From lars.tunkrans@bredband.net Fri May 2 00:54:33 2003 Received: with ECARTIS (v1.0.0; list fam); Fri, 02 May 2003 00:54:38 -0700 (PDT) Received: from mf1.bredband.net (mf1.bredband.net [195.54.106.36]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h427sVFu013451 for ; Fri, 2 May 2003 00:54:33 -0700 Received: from bredband.net ([213.112.167.36]) by mf1.bredband.net with ESMTP id <20030502075425.WPBR269.mf1@bredband.net> for ; Fri, 2 May 2003 09:54:25 +0200 Message-ID: <3EB223B2.40502@bredband.net> Date: Fri, 02 May 2003 09:52:18 +0200 From: Lars Tunkrans Organization: None User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.3) Gecko/20030314 X-Accept-Language: en-us, en MIME-Version: 1.0 To: fam@oss.sgi.com Subject: FAM-2.6.10 on SPARC/Solaris 9 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) X-archive-position: 126 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: lars.tunkrans@bredband.net Precedence: bulk X-list: fam Hello I have identified three problems which needs to be fixed If it is going to be possible to ever use FAM on a SUN Workstation. 1) location of mntent.h As Rolf Sponsel has reported earlier: mntent.h is in /usr/include/sys/mntent.h on Solaris 7/8/9 2) Declaration for unsigned integrals. In several places in the source " u_int32_t " is used. Solaris uses the other style " uint32_t " It is arguable that *NIX operatingsystem should be able to handle both as FreeBSD does. Hovewver the /usr/inlclude/sys/types.h in solaris does not include the former style, only the latter of these above. 3) The string "sun" is a reserved word. in ./libfam/Client.c++ the string "sun" is used in eight places. on a SUN/Solaris system the string "sun" is a preprocessor define and cant be used for program variables. The string "sun" is also used in ./fam/ClientConnection.c++ ./fam/ClientConnection.h ./fam/Listener.c++ ./fam/LocalClient.c++ ./fam/LocalClient.h ./fam/TCP_Client.h An maybe other places I have not found. -- ======================================================== Lars Tunkrans From chris@10191.com Fri May 2 16:06:38 2003 Received: with ECARTIS (v1.0.0; list fam); Fri, 02 May 2003 16:06:44 -0700 (PDT) Received: from mx1.ovh.net (b4.ovh.net [213.186.33.54]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h42N6aFu000448 for ; Fri, 2 May 2003 16:06:37 -0700 Received: (qmail 6786 invoked by uid 503); 2 May 2003 23:06:32 -0000 Received: from unknown (HELO 10191.com) (chris%10191.com@81.48.247.182) by ns0.ovh.net with SMTP; 2 May 2003 23:06:32 -0000 Date: Sat, 3 May 2003 01:06:52 +0200 Mime-Version: 1.0 (Apple Message framework v552) Content-Type: text/plain; charset=US-ASCII; format=flowed Subject: Compile error on OSX Jaguar From: Chris To: fam@oss.sgi.com Content-Transfer-Encoding: 7bit Message-Id: X-Mailer: Apple Mail (2.552) X-Spam-Checker-Version: SpamAssassin 2.50 (1.173-2003-02-20-exp) X-archive-position: 127 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: chris@10191.com Precedence: bulk X-list: fam Hello, I'm trying to compile FAM 2.6.10 on OS X 10.2.5 (gcc version 3.1 20020420). For that i applied the freeBSD patch. the ./configure seems works well (without running autoconf). but i have errors with make : any idea ? Thx a lot. C. ---- output --- % make cd . && aclocal-1.6 cd . && \ automake-1.6 --gnu Makefile cd . && autoconf /bin/sh ./config.status --recheck running /bin/sh ./configure --no-create --no-recursion checking for a BSD compatible install... /usr/bin/install -c checking whether build environment is sane... yes /Users/kubernan/Desktop/fam-2.6.10/missing: Unknown `--run' option Try `/Users/kubernan/Desktop/fam-2.6.10/missing --help' for more information configure: WARNING: `missing' script is too old or missing checking for mawk... no checking for gawk... no checking for nawk... no checking for awk... awk checking whether make sets ${MAKE}... yes checking build system type... powerpc-apple-darwin6.5 checking host system type... powerpc-apple-darwin6.5 checking for style of include used by make... GNU checking for gcc... gcc checking for C compiler default output... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for executable suffix... checking for object suffix... o checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking dependency style of gcc... none checking for ld used by GCC... /usr/bin/ld checking if the linker (/usr/bin/ld) is GNU ld... no checking for /usr/bin/ld option to reload object files... -r checking for BSD-compatible nm... /usr/bin/nm -p checking whether ln -s works... yes checking how to recognise dependant libraries... file_magic Mach-O dynamically linked shared library checking command to parse /usr/bin/nm -p output... ok checking how to run the C preprocessor... gcc -E checking for dlfcn.h... no checking for ranlib... ranlib checking for strip... strip checking for objdir... .libs checking for gcc option to produce PIC... -fno-common checking if gcc PIC flag -fno-common works... yes checking if gcc static flag -static works... no checking if gcc supports -c -o file.o... yes checking if gcc supports -c -o file.lo... yes checking if gcc supports -fno-rtti -fno-exceptions... yes checking whether the linker (/usr/bin/ld) supports shared libraries... yes checking how to hardcode library paths into programs... unsupported checking whether stripping libraries is possible... no checking dynamic linker characteristics... darwin6.5 dyld checking if libtool supports shared libraries... yes checking whether to build shared libraries... yes checking whether to build static libraries... yes creating libtool checking for g++... g++ checking whether we are using the GNU C++ compiler... yes checking whether g++ accepts -g... yes checking dependency style of g++... none checking how to run the C++ preprocessor... g++ -E checking for ranlib... (cached) ranlib checking for a BSD compatible install... /usr/bin/install -c checking whether make sets ${MAKE}... (cached) yes checking for tar... /usr/bin/tar checking for ldconfig... no checking for mntent.h... yes checking for linux/imon.h... no Using imon support module IMonNone checking for sys/statvfs.h... no checking for sys/syssgi.h... no checking for sys/fs/nfs_clnt.h... no checking SGI_NOHANG... no checking for _daemonize... no checking for daemon... no checking for getgrmember... no checking for getgroups... yes checking for stat.st_fstype string... no checking for stat.st_ctim.tv_nsec (nanoseconds)... no checking for mountlist struct... no checking for mountlist pointer... yes checking netinet/in.h for bindresvport()... no checking for prmap_sgi_t... no checking for socklen_t... no checking for built-in bool... yes checking for C++ namespace support... yes checking for echo... /bin/echo checking for echo -e flag... /bin/echo configure: creating ./config.status ****************************************************************** * * * Note! Kernel support for imon was not detected. fam should * * still work, but it will have to poll monitored filesystems * * instead of receiving file operation events from the kernel. * * * * If you think your kernel has imon, see what this script is * * looking for (on Linux, probably /usr/include/linux/imon.h), * * fix the problem (perhaps by adding a link), and run this * * again. (You probably need to remove config.cache first.) * * * * See http://oss.sgi.com/projects/fam/ for more information. * * * ****************************************************************** cd . && /bin/sh ./config.status Makefile config.status: creating Makefile cd . && autoheader touch ./config.h.in cd . && /bin/sh ./config.status config.h config.status: creating config.h make all-recursive Making all in util cd .. && \ automake-1.6 --gnu util/Makefile cd .. && /bin/sh ./config.status util/Makefile config.status: error: invalid argument: util/Makefile make[2]: *** [Makefile] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 --- end of outpt --- From jbaxter@badlandsgames.com Sun May 18 15:56:58 2003 Received: with ECARTIS (v1.0.0; list fam); Sun, 18 May 2003 15:57:04 -0700 (PDT) Received: from natasha.asdf456.com (natasha.asdf456.com [64.235.37.2]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4IMuv2x020981 for ; Sun, 18 May 2003 15:56:58 -0700 Received: (qmail 32016 invoked by uid 0); 18 May 2003 22:54:34 -0000 Received: from unknown (HELO badlandsgames.com) (66.92.24.37) by natasha.asdf456.com with SMTP; 18 May 2003 22:54:34 -0000 Message-ID: <3EC80F8E.900@badlandsgames.com> Date: Sun, 18 May 2003 15:56:14 -0700 From: Joel Baxter Reply-To: jbaxter@neogeographica.com User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.3) Gecko/20030304 X-Accept-Language: en-us, en MIME-Version: 1.0 To: fam@oss.sgi.com Subject: fam error "failed to accept new client" Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 128 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: jbaxter@badlandsgames.com Precedence: bulk X-list: fam I'm trying to get fam working on a RedHat 7.1 box (x86). In the messages log I'm seeing this: fam[10247]: failed to accept new client: Invalid argument I've experimented with using the RedHat-target source RPMs for fam 2.6.4, 2.6.7, and 2.6.8. There were no problems with the configure/compilation/installation as far as I can tell. I've tried running both the fam test app as well as the post-compile tests for Courier IMAP (which uses fam, and is the reason I'm trying to get fam installed). In all cases, when I try to run the app, the above error message is generated in the log, and the app fails to work as expected. chkconfig shows the fam service to be "on", and rpcinfo -p gives me: program vers proto port 100000 2 tcp 111 portmapper 100000 2 udp 111 portmapper 100024 1 udp 1024 status 100024 1 tcp 1024 status 391002 2 tcp 1129 sgi_fam So... on the surface this looks copacetic, but obviously something is not working. Any solution, or just a hint as where to start looking, would be appreciated. From jbaxter@badlandsgames.com Sun May 18 16:51:41 2003 Received: with ECARTIS (v1.0.0; list fam); Sun, 18 May 2003 16:51:45 -0700 (PDT) Received: from natasha.asdf456.com (natasha.aplushosting.com [64.235.37.2]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4INpe2x022067 for ; Sun, 18 May 2003 16:51:40 -0700 Received: (qmail 18744 invoked by uid 0); 18 May 2003 23:49:16 -0000 Received: from unknown (HELO badlandsgames.com) (66.92.24.37) by natasha.asdf456.com with SMTP; 18 May 2003 23:49:16 -0000 Message-ID: <3EC81C5E.70704@badlandsgames.com> Date: Sun, 18 May 2003 16:50:54 -0700 From: Joel Baxter Reply-To: jbaxter@neogeographica.com User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.3) Gecko/20030304 X-Accept-Language: en-us, en MIME-Version: 1.0 To: fam@oss.sgi.com Subject: Re: fam error "failed to accept new client" References: <3EC80F8E.900@badlandsgames.com> In-Reply-To: <3EC80F8E.900@badlandsgames.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 129 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: jbaxter@badlandsgames.com Precedence: bulk X-list: fam Joel Baxter wrote: > I'm trying to get fam working on a RedHat 7.1 box (x86). In the > messages log I'm seeing this: > > fam[10247]: failed to accept new client: Invalid argument Well... fixed. I updated my xinetd version, and also tried the RedHat 7.2 binary RPM for fam 2.6.4 rather than compiling it myself. The Courier IMAP post-compile tests now run successfully. Not sure which of the two changes was key, but don't really care at the moment. :-) If someone knows specifically why I was getting that error message, I'd still be interested in hearing it, but it's no longer a blocker. Also, if there's some nasty known problem with 2.6.4, please tell. From michael.wardle@adacel.com Sun May 18 17:35:39 2003 Received: with ECARTIS (v1.0.0; list fam); Sun, 18 May 2003 17:35:46 -0700 (PDT) Received: from adacel.com (gunsmoke.adacel.com.au [210.11.130.7]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4J0Zb2x022526 for ; Sun, 18 May 2003 17:35:38 -0700 Received: from nexus.adacel.com (Not Verified[10.32.240.1]) by adacel.com with MailMarshal (v5,0,3,91) id ; Mon, 19 May 2003 10:32:43 +1000 Received: (qmail 980 invoked from network); 19 May 2003 00:35:11 -0000 Received: from unknown (HELO hera.wodonga.adacel.com.au) (192.168.75.251) by nexus.adacel.com with SMTP; 19 May 2003 00:35:11 -0000 Received: from selene.wodonga.adacel.com.au (selene.wodonga.adacel.com.au [192.168.75.55]) by hera.wodonga.adacel.com.au (8.11.6/8.11.6) with ESMTP id h4J0aME04488; Mon, 19 May 2003 10:36:23 +1000 From: Michael Wardle Organization: Adacel Technologies To: jbaxter@neogeographica.com Subject: Re: fam error "failed to accept new client" Date: Mon, 19 May 2003 10:36:17 +1000 User-Agent: KMail/1.5 References: <3EC80F8E.900@badlandsgames.com> <3EC81C5E.70704@badlandsgames.com> In-Reply-To: <3EC81C5E.70704@badlandsgames.com> Cc: fam@oss.sgi.com MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200305191036.17936.michael.wardle@adacel.com> X-archive-position: 130 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: michael.wardle@adacel.com Precedence: bulk X-list: fam On Monday 19 May 2003 09:50, Joel Baxter wrote: > Joel Baxter wrote: > > I'm trying to get fam working on a RedHat 7.1 box (x86). In the > > messages log I'm seeing this: > > > > fam[10247]: failed to accept new client: Invalid argument > > Well... fixed. I updated my xinetd version, and also tried the RedHat > 7.2 binary RPM for fam 2.6.4 rather than compiling it myself. The > Courier IMAP post-compile tests now run successfully. > > Not sure which of the two changes was key, but don't really care at the > moment. :-) If my memory serves me correctly, there was an issue with xinetd handling RPC version 1 connections for FAM. I think Alex Larsson (the Red Hat FAM maintainer) made a change to the FAM service definition in /etc/xinetd.d/sgi_fam to make it only use RPC version 2. It could also be that upgrading to a later version of xinetd fixed the problem in xinetd. You might like to look at the respective changelogs for more details (example: rpm --query --changelog fam). > If someone knows specifically why I was getting that error message, I'd > still be interested in hearing it, but it's no longer a blocker. Also, > if there's some nasty known problem with 2.6.4, please tell. Users were able to monitor files that belonged to the group the FAM daemon was running as (group nobody by default on Red Hat Linux). It's not a huge issue, but one you should be aware of nonetheless. There were also issues when monitoring several thousand files that have only recently been fixed with a new version of glibc. See the FAM changelog for more details http://oss.sgi.com/projects/fam/ChangeLog. -- Michael Wardle Adacel Technologies From jbaxter@badlandsgames.com Sun May 18 18:57:22 2003 Received: with ECARTIS (v1.0.0; list fam); Sun, 18 May 2003 18:57:26 -0700 (PDT) Received: from natasha.asdf456.com (natasha.aplushosting.com [64.235.37.2]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4J1vL2x025438 for ; Sun, 18 May 2003 18:57:22 -0700 Received: (qmail 27529 invoked by uid 0); 19 May 2003 01:54:55 -0000 Received: from unknown (HELO badlandsgames.com) (66.92.24.37) by natasha.asdf456.com with SMTP; 19 May 2003 01:54:55 -0000 Message-ID: <3EC839D5.10909@badlandsgames.com> Date: Sun, 18 May 2003 18:56:37 -0700 From: Joel Baxter Reply-To: jbaxter@neogeographica.com User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.3) Gecko/20030304 X-Accept-Language: en-us, en MIME-Version: 1.0 To: fam@oss.sgi.com Subject: Re: fam error "failed to accept new client" References: <3EC80F8E.900@badlandsgames.com> <3EC81C5E.70704@badlandsgames.com> <200305191036.17936.michael.wardle@adacel.com> In-Reply-To: <200305191036.17936.michael.wardle@adacel.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 131 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: jbaxter@badlandsgames.com Precedence: bulk X-list: fam Thanks for the response. Just to finish this off: it was indeed the xinetd version that was the culprit. I now have FAM running OK as built from the 2.6.7 source RPM. In fact using the binary RPM turned out to be a problem of a different sort. None of the RedHat binary RPMs for FAM include the static library libfam.a, and at least one of the many configure scripts in the Courier IMAP build actually checks for the existence of libfam.a at one point, as part of a does-FAM-exist sequence. (Even though, as far as I can tell, only the dynamic library is actually used by the compile end products.) So that wasn't good. Rather than hacking at the config scripts to make them get the correct answer, I thought I'd try going back to the source builds, so that libfam.a would exist, and all seems to be well. From michael.wardle@adacel.com Sun May 18 19:58:45 2003 Received: with ECARTIS (v1.0.0; list fam); Sun, 18 May 2003 19:58:58 -0700 (PDT) Received: from adacel.com (gunsmoke.adacel.com.au [210.11.130.7]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4J2wh2x026565 for ; Sun, 18 May 2003 19:58:44 -0700 Received: from nexus.adacel.com (Not Verified[10.32.240.1]) by adacel.com with MailMarshal (v5,0,3,91) id ; Mon, 19 May 2003 12:55:48 +1000 Received: (qmail 19857 invoked from network); 19 May 2003 02:58:15 -0000 Received: from unknown (HELO hera.wodonga.adacel.com.au) (192.168.75.251) by nexus.adacel.com with SMTP; 19 May 2003 02:58:15 -0000 Received: from selene.wodonga.adacel.com.au (selene.wodonga.adacel.com.au [192.168.75.55]) by hera.wodonga.adacel.com.au (8.11.6/8.11.6) with ESMTP id h4J2xRE04604; Mon, 19 May 2003 12:59:27 +1000 From: Michael Wardle Organization: Adacel Technologies To: jbaxter@neogeographica.com Subject: Re: fam error "failed to accept new client" Date: Mon, 19 May 2003 12:59:22 +1000 User-Agent: KMail/1.5 References: <3EC80F8E.900@badlandsgames.com> <200305191036.17936.michael.wardle@adacel.com> <3EC839D5.10909@badlandsgames.com> In-Reply-To: <3EC839D5.10909@badlandsgames.com> Cc: fam@oss.sgi.com MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200305191259.22553.michael.wardle@adacel.com> X-archive-position: 132 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: michael.wardle@adacel.com Precedence: bulk X-list: fam On Monday 19 May 2003 11:56, Joel Baxter wrote: > Thanks for the response. Not a problem. > I now have FAM running OK as built from the 2.6.7 source RPM. Good. > In fact using the binary RPM turned out to be a problem of a different > sort. None of the RedHat binary RPMs for FAM include the static library > libfam.a I suspect this file is in the "fam-devel" package. -- Michael Wardle Adacel Technologies From frank@hemer.org Thu May 22 03:12:01 2003 Received: with ECARTIS (v1.0.0; list fam); Thu, 22 May 2003 03:12:39 -0700 (PDT) Received: from mailout11.sul.t-online.com (mailout11.sul.t-online.com [194.25.134.85]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4MABx2x025133 for ; Thu, 22 May 2003 03:12:01 -0700 Received: from fwd06.sul.t-online.de by mailout11.sul.t-online.com with smtp id 19In3H-0005DK-02; Thu, 22 May 2003 12:11:59 +0200 Received: from server1.home.de (520058669942-0001@[217.86.3.196]) by fmrl06.sul.t-online.com with esmtp id 19In2n-2GEOLwC; Thu, 22 May 2003 12:11:29 +0200 Received: from master.home.de ([192.168.0.30]) by server1.home.de (8.11.6/8.11.6/SuSE Linux 0.5) with ESMTP id h4MACqH07579 for ; Thu, 22 May 2003 12:12:52 +0200 From: Frank Hemer Reply-To: frank@hemer.org To: fam@oss.sgi.com Subject: test Date: Thu, 22 May 2003 12:12:51 +0200 User-Agent: KMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200305221212.51671.frank@hemer.org> X-Sender: 520058669942-0001@t-dialin.net X-archive-position: 133 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: frank@hemer.org Precedence: bulk X-list: fam test From frank@hemer.org Thu May 22 10:37:13 2003 Received: with ECARTIS (v1.0.0; list fam); Thu, 22 May 2003 10:37:45 -0700 (PDT) Received: from mailout02.sul.t-online.com (mailout02.sul.t-online.com [194.25.134.17]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4MHbB2x003767 for ; Thu, 22 May 2003 10:37:12 -0700 Received: from fwd03.sul.t-online.de by mailout02.sul.t-online.com with smtp id 19Iu06-00066k-05; Thu, 22 May 2003 19:37:10 +0200 Received: from server1.home.de (520058669942-0001@[217.86.13.132]) by fmrl03.sul.t-online.com with esmtp id 19Itzy-1f9XUGC; Thu, 22 May 2003 19:37:02 +0200 Received: from master.home.de ([192.168.0.30]) by server1.home.de (8.11.6/8.11.6/SuSE Linux 0.5) with ESMTP id h4MHcZH08239 for ; Thu, 22 May 2003 19:38:36 +0200 Received: from [194.25.134.21] (helo=mailout10.sul.t-online.com) by mxng16.kundenserver.de with esmtp (Exim 3.35 #1) id 19IqMf-0004Ya-00 for frank@hemer.org; Thu, 22 May 2003 15:44:13 +0200 Received: from localhost (mail@) by mailout10.sul.t-online.com with local id 19IqMe-0008Bn-00; Thu, 22 May 2003 15:44:12 +0200 X-Failed-Recipients: fam@oss.sgi.com From: Frank Hemer To: fam@oss.sgi.com Subject: Mail delivery failed: returning message to sender Date: Thu, 22 May 2003 19:38:34 +0200 Reply-To: frank@hemer.org User-Agent: KMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200305221938.34932.frank@hemer.org> X-Sender: 520058669942-0001@t-dialin.net X-archive-position: 134 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: frank@hemer.org Precedence: bulk X-list: fam Hi, I just recently have added fam support to LinCVS (http://www.lincvs.org), a graphical frontend for cvs. Implementing this, I encountered problems with fam registering many (I tested with about 1200) dirs. The result in many cases is a crash. Still, for up to 200 dirs, fam works fine. This seems to be a problem with the polling mode, on my test system (SuSE 8.2) fam runs with fileschanged, obviously no IMon support. To fix this, I didn't want to patch my kernel, so I have written a kernel module (http://www.fwatch.org) that notifies about change-events, but without the need to recompile the kernel. I have added two files from the fam project, to be able to include this module into fam. I wonder whether anybody is interested to include fwatch-support into fam? The contrib files at the moment are just a hack since I was to lazy to change the autoconf setup. So it will only properly compile if IMon is not installed on the target system. Keep up the good work, Regards Frank From michael.wardle@adacel.com Thu May 22 17:56:16 2003 Received: with ECARTIS (v1.0.0; list fam); Thu, 22 May 2003 17:56:50 -0700 (PDT) Received: from adacel.com (gunsmoke.adacel.com.au [210.11.130.7]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4N0uE2x009750 for ; Thu, 22 May 2003 17:56:16 -0700 Received: from nexus.adacel.com (Not Verified[10.32.240.1]) by adacel.com with MailMarshal (v5,0,3,91) id ; Fri, 23 May 2003 10:52:51 +1000 Received: (qmail 20478 invoked from network); 23 May 2003 00:55:21 -0000 Received: from unknown (HELO hera.wodonga.adacel.com.au) (192.168.75.251) by nexus.adacel.com with SMTP; 23 May 2003 00:55:21 -0000 Received: from selene.wodonga.adacel.com.au (selene.wodonga.adacel.com.au [192.168.75.55]) by hera.wodonga.adacel.com.au (8.11.6/8.11.6) with ESMTP id h4N0vGE07737; Fri, 23 May 2003 10:57:16 +1000 From: Michael Wardle Organization: Adacel Technologies To: frank@hemer.org Subject: Re: FAM addon/patch Date: Fri, 23 May 2003 10:57:04 +1000 User-Agent: KMail/1.5 References: <200305221411.48248.frank@hemer.org> In-Reply-To: <200305221411.48248.frank@hemer.org> Cc: fam@oss.sgi.com MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200305231057.04958.michael.wardle@adacel.com> X-archive-position: 135 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: michael.wardle@adacel.com Precedence: bulk X-list: fam Hello Frank On Thursday 22 May 2003 22:11, Frank Hemer wrote: > I tried to post this to the mailing-list, but unfortunatelly, it didn't > appear. So I'll try direct ... As you might have noticed, your message has since appeared on the mailing list, albeit with the subject "Mail delivery failed: returning message to sender". > I'm core developer of a cvs frontend (http://www.lincvs.org), and have just > recently included fam support. Nice to hear. > When using fam, I realized the FamNextEvent routine throwing errors and > crashing when monitoring more than at about 200 directories. I've heard about such a problem before, but haven't made the time to investigate why it is occuring. It's definitely a problem I'd like to fix soon, so any input is most welcome. > It appeared to me, that fam (or fileschanged/dnotify???) was eating up a > lot of fd's. Yes, unfortunately this is the way that DNotify works, and FAM on most GNU/Linux distributions uses DNotify. > So finally I decided to write my own kernel notifier, and patch it into > fam. I built it as a kernel module (linux;-), that hooks into the > sys_calls, and writes events to /dev/fwatch. The required select routine to > read the events seemed to be quite close to the one imon uses, so I just > hacked the code into IMonNone.c++ and IMon.c++. When FAM was first open sourced, SGI made available a version of IMon for Linux (as you will see). After I became aware of the Linux DNotify API in Linux 2.4, I started recommending that people use Alex Larsson's DNotify patch on Linux so I wouldn't have to maintain an IMon kernel patch for each Linux kernel release and so users could get acceptable performance from FAM without patching their kernel. Since suggesting people use DNotify, it has become apparent that it has some limitations. :-( It's intriguing that you've provided another monitor. I'll have a look at it, and place it the contrib directory of our FTP site if it seems usable. Thanks for sending it. > With these changes, fam seems to be able to handle real huge projects > without problem;-). This suggests to me that the ~ 200 directory limit might be related to DNotify. I'll investigate that idea. > One more thing I reallized: > When removing a whole bunch of registered directories (I tried with about > 1200) and re-registering them right after, fam still reports changes to > these dirs. Sounds like a bug. I'll check it out when I find time. Thanks -- Michael Wardle Adacel Technologies From nwourms@myrealbox.com Fri May 23 08:23:35 2003 Received: with ECARTIS (v1.0.0; list fam); Fri, 23 May 2003 08:23:42 -0700 (PDT) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by oss.sgi.com (8.12.9/8.12.9) with SMTP id h4NFNY2x004859 for ; Fri, 23 May 2003 08:23:35 -0700 Received: from myrealbox.com nwourms@smtp-send.myrealbox.com [130.127.122.186] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.34 $ on Novell NetWare via secured & encrypted transport (TLS); Fri, 23 May 2003 08:23:34 -0700 Message-ID: <3ECE3C27.9060607@myrealbox.com> Date: Fri, 23 May 2003 11:20:07 -0400 From: Nicholas Wourms User-Agent: Mozilla/5.0 (Windows; U; Win 9x 4.90; en-US; rv:1.0.2) Gecko/20030208 Netscape/7.02 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Michael Wardle CC: fam@oss.sgi.com Subject: Re: FAM addon/patch References: <200305221411.48248.frank@hemer.org> <200305231057.04958.michael.wardle@adacel.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-archive-position: 136 X-ecartis-version: Ecartis v1.0.0 Sender: fam-bounce@oss.sgi.com Errors-to: fam-bounce@oss.sgi.com X-original-sender: nwourms@myrealbox.com Precedence: bulk X-list: fam Michael Wardle wrote: > Hello Frank [SNIP] >>When using fam, I realized the FamNextEvent routine throwing errors and >>crashing when monitoring more than at about 200 directories. > > > I've heard about such a problem before, but haven't made the time to > investigate why it is occuring. It's definitely a problem I'd like to fix > soon, so any input is most welcome. > > >>It appeared to me, that fam (or fileschanged/dnotify???) was eating up a >>lot of fd's. > > > Yes, unfortunately this is the way that DNotify works, and FAM on most > GNU/Linux distributions uses DNotify. Which is exactly why Imon should not only continue to be improved, but promoted as being superior to the crap which passes as Dnotify. >>So finally I decided to write my own kernel notifier, and patch it into >>fam. I built it as a kernel module (linux;-), that hooks into the >>sys_calls, and writes events to /dev/fwatch. The required select routine to >>read the events seemed to be quite close to the one imon uses, so I just >>hacked the code into IMonNone.c++ and IMon.c++. > > > When FAM was first open sourced, SGI made available a version of IMon for > Linux (as you will see). After I became aware of the Linux DNotify API in > Linux 2.4, I started recommending that people use Alex Larsson's DNotify > patch on Linux so I wouldn't have to maintain an IMon kernel patch for each > Linux kernel release and so users could get acceptable performance from FAM > without patching their kernel. Since suggesting people use DNotify, it has > become apparent that it has some limitations. :-( > > It's intriguing that you've provided another monitor. I'll have a look at it, > and place it the contrib directory of our FTP site if it seems usable. > Thanks for sending it. With all the negative feedback surrounding DNotify, perhaps it is time for the kernel people to face the reality that it is an inferior interface and an alternative choice should be made available to the Linux community. Cheers, Nicholas