Bug #12634 Updated: Segfaults due to possible memory leak??

ID: 12634 Updated by: yohgaki Reported By: katz@advanced.org Status: Open Bug Type: Reproducible crash Operating System: Linux 2.4 PHP Version: 4.0.6 New Comment: Do you still have problems with 4.1.0? Is it ok to be closed? Previous Comments: ------------------------------------------------------------------------ [2001-10-03 14:21:36] katz@advanced.org As a follow-up on this bug report .. Just today I installed a recent cvs version (called "4.0.7RC2" in the Debian unstable distribution)... So far the system has been running almost an entire day without a problem .. I noticed in the change log that there were some LDAP memory leak fixes ... that may have been the problem as our site also relies heavily upon ldap .. So far .. So good. ------------------------------------------------------------------------ [2001-08-19 21:40:28] katz@advanced.org Also .. keep in mind my stuff uses OCI-8 (with oracle 8.1.6), so that may be linked (since no one else has reported this problem).... ?? ------------------------------------------------------------------------ [2001-08-19 21:36:48] katz@advanced.org No, it appears to be almost random.... The script names (and line numbers) that appear in the logs are always different. I was able to force the problem for any script by using the apache benchmark utility (ab) and blasting a script with generated hits .. when doing this the seg faults start almost immediately (and again for any php script i point it at).. if this helps.... ------------------------------------------------------------------------ [2001-08-19 05:05:03] sniper@php.net Does this happen with specific script? ------------------------------------------------------------------------ [2001-08-07 17:13:46] katz@advanced.org Under 4.0.6, my apache processes start segfaulting with the following problem after running (stable) for a period of time: [Tue Aug 7 16:53:30 2001] [notice] child pid 7139 exit signal Segmentation fault (11) Once this starts happening, it continues until i kill and restart the server. It will then run fine for some time, then happen again. The period of time that it runs stable decreases as hits increase .. Running with --enable-debug I get the following in the apache error log (I had issues generating a gdb backtrace, but if needed I can attempt again): /usr/www/thinkquest/common/nav_menu.inc(330) : Fatal error - Allowed memory size of 12582912 bytes exhausted at ../../Zend/zend_opcode.c:48 (tried to allocate 184320 bytes) /usr/www/thinkquest/common/scripts.inc(3062) : Fatal error - Allowed memory size of 12582912 bytes exhausted at ../../Zend/zend_opcode.c:48 (tried to allocate 11520 bytes) [Tue Aug 7 16:53:23 2001] Script: '/usr/www/thinkquest/index.html' --------------------------------------- ../../Zend/zend_opcode.c(172) : Block 0x08138E90 status: Beginning: Overrun (magic=0x401E8080, expected=0x7312F8DC) End: Unknown --------------------------------------- /usr/www/thinkquest/common/nav_menu.inc(330) : Fatal error - Allowed memory size of 12582912 bytes exhausted at ../../Zend/zend_opcode.c:48 (tried to allocate 184320 bytes) /usr/www/thinkquest/common/scripts.inc(1870) : Fatal error - Allowed memory size of 12582912 bytes exhausted at ../../Zend/zend_opcode.c:48 (tried to allocate 184320 bytes) ... and similar, over and over again.. however the script that it occurs in, and the line, are always different.. I tried increasing the amount of memory allowed per script, but it just uses all that up too.. As a note... everything runs fine under 4.0.5 .... I have cvs updates up to July 20.. The configuration parameters I used are: "../configure" \ "--prefix=/usr" \ "--with-apxs=/usr/bin/apxs" \ "--with-regex=system" \ "--with-config-file-path=/etc/php4/apache" \ "--disable-rpath" \ "--enable-debug" \ "--enable-memory-limit" \ "--enable-calendar" \ "--enable-sysvsem" \ "--enable-sysvshm" \ "--enable-track-vars" \ "--enable-trans-sid" \ "--enable-bcmath" \ "--with-bz2" \ "--enable-ctype" \ "--with-db2" \ "--with-ndbm" \ "--with-ldap=/usr" \ "--with-oci8=/usr/oracle/product/8.1.6" \ "--enable-sigchild" \ "--enable-exif" \ "--enable-filepro" \ "--enable-ftp" \ "--with-gettext" \ "--enable-mbstring" \ "--enable-mbstr-enc-trans" \ "--with-pcre-regex=/usr" \ "--enable-shmop" \ "--enable-sockets" \ "--enable-wddx" \ "--with-xml=/usr" \ "--with-expat-dir=/usr" \ "--enable-yp" \ "--with-zlib" \ "--without-pgsql" \ "--disable-static" \ "--with-layout=GNU" \ "--with-curl=shared,/usr" \ "--with-dom=shared,/usr" \ "--with-gd=shared,/usr" \ "--with-jpeg-dir=shared,/usr" \ "--with-xpm-dir=shared,/usr/X11R6" \ "--with-png-dir=shared,/usr" \ "--with-freetype-dir=shared,/usr" \ "--with-imap=shared,/usr" \ "--with-mcal=shared,/usr" \ "--with-mhash=shared,/usr" \ "--with-mysql=shared,/usr" \ "--with-sablot=shared,/usr" \ "--with-snmp=shared" \ "--with-sybase-ct=shared,/usr" \ "--with-ttf=shared,/usr" \ "--with-t1lib=shared,/usr" The extension order i use is: gd.so imap.so mhash.so domxml.so mysql.so curl.so ( I tried removing domxml, mhash and curl and the problems were still there ) ... Also to throw in, I have OCI8 configured with Oracle 8.1.6 (as seen above) .. linked with pthreads and using Apache 1.3.20... Any more info if needed can be provided. ------------------------------------------------------------------------ Edit this bug report at http://bugs.php.net/?id=12634&edit=1