few weeks. (All >> the other machines I could use for testing are i386, too, one 8-STABLE >> and the others 8.0-RELEASE.) > > Thanks! I suspect that we are not quite safe for the SMP case, yet. I have upgraded almost all of my machines from i386 to amd64 and would like to repeat the OpenAFS testing -- even though you suspect it is SMP, which applies to all machines available to me. The port (that uses the master branch) currently only builds until: cc -I. -I.. -I../nfs -I/usr/ports/net/openafs-devel/work/openafs/src -I/usr/ports/net/openafs-devel/work/openafs/src/afs -I/usr/ports/net/openafs-devel/work/openafs/src/afs/FBSD -I/usr/ports/net/openafs-devel/work/openafs/src/config -I/usr/ports/net/openafs-devel/work/openafs/src/rx/FBSD -I/usr/ports/net/openafs-devel/work/openafs/src/rxkad -I/usr/ports/net/openafs-devel/work/openafs/src/util -I/usr/ports/net/openafs-devel/work/openafs/src -I/usr/ports/net/openafs-devel/work/openafs/src/afs -I/usr/ports/net/openafs-devel/work/openafs/src/afs/FBSD -I/usr/ports/net/openafs-devel/work/openafs/src/util -I/usr/ports/net/openafs-devel/work/openafs/src/rxkad -I/usr/ports/net/openafs-devel/work/openafs/src/config -I/usr/ports/net/openafs-devel/work/openafs/src/fsint -I/usr/ports/net/openafs-devel/work/openafs/src/vlserver -I/usr/ports/net/openafs-devel/work/openafs/include -I/usr/ports/net/openafs-devel/work/openafs/include/afs -O -I. -I.. -I/usr/ports/net/openafs-devel/work/openafs/src/config -DAFSDEBUG -DKERNEL -DAFS -DVICE -DNFS -DUFS -DINET -DQUOTA -DGETMOUNT -Wall -nostdinc -I/usr/include -D_KERNEL -DKLD_MODULE -elf -mno-align-long-strings -fno-common -ffreestanding -I/usr/obj/usr/src/sys/GENERIC -include opt_global.h -fno-strict-aliasing -fPIC -O2 -o afs_atomlist.o -c /usr/ports/net/openafs-devel/work/openafs/src/util/afs_atomlist.c :14600381:116: error: opt_global.h: No such file or directory *** Error code 1 Stop in /usr/ports/net/openafs-devel/work/openafs/src/libafs/MODLOAD. Do you have a git tag (or whatever, I do not really know git) that is known to compile? How could I make the port use anything else but master? (Probably "make extract ; cd work/openafs/ ; SOME GIT COMMAND ; cd ../../ ; make" or something similar.) Thanks, Jan Henrik From owner-freebsd-afs@FreeBSD.ORG Tue May 18 16:10:15 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 27D9C106566B for ; Tue, 18 May 2010 16:10:15 +0000 (UTC) (envelope-from kaduk@mit.edu) Received: from dmz-mailsec-scanner-7.mit.edu (DMZ-MAILSEC-SCANNER-7.MIT.EDU [18.7.68.36]) by mx1.freebsd.org (Postfix) with ESMTP id D20B38FC17 for ; Tue, 18 May 2010 16:10:14 +0000 (UTC) X-AuditID: 12074424-b7b9dae000002832-fe-4bf2bbe5ca64 Received: from mailhub-auth-1.mit.edu (MAILHUB-AUTH-1.MIT.EDU [18.9.21.35]) by dmz-mailsec-scanner-7.mit.edu (Symantec Brightmail Gateway) with SMTP id 36.48.10290.5EBB2FB4; Tue, 18 May 2010 12:10:14 -0400 (EDT) Received: from outgoing.mit.edu (OUTGOING-AUTH.MIT.EDU [18.7.22.103]) by mailhub-auth-1.mit.edu (8.13.8/8.9.2) with ESMTP id o4IGADgV028312; Tue, 18 May 2010 12:10:13 -0400 Received: from multics.mit.edu (MULTICS.MIT.EDU [18.187.1.73]) (authenticated bits=56) (User authenticated as kaduk@ATHENA.MIT.EDU) by outgoing.mit.edu (8.13.6/8.12.4) with ESMTP id o4IGABwQ017491 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Tue, 18 May 2010 12:10:12 -0400 (EDT) Received: (from kaduk@localhost) by multics.mit.edu (8.12.9.20060308) id o4IGABOj024320; Tue, 18 May 2010 12:10:11 -0400 (EDT) Date: Tue, 18 May 2010 12:10:10 -0400 (EDT) From: Benjamin Kaduk To: Jan Henrik Sylvester In-Reply-To: <4BF2B962.50408@janh.de> Message-ID: References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> User-Agent: Alpine 1.10 (GSO 962 2008-03-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-Brightmail-Tracker: AAAAARQqCq8= Cc: afs-list freebsd Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 16:10:15 -0000 On Tue, 18 May 2010, Jan Henrik Sylvester wrote: > On 05/04/10 18:45, Benjamin Kaduk wrote: >> On Tue, 4 May 2010, Jan Henrik Sylvester wrote: >> > I have upgraded almost all of my machines from i386 to amd64 and would like > to repeat the OpenAFS testing -- even though you suspect it is SMP, which > applies to all machines available to me. I expect that you will find AFS more usable if you set kern.smp.disabled=1 in /boot/loader.conf, which will make your machine effectively single-processor. > > The port (that uses the master branch) currently only builds until: [...] > -fno-common -ffreestanding -I/usr/obj/usr/src/sys/GENERIC -include > opt_global.h -fno-strict-aliasing -fPIC -O2 -o afs_atomlist.o -c > /usr/ports/net/openafs-devel/work/openafs/src/util/afs_atomlist.c > :14600381:116: error: opt_global.h: No such file or directory > *** Error code 1 > Stop in /usr/ports/net/openafs-devel/work/openafs/src/libafs/MODLOAD. > opt_global.h should be coming from your kernel's build tree ... > Do you have a git tag (or whatever, I do not really know git) that is known > to compile? > I compiled recently with head at commit af6b80f257 > How could I make the port use anything else but master? (Probably "make > extract ; cd work/openafs/ ; SOME GIT COMMAND ; cd ../../ ; make" or > something similar.) I would change the makefile to have the do-patch: target actually do something, i.e.: -------------- do-patch: cd ${WRKDIR} && git checkout af6b80f257 -------------- Note that that must be a tab character, as per all Makefiles. I don't think any of the changes between af6b80f and the current master could cause your build failure, though. -Ben From owner-freebsd-afs@FreeBSD.ORG Tue May 18 16:11:54 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 07928106567A for ; Tue, 18 May 2010 16:11:54 +0000 (UTC) (envelope-from shadow@gmail.com) Received: from mail-pw0-f54.google.com (mail-pw0-f54.google.com [209.85.160.54]) by mx1.freebsd.org (Postfix) with ESMTP id C9ED08FC0C for ; Tue, 18 May 2010 16:11:53 +0000 (UTC) Received: by pwi9 with SMTP id 9so3624125pwi.13 for ; Tue, 18 May 2010 09:11:53 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:received:in-reply-to :references:date:message-id:subject:from:to:cc:content-type :content-transfer-encoding; bh=Pa1Bd6Pwjcp65PE1IP8qCJHJdPFq2JSHywXXdzwc52o=; b=cJX5jXori28ZZaGiVRFPobUqho7f+DN/fIWF6g7qrrQIT3YqBBnTcTRVERS0tE4YT0 bqNxBjDiND6fANkFo7/T31u+8RagYBsd6B8TsgPf2QbAljr1orKgW2uiAUjRQyQ6QWDG Iio4o+zC+qdZ1sMVIjpvs7FVSxqdM9t1U98I4= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; b=CKkSqCTs6GsXvoKyn7ewOBlXgPxfgE2SSDL22HxCe2EBlfWvK+OKTMAMEjTPUH3d0a kS9BaSLeKdARyw7WfmAD/FI2OkV/52oqYdjvdpU4vEqGrKQTGY+UzyBCiHkS/Qya9Ngn wvjhGKwxu4oyJXeWiSykf10f982ZIqXghWib8= MIME-Version: 1.0 Received: by 10.141.14.15 with SMTP id r15mr5155359rvi.139.1274199113231; Tue, 18 May 2010 09:11:53 -0700 (PDT) Received: by 10.140.161.18 with HTTP; Tue, 18 May 2010 09:11:53 -0700 (PDT) In-Reply-To: <4BF2B962.50408@janh.de> References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> Date: Tue, 18 May 2010 12:11:53 -0400 Message-ID: From: Derrick Brashear To: Jan Henrik Sylvester Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Cc: afs-list freebsd , Benjamin Kaduk Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 16:11:54 -0000 On Tue, May 18, 2010 at 11:59 AM, Jan Henrik Sylvester wrote: > On 05/04/10 18:45, Benjamin Kaduk wrote: >> >> On Tue, 4 May 2010, Jan Henrik Sylvester wrote: >> >>> On 04/14/2010 18:52, Benjamin Kaduk wrote: >>>> >>>> Could you send a bit of detail about your machine, though? Is it >>>> single-processor, 32-bit, etc.? >>> >>> I just saw that I never answered this question, sorry. >>> >>> The machine is a Core2Quad (four cores) currently running on a GENERIC >>> FreeBSD-8.0-RELEASE i386. I plan to test amd64 in a few weeks. (All >>> the other machines I could use for testing are i386, too, one 8-STABLE >>> and the others 8.0-RELEASE.) >> >> Thanks! I suspect that we are not quite safe for the SMP case, yet. > > I have upgraded almost all of my machines from i386 to amd64 and would li= ke > to repeat the OpenAFS testing -- even though you suspect it is SMP, which > applies to all machines available to me. > > The port (that uses the master branch) currently only builds until: > cc -I. -I.. -I../nfs =A0-I/usr/ports/net/openafs-devel/work/openafs/src > -I/usr/ports/net/openafs-devel/work/openafs/src/afs > -I/usr/ports/net/openafs-devel/work/openafs/src/afs/FBSD > -I/usr/ports/net/openafs-devel/work/openafs/src/config > -I/usr/ports/net/openafs-devel/work/openafs/src/rx/FBSD > -I/usr/ports/net/openafs-devel/work/openafs/src/rxkad > -I/usr/ports/net/openafs-devel/work/openafs/src/util > -I/usr/ports/net/openafs-devel/work/openafs/src > -I/usr/ports/net/openafs-devel/work/openafs/src/afs > -I/usr/ports/net/openafs-devel/work/openafs/src/afs/FBSD > -I/usr/ports/net/openafs-devel/work/openafs/src/util > -I/usr/ports/net/openafs-devel/work/openafs/src/rxkad > -I/usr/ports/net/openafs-devel/work/openafs/src/config > -I/usr/ports/net/openafs-devel/work/openafs/src/fsint > -I/usr/ports/net/openafs-devel/work/openafs/src/vlserver > -I/usr/ports/net/openafs-devel/work/openafs/include > -I/usr/ports/net/openafs-devel/work/openafs/include/afs =A0-O -I. -I.. > -I/usr/ports/net/openafs-devel/work/openafs/src/config =A0-DAFSDEBUG -DKE= RNEL > -DAFS -DVICE -DNFS -DUFS -DINET -DQUOTA -DGETMOUNT -Wall -nostdinc > -I/usr/include -D_KERNEL -DKLD_MODULE =A0-elf -mno-align-long-strings > -fno-common -ffreestanding -I/usr/obj/usr/src/sys/GENERIC -include > opt_global.h -fno-strict-aliasing -fPIC -O2 =A0-o afs_atomlist.o -c > /usr/ports/net/openafs-devel/work/openafs/src/util/afs_atomlist.c > :14600381:116: error: opt_global.h: No such file or directo= ry You get to provide that, from your kernel build. It doesn't come from openafs. No version of OpenAFS will ever provide that. configure --with-bsd-kernel-build=3D/usr/obj/usr/src/sys/GENERIC or whatever path to your kernel. From owner-freebsd-afs@FreeBSD.ORG Tue May 18 17:54:52 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 059A4106566C for ; Tue, 18 May 2010 17:54:52 +0000 (UTC) (envelope-from me@janh.de) Received: from mailhost.uni-hamburg.de (mailhost.uni-hamburg.de [134.100.32.155]) by mx1.freebsd.org (Postfix) with ESMTP id 2EE558FC1F for ; Tue, 18 May 2010 17:54:50 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by mailhost.uni-hamburg.de (Postfix) with ESMTP id 2DA6E902F8; Tue, 18 May 2010 19:54:50 +0200 (CEST) X-Virus-Scanned: by University of Hamburg (RRZ/mailhost) Received: from mailhost.uni-hamburg.de ([127.0.0.1]) by localhost (mailhost.uni-hamburg.de [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 0eu8zcVvmhcA; Tue, 18 May 2010 19:54:50 +0200 (CEST) Received: from pc861.math.uni-hamburg.de (pc861.math.uni-hamburg.de [134.100.222.11]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) (Authenticated sender: fmjv004) by mailhost.uni-hamburg.de (Postfix) with ESMTPSA id 0EC64902DA; Tue, 18 May 2010 19:54:49 +0200 (CEST) Message-ID: <4BF2D461.8030302@janh.de> Date: Tue, 18 May 2010 19:54:41 +0200 From: Jan Henrik Sylvester User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.1.9) Gecko/20100514 Thunderbird/3.0.4 MIME-Version: 1.0 To: Benjamin Kaduk References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: afs-list freebsd Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 17:54:52 -0000 On 05/18/10 18:10, Benjamin Kaduk wrote: > opt_global.h should be coming from your kernel's build tree ... Yeah, sorry, I asked the same question before. (I never did a buildworld or buildkernel on this machine after installing from the release build and using freebsd-update.) I try to remember next time. > On Tue, 18 May 2010, Jan Henrik Sylvester wrote: >> I have upgraded almost all of my machines from i386 to amd64 and would >> like to repeat the OpenAFS testing -- even though you suspect it is >> SMP, which applies to all machines available to me. On amd64, the listing of the directories was instant unlike on i386. Reading (small) files caused no problems, but I was able to crash the machine accessing some weird entry (some object that OpenAFS on Linux shows with "?" for the file permissions) and -- more realistically -- deadlock all access to afs upon trying to copy a large file to afs. > I expect that you will find AFS more usable if you set > kern.smp.disabled=1 in /boot/loader.conf, which will make your machine > effectively single-processor. I just tried that and copying a large file to afs caused a deadlock of afs, too. I guess it is not just SMP. I have not tried copying smaller files or starting afsd with different parameters than suggested in the ports README (-dynroot -fakestat-all -afsdb -memcache -daemons 4). Cheers, Jan Henrik From owner-freebsd-afs@FreeBSD.ORG Tue May 18 18:14:56 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 84A2B1065674 for ; Tue, 18 May 2010 18:14:56 +0000 (UTC) (envelope-from shadow@gmail.com) Received: from mail-pw0-f54.google.com (mail-pw0-f54.google.com [209.85.160.54]) by mx1.freebsd.org (Postfix) with ESMTP id 54AFD8FC27 for ; Tue, 18 May 2010 18:14:55 +0000 (UTC) Received: by pwi9 with SMTP id 9so3690697pwi.13 for ; Tue, 18 May 2010 11:14:55 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:received:in-reply-to :references:date:message-id:subject:from:to:cc:content-type; bh=roaPUY1GY2EYYVxCh3U7hCWrINOQ5JC4fyN/eQCEeDc=; b=r4iqgxYPT/TH/R/0tU0LoDkxErQJuGV48hOe3dLsx9F4wNnixAnQJ46LpiR0c/nd7k KN8Jol5EVlHaL9EOKvzgUL2pCowf40mv4PZwl8NS5bcms4ZdJrbik1aqWKrNYPh+Pivg nDTwbt98+E4ntMPViYNXkHUKGaBCC7OM4GMGs= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=Og+hAFZLaM1wQlsTjKxPOifEbINQaMt5UzKwqfVE4vY9Pa7OqJs0JEm28p3pJoO/+P QE5QFwWNUbVkWgfAnTtCvqIeW0TwypPllmmSOMc5BpII8195EZpHcCw3j5Rh7EROmdEP s50E7GAPXR6xGMNdEsV0xjUBiVBC1aM0j4ZKo= MIME-Version: 1.0 Received: by 10.141.213.5 with SMTP id p5mr5334482rvq.14.1274206495757; Tue, 18 May 2010 11:14:55 -0700 (PDT) Received: by 10.140.161.18 with HTTP; Tue, 18 May 2010 11:14:55 -0700 (PDT) In-Reply-To: <4BF2D461.8030302@janh.de> References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> <4BF2D461.8030302@janh.de> Date: Tue, 18 May 2010 14:14:55 -0400 Message-ID: From: Derrick Brashear To: Jan Henrik Sylvester Content-Type: text/plain; charset=ISO-8859-1 Cc: afs-list freebsd , Benjamin Kaduk Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 18:14:56 -0000 On Tue, May 18, 2010 at 1:54 PM, Jan Henrik Sylvester wrote: > On 05/18/10 18:10, Benjamin Kaduk wrote: >> opt_global.h should be coming from your kernel's build tree ... > > Yeah, sorry, I asked the same question before. (I never did a buildworld or > buildkernel on this machine after installing from the release build and > using freebsd-update.) I try to remember next time. > >> On Tue, 18 May 2010, Jan Henrik Sylvester wrote: >>> >>> I have upgraded almost all of my machines from i386 to amd64 and would >>> like to repeat the OpenAFS testing -- even though you suspect it is >>> SMP, which applies to all machines available to me. > > On amd64, the listing of the directories was instant unlike on i386. Reading > (small) files caused no problems, but I was able to crash the machine > accessing some weird entry (some object that OpenAFS on Linux shows with "?" > for the file permissions) and -- more realistically -- deadlock all access > to afs upon trying to copy a large file to afs. does cmdebug (yourhost) return anything? >> I expect that you will find AFS more usable if you set >> kern.smp.disabled=1 in /boot/loader.conf, which will make your machine >> effectively single-processor. > > I just tried that and copying a large file to afs caused a deadlock of afs, > too. I guess it is not just SMP. > > I have not tried copying smaller files or starting afsd with different > parameters than suggested in the ports README (-dynroot -fakestat-all -afsdb > -memcache -daemons 4). -- Derrick From owner-freebsd-afs@FreeBSD.ORG Thu May 20 17:19:53 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 843B7106566B for ; Thu, 20 May 2010 17:19:53 +0000 (UTC) (envelope-from me@janh.de) Received: from mailhost.uni-hamburg.de (mailhost.uni-hamburg.de [134.100.32.155]) by mx1.freebsd.org (Postfix) with ESMTP id 03E918FC14 for ; Thu, 20 May 2010 17:19:52 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by mailhost.uni-hamburg.de (Postfix) with ESMTP id 40C5B90215; Thu, 20 May 2010 19:19:51 +0200 (CEST) X-Virus-Scanned: by University of Hamburg (RRZ/mailhost) Received: from mailhost.uni-hamburg.de ([127.0.0.1]) by localhost (mailhost.uni-hamburg.de [127.0.0.1]) (amavisd-new, port 10024) with LMTP id khAkEI2Jgj+4; Thu, 20 May 2010 19:19:51 +0200 (CEST) Received: from pc861.math.uni-hamburg.de (pc861.math.uni-hamburg.de [134.100.222.11]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) (Authenticated sender: fmjv004) by mailhost.uni-hamburg.de (Postfix) with ESMTPSA id 06A96901E1; Thu, 20 May 2010 19:19:51 +0200 (CEST) Message-ID: <4BF56F38.4040008@janh.de> Date: Thu, 20 May 2010 19:19:52 +0200 From: Jan Henrik Sylvester User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.1.9) Gecko/20100514 Thunderbird/3.0.4 MIME-Version: 1.0 To: Derrick Brashear References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> <4BF2D461.8030302@janh.de> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: afs-list freebsd , Benjamin Kaduk Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 20 May 2010 17:19:53 -0000 On 05/18/10 20:14, Derrick Brashear wrote: > On Tue, May 18, 2010 at 1:54 PM, Jan Henrik Sylvester wrote: >> On amd64, the listing of the directories was instant unlike on i386. Reading >> (small) files caused no problems, but I was able to crash the machine >> accessing some weird entry (some object that OpenAFS on Linux shows with "?" >> for the file permissions) and -- more realistically -- deadlock all access >> to afs upon trying to copy a large file to afs. > > does > cmdebug (yourhost) > return anything? - change to some afs-directory cmdebug: nothing - change to some afs-directory that uses krb login cmdebug: crashes the machine instantly - repeat both steps after reboot cmdebug: nothing - copying a small file cmdebug: reports something (I lost it, looked as if it was expected) - eventually copying of the small file finished (took very long) cmdebug: nothing - copying a large file cmdebug: nothing - copying takes forever (does nothing after a second, no reading) cmdebug: nothing >>> I expect that you will find AFS more usable if you set >>> kern.smp.disabled=1 in /boot/loader.conf, which will make your machine >>> effectively single-processor. >> >> I just tried that and copying a large file to afs caused a deadlock of afs, >> too. I guess it is not just SMP. The "protocol" above was with SMP enabled. I could repeat it with SMP disabled if you expect anything from it. (I did observe the same behavior with SMP disabled.) On a different machine, I was just changing afs directories and listing them calling cmdebug a few times in between. Nothing happened, but when I came back to the machine a few minutes later, it had deadlocked completely -- without copying data at all. I just found the first machine that I tried to reboot with its stalled copy process is now hanging completely, too. I would have tried to narrow down the circumstances that cause deadlocks, if both of the machines in front of me were not crashing so often... Cheers, Jan Henrik From owner-freebsd-afs@FreeBSD.ORG Thu May 20 17:24:07 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 7001F1065677 for ; Thu, 20 May 2010 17:24:07 +0000 (UTC) (envelope-from kaduk@mit.edu) Received: from dmz-mailsec-scanner-1.mit.edu (DMZ-MAILSEC-SCANNER-1.MIT.EDU [18.9.25.12]) by mx1.freebsd.org (Postfix) with ESMTP id 2458B8FC2C for ; Thu, 20 May 2010 17:24:06 +0000 (UTC) X-AuditID: 1209190c-b7bd2ae000005d05-e9-4bf57036bed9 Received: from mailhub-auth-2.mit.edu (MAILHUB-AUTH-2.MIT.EDU [18.7.62.36]) by dmz-mailsec-scanner-1.mit.edu (Symantec Brightmail Gateway) with SMTP id 52.18.23813.63075FB4; Thu, 20 May 2010 13:24:06 -0400 (EDT) Received: from outgoing.mit.edu (OUTGOING-AUTH.MIT.EDU [18.7.22.103]) by mailhub-auth-2.mit.edu (8.13.8/8.9.2) with ESMTP id o4KHO5Iv021696; Thu, 20 May 2010 13:24:06 -0400 Received: from multics.mit.edu (MULTICS.MIT.EDU [18.187.1.73]) (authenticated bits=56) (User authenticated as kaduk@ATHENA.MIT.EDU) by outgoing.mit.edu (8.13.6/8.12.4) with ESMTP id o4KHO2W3027175 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Thu, 20 May 2010 13:24:04 -0400 (EDT) Received: (from kaduk@localhost) by multics.mit.edu (8.12.9.20060308) id o4KHO2AS004918; Thu, 20 May 2010 13:24:02 -0400 (EDT) Date: Thu, 20 May 2010 13:24:02 -0400 (EDT) From: Benjamin Kaduk To: Jan Henrik Sylvester In-Reply-To: <4BF56F38.4040008@janh.de> Message-ID: References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> <4BF2D461.8030302@janh.de> <4BF56F38.4040008@janh.de> User-Agent: Alpine 1.10 (GSO 962 2008-03-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-Brightmail-Tracker: AAAAAA== Cc: afs-list freebsd Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 20 May 2010 17:24:07 -0000 On Thu, 20 May 2010, Jan Henrik Sylvester wrote: > I would have tried to narrow down the circumstances that cause deadlocks, if > both of the machines in front of me were not crashing so often... I fixed a bug that was causing very quick deadlocks on my system, in revision 42a280f50daf6e4dc65873150c4738aacf2c3a86 ( Wed, 19 May 2010 10:39:35 +0000 (03:39 -0700)). Now the most common failure mode I am seeing is kernel panics that seem to be due to some form of memory corruption. -Ben From owner-freebsd-afs@FreeBSD.ORG Thu May 20 17:31:29 2010 Return-Path: Delivered-To: freebsd-afs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id C91FD106566C for ; Thu, 20 May 2010 17:31:29 +0000 (UTC) (envelope-from me@janh.de) Received: from mailhost.uni-hamburg.de (mailhost.uni-hamburg.de [134.100.32.155]) by mx1.freebsd.org (Postfix) with ESMTP id 72B418FC0C for ; Thu, 20 May 2010 17:31:29 +0000 (UTC) Received: from localhost (localhost [127.0.0.1]) by mailhost.uni-hamburg.de (Postfix) with ESMTP id 70C2C90215; Thu, 20 May 2010 19:31:28 +0200 (CEST) X-Virus-Scanned: by University of Hamburg (RRZ/mailhost) Received: from mailhost.uni-hamburg.de ([127.0.0.1]) by localhost (mailhost.uni-hamburg.de [127.0.0.1]) (amavisd-new, port 10024) with LMTP id c7llTw9Ybwl1; Thu, 20 May 2010 19:31:28 +0200 (CEST) Received: from pc861.math.uni-hamburg.de (pc861.math.uni-hamburg.de [134.100.222.11]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) (Authenticated sender: fmjv004) by mailhost.uni-hamburg.de (Postfix) with ESMTPSA id 51C849012E; Thu, 20 May 2010 19:31:28 +0200 (CEST) Message-ID: <4BF571F1.8090706@janh.de> Date: Thu, 20 May 2010 19:31:29 +0200 From: Jan Henrik Sylvester User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.1.9) Gecko/20100514 Thunderbird/3.0.4 MIME-Version: 1.0 To: Benjamin Kaduk References: <558186334.2619.1271175586732.JavaMail.root@thunderbeast.private.linuxbox.com> <4BC4A5F1.6050406@janh.de> <4BC5EEB5.5050109@janh.de> <4BE01AF0.3080309@janh.de> <4BF2B962.50408@janh.de> <4BF2D461.8030302@janh.de> <4BF56F38.4040008@janh.de> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: afs-list freebsd Subject: Re: AFS on FreeBSD 8? X-BeenThere: freebsd-afs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: The Andrew File System and FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 20 May 2010 17:31:29 -0000 On 05/20/10 19:24, Benjamin Kaduk wrote: > On Thu, 20 May 2010, Jan Henrik Sylvester wrote: > >> I would have tried to narrow down the circumstances that cause >> deadlocks, if both of the machines in front of me were not crashing so >> often... > > I fixed a bug that was causing very quick deadlocks on my system, in > revision 42a280f50daf6e4dc65873150c4738aacf2c3a86 ( Wed, 19 May 2010 > 10:39:35 +0000 (03:39 -0700)). Now the most common failure mode I am > seeing is kernel panics that seem to be due to some form of memory > corruption. Thanks for that information. My build is 18 hours older than your commit. I will try again, soon, but not today. Cheers, Jan Henrik