From owner-freebsd-ports@FreeBSD.ORG Mon May 17 01:30:32 2010 Return-Path: Delivered-To: freebsd-ports@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 70A3F106566C for ; Mon, 17 May 2010 01:30:32 +0000 (UTC) (envelope-from Hans.F.Nordhaug@hiMolde.no) Received: from malle.himolde.no (malle.hiMolde.no [158.38.68.22]) by mx1.freebsd.org (Postfix) with ESMTP id EC3BA8FC1A for ; Mon, 17 May 2010 01:30:31 +0000 (UTC) Received: from harr.himolde.no (harr.hiMolde.no [158.38.68.20]) by malle.himolde.no (8.13.8/8.13.8) with ESMTP id o4H1UTNq003915 for ; Mon, 17 May 2010 03:30:29 +0200 Received: from harr.himolde.no (harr.himolde.no [127.0.0.1]) by harr.himolde.no (8.13.1/8.13.1) with ESMTP id o4H1UT4B015297 for ; Mon, 17 May 2010 03:30:29 +0200 Received: (from nordhaug@localhost) by harr.himolde.no (8.13.1/8.13.1/Submit) id o4H1US0m015296 for freebsd-ports@freebsd.org; Mon, 17 May 2010 03:30:28 +0200 Date: Mon, 17 May 2010 03:30:28 +0200 From: "Hans F. Nordhaug" To: freebsd-ports@freebsd.org Message-ID: <20100517013028.GA14513@hiMolde.no> References: <20100509162335.GA934@hiMolde.no> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20100509162335.GA934@hiMolde.no> User-Agent: Mutt/1.4.1i Subject: Re: Apache 22 - FreeBSD 8.0 - (httpd), uid 80: exited on signal 11 X-BeenThere: freebsd-ports@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting software to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 17 May 2010 01:30:32 -0000 * Hans F. Nordhaug [2010-05-09]: > * Helmut Schneider [2010-05-03]: > > Hans F. Nordhaug wrote: > > > > > * Helmut Schneider [2010-05-02]: > > > > Hans F. Nordhaug wrote: > > > > > > > > > I recently upgrade to FreeBSD 8.0 (from 7.2) and suddenly I get a > > > > > lot of "(httpd), uid 80: exited on signal 11" in my logs. I have > > > > > similar problems with amavisd - see > > > > > http://lists.freebsd.org/pipermail/freebsd-questions/2010-April/thread.html#215757 > > > > > I'm have updated and recompiled all ports. The logs > > > > > /var/log/messages and the httpd error log both just report > > > > > "exited on signal 11" or "Segmentation fault (11)" > > > > > > > > > > Any hints? > > > > > > > > Find the .core file, start gdb with the core file, type "bt", post > > > > the output. > > > > > > Sorry, I should have mentioned that I can't find any core files. > > > "find / -name '*.core'" returns nothing. > > > > I guess one can assume that > > > > - you already checked RAM with memtest or so. > > - you already check your hard drive (you mentiened you fsck'ed your > > hard drive and found errors that wer corrected) > > - you are running one or more perl apps within apache? > > Thx to you and everybody else who have answered so far: > > - I have (now) checked RAM with memtest for several days. No issues. > - Yes, I have already checked my hard drive using fsck, and the few > problems that were reported was fixed a week ago. > - My websites are using PHP, not Perl. I have checked/fixed the > extension loading order using fixphpextorder.sh as posted on > http://www.pingle.org/2007/09/22/php-crashes-extensions-workaround > It didn't help. I don't have that many extensions - the order is: [cut] OK, this problem is just not going away. Doing a reload of a page, the segmentation fault will appear from 20-50% of the time. It's getting more and more clear that this is a PHP bug in some way. I have compiled PHP and Apache with debugging, disabled the Suhosin patch and only suing two extensions: php5-xml and php5-session. All ports are up-to-date. This is still on FreeBSD 8 with PHP 5.3.2. The backtrace is: gdb /usr/local/sbin/httpd (gdb) run -X Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 28501140 (LWP 100131)] 0x28ad2030 in zend_mm_check_ptr (heap=0x28e4f900, ptr=0x29994d0c, silent=1, __zend_filename=0x28c53d6b "Zend/zend_language_scanner.l", __zend_lineno=685, __zend_orig_filename=0x0, __zend_orig_lineno=0) at /usr/ports/lang/php5/work/php-5.3.2/Zend/zend_alloc.c:1347 1347 if (p->info._size != ZEND_MM_NEXT_BLOCK(p)->info._prev) { (gdb) bt #0 0x28ad2030 in zend_mm_check_ptr (heap=0x28e4f900, ptr=0x29994d0c, silent=1, __zend_filename=0x28c53d6b "Zend/zend_language_scanner.l", __zend_lineno=685, __zend_orig_filename=0x0, __zend_orig_lineno=0) at /usr/ports/lang/php5/work/php-5.3.2/Zend/zend_alloc.c:1347 #1 0x28ad366c in _zend_mm_free_int (heap=0x28e4f900, p=0x29994d0c, __zend_filename=0x28c53d6b "Zend/zend_language_scanner.l", __zend_lineno=685, __zend_orig_filename=0x0, __zend_orig_lineno=0) at /usr/ports/lang/php5/work/php-5.3.2/Zend/zend_alloc.c:1983 #2 0x28ad465c in _efree (ptr=0x29994d0c, __zend_filename=0x28c53d6b "Zend/zend_language_scanner.l", __zend_lineno=685, __zend_orig_filename=0x0, __zend_orig_lineno=0) at /usr/ports/lang/php5/work/php-5.3.2/Zend/zend_alloc.c:2351 #3 0x28abb74c in zend_multibyte_read_script ( buf=0x284ff000 ", "\n//\n// PIVOTX - LICENSE:\n//\n// This file is part of PivotX. PivotX and all its parts are licensed under\n// the GPL "..., n=658) at zend_language_scanner.l:685 #4 0x28aba690 in open_file_for_scanning (file_handle=0xbfbfe934) at zend_language_scanner.l:263 #5 0x28aba8a8 in compile_file (file_handle=0xbfbfe934, type=8) at zend_language_scanner.l:331 #6 0x28af6fe7 in zend_execute_scripts (type=8, retval=0x0, file_count=3) at /usr/ports/lang/php5/work/php-5.3.2/Zend/zend.c:1186 #7 0x28a8a9fd in php_execute_script (primary_file=0xbfbfe934) at /usr/ports/lang/php5/work/php-5.3.2/main/main.c:2260 #8 0x28bc3173 in php_handler (r=0x290db058) at /usr/ports/lang/php5/work/php-5.3.2/sapi/apache2handler/sapi_apache2.c:655 And so on ... I also have the full backtrace for those interested. Regards, Hans PS! Should I file PR?