新聞財經娛樂♬資訊 News Economy

good trick using n++ search, although none of the php file contains 500 other than part of something else.

I'll have to go through turning off each include to find it indeed....
Will keep u posted, but off to bed now.

Download the whole site locally, if you don't already have it download Notepad++
Load all files into Notepad and do a global search for 500 - you may get lucky and find a page that actually has it as an header response - but I doubt it.

Reality is you'll have to probably go through disabling stuff til you find it.

i renamed it and touched an empty index.php
that resulted in:
HTTP request sent, awaiting response... 200 OK
Length: 0 [text/html]
Saving to: “index.html.8”


so, definitely caused by something in the script!

Clear everything out of index.php so it is just blank and see if you still get a 500
Obviously do it quick so you can get the site back up.

i commented the initialization files (includes) at the start of the index.php which resulted in an empty website, but still with error 500 via wget.

I recon you'd advise me to try turning off each include I find more?

Disable all your plugins and see if the error is still there

all are 3 different vhosts with a own copy of the php files for each domain, and also their own mysql db for each domain.

Are all the sites on the same server? Running from the same same installation of OSCommerce?

Check your vhosts file.

well so far it's been 2 IT professionals that went though it all (site owner and myself), but we're not php experts... my buddy keeps insisting it must be a language setting because its only happening for the french and dutch websites, and not on the main one.

and we both tried various in the php files, but always failed or site stopped working

You really need someone who can go through your files - we can only guess so much.

no worries, i just wanted to pass the website(s) details via mail.
I have commented out following in .htaccess file of the .be site

#php_flag register_globals on
#php_flag session.use_trans_sid off
 
site still works fine, although still error 500
also after a restart of apache

@warddhooghe,

This is already set as a private question so it will not get indexed, If you are concerned about posting the link, it can be deleted after the problem is solved and the question is closed.

COBOLdinosaur, Topic Advisor

Offline email communication is against the rules

drop me a line on ward.dhooghe ( at ) gmail ( dot ) com

Can you give a link - promise I won't tell anyone else, use a url shortener if you need

the website has never failed using a browser. it's a online shop and we get orders nearly every hour of the day.
For the xmas period we really like to use google adwords,
Just that error 500 keeps showing when using wget or fiddlering the browser.

can't continue much getting into the 'constants' errors, needs some sleep too.
will keep you posted on progress.

To just backtrack does the site work normally if you access it?

an old version of oscommerce which was custom tweaked for us many years ago.

got the timezone errors fixed, thanks for that suggestion.

Can these constant errors still cause the error 500?

Well at least we know the site is working (of a sorts)

You can fix the timezone errors by adding to your page
date_default_timezone_set('Europe/Berlin');
I'm assuming Berlin as PHP thinks that's where you are

For the Constant FILENAME_DEFAULT already defined errors you must be trying to define the values when they are already defined in your vhosts file


Are you using OSCommerce or something?

drop me a line on ward.dhooghe ( at ) gmail ( dot ) com for the site details.

My index.php file now starts with:
<?php
error_reporting(E_ALL);

However, no errors are shown in IE.
Now that I got php_error.log turned, it does show quiet some errors that might have something to do with it:

PHP Notice:  Constant FILENAME_DEFAULT already defined in /var/www/vhosts/
PHP Notice:  Constant FILENAME_CATEGORIES already defined in
PHP Notice:  Constant FILENAME_CMS_CATEGORIES already defined in
PHP Notice:  Constant FILENAME_DEFAULT already defined in
PHP Notice:  Constant FILENAME_CATEGORIES already defined in
PHP Notice:  Constant FILENAME_CMS_CATEGORIES already defined in

PHP Warning:  date() [<a href='function.date'>function.date</a>]: It is not safe to rely on the system's timezone settings. You are *required* to use the date.timezone setting or the
 date_default_timezone_set() function. In case you used any of those methods and you are still getting this warning, you most likely misspelled the timezone identifier. We selected 'Europe/Berlin' for 'CET
/1.0/no DST' instead in /httpdocs/includes/modules/page_modules/product_info.php on line 297

PHP Warning:  mktime() [<a href='function.mktime'>function.mktime</a>]: It is not safe to rely on the system's timezone settings. You are *required* to use the date.timezone setting
or the date_default_timezone_set() function. In case you used any of those methods and you are still getting this warning, you most likely misspelled the timezone identifier. We selected 'Europe/Berlin' for 'CET/1.0/no DST' instead in /httpdocs/includes/functions/general.php on line 885

 PHP Warning:  strftime() [<a href='function.strftime'>function.strftime</a>]: It is not safe to rely on the system's timezone settings. You are *required* to use the date.timezone se
tting or the date_default_timezone_set() function. In case you used any of those methods and you are still getting this warning, you most likely misspelled the timezone identifier. We selected 'Europe/Berl
in' for 'CET/1.0/no DST' instead in /httpdocs/includes/functions/general.php on line 885


All of these repeat quiet a lot.

Did you do everything else I said? The phpinfo.php was just to make sure your site is configured and working properly.
Can you post a link to the site - would be so much easier.

phpinfo sure shows a lot of info.
core directive error_log has no value, guess that explains why i couldnt find anything in logs.

hope we're onto something here, unfortunately i dont have much time today, will continue later and keep you posted.

Yes thanks GaryC123, reminder or trigger that
let me have other thinking

If the index page is php script which will
cause fatal error by false php scripting not reported by browser   and it might  cause
500 error. If error_log file is no issue, Gary's suggestion, it is
good to try

warddhooghe,I'll quit this thread since I have no more further idea
if error_log is no issue

Duncan

Throw a phpinfo.php page on the server and access it through the browser or wget.
If successful it probably means you have a programming/logic error in your php pages.
Make sure you have error reporting turned on in your pages - add this to the top of your php
page.
error_reporting(E_ALL);

In your browser (IE) click Tools > Internet Options
Click the Advanced tab
Scroll down and under Browsing make sure Show friendly HTTP error messages is not checked

For Firefox
Type in the addressbar about:config
In the search bar type error_pages.enabled and press enter.
Now set the value of browser.xul.error_pages.enabled to true. Double clicking will do this

what is  your error_log file say about when you wget www.yoursite.com ?
Please display the file
by tac  /usr/loca/apache/logs >junk.txt and copy us the
top of 20 lines in junk.txt, if the file is too big , just ctrl-c  during command executing
Probably the logs will give you some hint if you read it  carefully

the file is at  /usr/local/apache/logs if you have root access,
I do  not remember where is log file if you are using shared-server

Duncan

removed it, restarted httpd, and still the same problem

so I feel it is a little bit  strange that  I have NO 500 error but you have
now you explained that is misunderstanding only.

Now it seems and comes to final, probably it should be caused
from   Options +FollowSymLinks, in .htaccess file


Please delete it and try

Duncan

test.com isn't mine, it was an example.
Why did I wrote this:
1) Please drop me a line on ward.dhooghe ( at ) gmail ( dot ) com for the details.

Dear Ward,

I thought your worksite is example website for asking question purpose only (not real
site such as www.testing-yourworksite.com or www.testing-example.com) ,
Actually it is your real website, www.test.com

I try www.test.com & www.test.com/test in my browser it works fine
and no 500 error ?  Please see test.gif for your site image, www.test.com

And I wget www.test.com and www.test.com/test, it also work fine and its index.html page
loaded into my linux server.

FollowSymLinks means if a dir is a symbol link, follow the link
Please try it as  follows:


And the question I still confuse
==========================
-1 Just make sure when you wget www.test.com on your server , it reported 500 error, Right ?
-2- Your server must be shared-server, Right ?


Hope understand your question and it will be final soon to solve your issue
Duncan
-

Dear Duncan,

1) Please drop me a line on ward.dhooghe ( at ) gmail ( dot ) com for the details.

2) going directly to http://www.test.com/index.php also gives error 500 with wget, and also 500 visible in fiddler when using browser.

3) .htaccess file present in the root of website:

RewriteEngine on
RewriteBase /
Options +FollowSymLinks
 
php_flag register_globals on
php_flag session.use_trans_sid off

RewriteRule includes/.*.php - [F]
 
RewriteCond $1 !/$
RewriteRule ^([^.]+)$ /$1/ [R=301,L]

RewriteCond %{HTTP_HOST}   !^www
RewriteRule (.*)   http://www.%{HTTP_HOST}/$1   [L,R=301]

RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{QUERY_STRING} =""
RewriteCond %{REQUEST_URI} !^admin
RewriteRule (.*) index.php

# Error Document
ErrorDocument 404 /index.php?module=error404

Dear Ward,

The 500 Internal Server Error is a very general HTTP status code that means something has gone wrong on the web site's server but the server could not be more specific on what the exact problem is.

Most of the time, "gone wrong" means an issue with the page or site's programming, nothing
clients have anything to do with. the sever owner needs to take a look into all programming script related  to that  site

Just suggestion or cross-check only
==============================
1-I am also using fiddler and Charles http proxy debugger and Linux 64-bit server.
Do you think we could try to do wget yourwebsite at our side to see what happen
if it doesn't require password? Probably we will get the same 500 error. But at least
try and check it at our side) what is exact website path ?

2-could you read the index page(index.php/index.html) at the problem site at
www.test.com && www.test.com/test/ in oder to see what script error problem
(such as there is php script  fatal error but not reported by browsers so it might create 500 error) ?
 And also need to check php error log file on the site .



3-if possible, could we see the website's .htaccess file ?

Duncan

Hi Duncan,

To start from the beginning, all functionalities on the websites are working as they should.
I wanted to use Google Adwords (which works fine for other domains that we are using) for the website with the issue here, and it got declined with reason something in the terms of "Technical problems with website, error 500". I followed a help link from Google where they suggest to use Fiddler which is a tool that monitors what your browser is doing.

So when I am visiting the site in question, fiddler is showing on the very first line, this error 500 for GET / HTTP/1.1
the next lines are fine with result 200 for GET /templates/stylesheet.css HTTP/1.1
and it continues with more of result 200 for images, js files, etc

wget which is a tool similar to what google and general webcrawlers are using to access websites, hence me trying that. And its also gives the error 500

the domain root (eg www.test.com) and subfolders (e.g. www.test.com/test/) both gives the error 500 when using wget and it is also visible in fiddler when using a browser.

however a direct file (e.g. www.test.com/somefile.php) is working like normal (status 200).

To answer your questions:
im not using an alias. and I've used wget on several machines, diffirent linux distributions, all the same results.
i dont know curl, but i'm betting it will give the same results. I'll look into it later case it gives more information than wget.

Any other suggestions?
Best regards,
Ward

Dear warddhooghe,

wget http://www.xxxxxxx.xx


From your question,  http://www.xxxxxxx.xx, is it your Google Adword acount website ?
if not , why you said impossible to use google adwords in this situation as it detects a failing website. ? Hope you can clarify this to make sure we won't misunderstand your question

Based on your information, Fidder is working for the site without 500 error
and only happen with wget on your server. Right ?

Did you try curl on your server  to see  it can echo out the output file and see what it looks like ?
Did you use alias for  wget command before or alias wget in .bashrc file  if so unalias it first, and then using wget with your target option correctly ?
There is one article, just scan it only if have time, otherwise ignore it . That
I search "Wget Google Adwords 500 HTTP error" in google.
https://developers.google.com/gdata/articles/using_cURL
and look at the bottom of the link

Hope it will help
Duncan

The browser is also getting the error 500 but not showing it to the user, it shows the website as if nothing is wrong with it.

So your browser also gets the 500 error?

I thought this worked in your browser but failed when using wget?

sorry for delay, our websites dont use https, for payments https is used but all in domains from the financial service supplier

httpfox shows similar to fiddler, this is the info of the line with the error 500.
request header:
(Request-Line)      GET / HTTP/1.1
Host      www.sitewiththeissue.xx
User-Agent      Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0) Gecko/20100101 Firefox/12.0
Accept      text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language      en-us,en;q=0.5
Accept-Encoding      gzip, deflate
DNT      1
Connection      keep-alive
Cookie      osCsid=pup79njl243a1io35lgremrgk1

response header:
(Status-Line)      HTTP/1.0 500 Internal Server Error
Date      Mon, 18 Nov 2013 23:30:16 GMT
Server      Apache
Expires      Thu, 19 Nov 1981 08:52:00 GMT
Cache-Control      no-store, no-cache, must-revalidate, post-check=0, pre-check=0
Pragma      no-cache
X-Powered-By      PleskLin
Connection      close
Content-Type      text/html

cookiestab:
osCsid      pup79njl243a1io35lgremrgk1      /      .www.sitewiththeissue.xx      End Of Session

querystring and postdata have no parameters

Does the URL you are accessing contain cgi GET parameters encoded in a query string?
Does the URL you are accessing expect http POST parameters?
Does it expect to receive cookies?
Have you tried the firefox extension HttpFox I mentioned to inspect the http session?

I'm unclear on what this means. Does this mean you are not using HTTPS/SSL but fiddler still reports

You say you have other sites which do not have this error 500 and they use HTTPS/SSL?
Do they use the same apache server instance and therefore the same config file or do they have their own config files?

Can you inspect the apache config file of a working HTTPS/SSL server instance and compare it with the broken HTTPS/SSL apache server config file, paying particular attention to the mod_ssl config options
http://httpd.apache.org/docs/2.2/mod/mod_ssl.html
Specifically, looking at the cipher suite options such as SSLCipherSuite

its happening in http  
with the -no-check-cert option wget also give the same

whole website is using php and used a mysql db, everything works fine on the site.
it's running on linux and using plesk.
i've been looking for hours through everything that has .log extension.
havent checked for core dumps, but that would imply disk filling up, and it doesnt. we have about 3k+ visitors a day, disk usage is 'normal'. other sites that we run (also using same php/mysql), dont have this error 500, so baffles me a lot where this is coming from.

> ClientHello

So does this mean you are using https/ssl ?
Are your certs valid? Are they signed?
Have you tried wget with the --no-check-certificate option?

Are you accessing static html, or are you accessing some kind of CGI script like perl or php?

Would those scripts log into some location other than your web server's access/error.log?

Could they be dumping core? Are you on a *nix OS? Can you truss or strace the httpd process which handles the request?

using latest fiddler log shows something interesting:
00:46:44:5508 HTTPSLint> Warning: ClientHello record was 355 bytes long. Some servers have problems with ClientHello's greater than 255 bytes. https://github.com/ssllabs/research/wiki/Long-Handshake-Intolerance
00:46:44:6978 HTTPSLint> Warning: ClientHello record was 353 bytes long. Some servers have problems with ClientHello's greater than 255 bytes. https://github.com/ssllabs/research/wiki/Long-Handshake-Intolerance
00:46:45:3598 HTTPSLint> Warning: ClientHello record was 353 bytes long. Some servers have problems with ClientHello's greater than 255 bytes. https://github.com/ssllabs/research/wiki/Long-Handshake-Intolerance

the logs show the error 500 but no details to why.
no proxies involved
i run wget from 3 other locations with direct internet connection (no NAT) and it's all the same.

this came to light when i tried to use google adwords, it failed due to error 500. Google FAQ suggested to use fiddler to find out the problem.

it's very strange since the website is accessible with the browser (tested Firefox, Chrome, IE)

Is the website accessible with your browser?
Are you using an http proxy to access the site?
Is your wget command configured to use that http proxy?

Oh so you do have access to the server logs?

What does the access.log and error.log report around the time of you receiving error 500 from the wget client? Note if you are using ssl you might need to look in ssl_access.log and ssl_error.log.

the apache logs of the site and also the general apache logs dont show anything related to this

using the spoofed example from you, I still get the same Error 500

There are probably a gazillion possible reasons for this but whenever I've struck the same issue, whether it be using wget or some other downloader, it's usually because there is code on the site which is designed to respond in certain ways to "normal" web browsers. Along comes wget, or similar, and the code breaks because it doesn't have code for it, nor a proper fallback. Such site usually break for very old browsers as well. In such cases you have to get the downloader to mimic a common browser to get it to work correctly.


In fiddler, did you see the last  re-direct website address.

If so, why use not  wget "http:redirect.site"

Duncan