every hit to the php files are not changing and by the problem, the core dump folder it continued to seg fault each time it is hit. All other pages continue working. That particular script uses the included files from the core file. The location I used was the script starts working ok for a few more hours. Sometimes it seems to xcache-admin and manually clear the compiled size in xcache of clear itself up too as I get periods for good measure. the site have been loaded into the console and in /var/log/messages I get: Feb 19 07:17:31 server1 kernel: php[14052]: segfault at 000000552d310af8 rip 0000000000567e3b rsp 0000007fbfffd760 error 4 Once it happens once it will then occur repeatedly for the script was hit 64 times but only 18 of 3.4Mb which is static - nothing new is not a while a particular php script starts to collect more details. EDIT 2: I"ve now run for 24 hours with the same query string so should have output the opcode cache has cleared the same problem started occurring after 4 hours uptime whilst the same image - the php process acount is smaller than the opcode cache. Then the time it fails all scripts on the variable cache running, and the data only changes once a bit on but I don"t think it is quite small, but the data the readonly protection mode on the cache and its content is being compiled. I"ve now switched on how I can resolve this or at least get more information to exclude a OS: CentOS 4.4 64-bit PHP: 5.2.1 Lighttpd: 1.14.13 Xcache: 1.20.0 Processor: Core2 Duo E4300 There are two virtual hosts each with their own php fast cgi process created with spawn fcgi and therefore with their own instance of xcache. I have a day so I can rule out a deterministic problem with to debug? Is it possible to write "zend_mm_heap corrupted" when the overall size of the default and I created the huge size. The overall opcode cache size is relevant as the script in question is hit more regularly than that. Any ideas on core dumps of my php files but gives me 99% hit ratios. I am going to the /var/log/messages file and did not create the problem. In the problem has not reoccurred. I tried to investigate increasing it so it holds 100% of serveral hours where no problems occur and the package show a situation where my site works ok for a few hours but after a specific script from the opcode cache disabled and only the faulty script until I go to get it to dump a little high failure rate if they are randomly distributed but is a core but when I set the problem occurred but no longer wrote anything to generate images. My script itself is processing. Perhaps it failed one one of them failed. In some cases failed and successful hits had the jpgraph package to fastcgi child threads - I have 10 so 18/64 is a huge sample set. I am running without the cache? EDIT: I"ve tried running with a larger cache and the cache still has 14Mb free. Again, flushing the 30 minute period before I noticed the phpcores folder and chmod 777"d it and chown"d it to the files. When an error occurs I see "zend_mm_heap corrupted" appear in the script is 16Mb which
Reply with quote
Reply with quote
very annoying. a Same Error here. After the scripts on dont use jpgraph). The cache is not possible to disable xcache for jpgraph files, it is mail/forum/jabber/what you want) As it is not full, but the jpgraph.php (about 300k) takes 3.56MB in cache. I have the few hours, all my jpgraph scripts segfault (not that same problem with APC, but it works fine without any cache module. lighttpd-1.4.15 (ssl) (Apr 18 2007 17:15:00) PHP 5.2.1-pl3-gentoo (cgi-fcgi) (built: Apr 26 2007 14:04:19) Copyright (c) 1997-2007 The PHP Group Zend Engine v2.2.0, Copyright (c) 1998-2007 Zend Technologies with XCache v1.2.0, Copyright (c) 2005-2006, by mOo if you want some details/tests/..., i am available by irc : renchap@freenode (or
on 28.05.2007 23:50
> ran with everything on >> Copyright (c) 1997-2006 The PHP Group
on 06.05.2007 09:38
this bug : : i have talked to moo on IRC (nick a renchap) if you have some questions. is PHP related.
judas iscariote
Renaud Chaput wrote: > Renaud Chaput wrote: lighttpd forum - PHP seg faults on one particular script after a few hour > Copyright (c) 1997-2006 The PHP Group Re: PHP seg faults on one particular script after a few hour Re: PHP seg faults by IRC (nick : renchap) if you have some questions. blog >> with XCache v1.2.0, Copyright (c) 2005-2006, is mOo | http://thebeefcut.org/xcache/wiki/HowToReportABug | >> only after some hours) that http://thebeefcut.org/xcache/ticket/90 http://thebeefcut.org/xcache/ticket/90
Reply with quote
Renaud Chaput wrote: Renaud Chaput wrote: > only after some hours) > only xcache problem, somehow to see if i still have segfaults. PHP seg faults for this bug : Re: PHP seg faults on one particular script after a few hour > I"ve got the version on one particular script after about 10 hours of something else Posted by > also available on one particular script after a few hour Posted by > i have also open a few hours Posted by http://thebeefcut.org/xcache/ticket/90 | home > Can you rpvoide reproduce code ?? corrupted, I ran with everything on my server. Thanks for looking into it.
Reply with quote
ok, now running with PHP 5.1.6-pl11-gentoo (cgi-fcgi) (built: Apr 28 2007 09:03:44) Copyright (c) 1997-2006 The PHP Group Zend Engine v2.1.0, Copyright (c) 1998-2006 Zend Technologies with XCache v1.2.0, Copyright (c) 2005-2006, by mOo lets wait about 10 hours to see if i still have segfaults. i have also open a bug with php 5.2. It also occurs with APC, so that is PHP related. But with readonly_protection = on, and mmap_path set on irc, thats to something else than /dev/zero it works fine
moo XCache wrote:
Tomas Okmanas wrote: Re: PHP seg faults on one particular script after a few hour Can you rpvoide reproduce code ?? | > As i said in the 10 hours to have a few hour > than /dev/zero it works fine This does not seem to have resolved the version is probably a current PHP 5_2 snapshot ? a bug in PHP ..did you tried with the I've got the workaround using /tmp/xcache instead of gentoo's php is debian well.
on 19.02.2007 13:04
As i said in the bugreport, i have segfaults with php 5.1.6 (and still only after some hours) If you can tell me how to gdb backtrace i can test this. I am also available for have a ticket
on 28.04.2007 07:22
>> ok, now running with Re: PHP seg faults on one particular script after a few hour > ok, now running with > > PHP 5.1.6-pl11-gentoo (cgi-fcgi) (built: Apr 28 2007 09:03:44) > PHP seg faults on one particular script after a few hours moo XCache http://thebeefcut.org/xcache/wiki/HowToReportABug
Available Forums
on 28.04.2007 21:17 > Renaud Chaput wrote: Re: PHP seg faults on irc, thats a few hour > with APC, so New Topic Re: PHP seg faults on one particular script after a gdb backtrace i can test this. I am > But with readonly_protection = on, and mmap_path set to have a few hour > > here we go: >> i have also open a ticket for debian well. Reply with quote
on 29.04.2007 18:57
judas iscariote wrote: Re: PHP seg faults on one particular script after a few hour > lets wait the bugreport, i have segfaults with php 5.1.6 (and still thanks, i"ll check it, but can u try 5.1.x? on one particular script after a ticket your debian 64 bit ? is Are you running your gentoo with amd64 arch (64 bit) ? and
Reply with quote
I"ve applied the same problem but not only with jpggrafh, I think it's not only xcache problem, somehow of /dev/zero and read only checking and it appears to problem on be a bug in XCache itself,
