[Apache@Win] After Submitting a search: "Premature end of script headers"

Ray · Dec-21-2011, 03:30 AM

sazizrahman - I answered the same question you also posted here:
http://www.wrensoft.com/forum/showthread.php?t=4816

sazizrahman · Dec-20-2011, 05:06 PM

Originally posted by wrensoft View Post

You normally get this error is the server is configured to run .CGI scripts as Perl. Zoom is not a Perl script, thus you get an error.

See also this FAQ for other details of how to get the CGI setup.

Also check that you have use the Windows version of the search.cgi (not the Unix or Linux release). You can rename the search.cgi file to be search.exe and run it in DOS to check that it is the Windows version.

Can i used the renamed "search.exe" to get the results in DOS somehow. e.g.
search.exe zoom_query=Engineering

Ray · Oct-09-2009, 01:07 AM

We've just tested with a copy of XAMPP (both an old version V1.1.1.2, and the latest V1.7.2) and the CGI ran successfully with the default httpd.conf settings with no changes needed in the configuration file. We also tried a large index (over 100,000 pages) and it ran without issue.

Originally posted by toby View Post

But I'm not sure: Does the PHP Memory Limit also take effect on the Memory Limit for CGI's?

Or is my httpd.conf wrong? Do I have to put the three directives (RLimitMem, RLimitCPU and Timeout) somewhere else?

Try placing it globally, outside the mime_module settings.

These limits are not PHP limits exclusively. They are Apache limits for any process that it spawns. Since PHP is a script engine that is sometimes installed as a CGI and sometimes as a module, it too has to obey these limits.

You might want to check if you have any anti-virus/security programs which might be interferring as well.

If you still have trouble, zip up the set of index files which is not working, and e-mail them to us. We can verify if the files are okay.

toby · Oct-08-2009, 12:22 PM

Yes, this makes sense... I have to admit, that I'm fairly new to CGI and configuring an Apache.

I changed my httpd.conf to this:

Code:

<IfModule mime_module>
    TypesConfig conf/mime.types

    AddType application/x-compress .Z
    AddType application/x-gzip .gz .tgz
    AddType application/x-httpd-php .php
    AddType application/x-httpd-php .php3

    RLimitMEM 1073741824 1610612736
    RLimitCPU 300 420
    Timeout 300
    AddHandler cgi-script .cgi

    AddType text/html .shtml
    AddOutputFilter INCLUDES .shtml
</IfModule>

But the result is still the same :/

So, I've slightly modified your PHP "Test" script to this:

PHP Code:

<?php
$num_of_int = 20000000;
print("<h1>Begin test loop to allocate memory for {$num_of_int} integers</h1></ul>");

for ($i = 0; $i < $num_of_int; $i++)
{
    $dict[$i] = $i;

    if ($i % 5000 == 0)
    {
        print("<li>allocated memory for {$i} integers.</li>");
    }
}
print('</ul><h2>End test loop</h2>');
?>

The script end; I see the "End test loop"-line.

But I'm not sure: Does the PHP Memory Limit also take effect on the Memory Limit for CGI's?

Or is my httpd.conf wrong? Do I have to put the three directives (RLimitMem, RLimitCPU and Timeout) somewhere else?

David · Oct-08-2009, 09:49 AM

You might also want to have a look at this old post, which is an Apache configuration issues that causes PHP scripts and CGI's to die before they can finish executing.
http://www.wrensoft.com/forum/showthread.php?t=12

The larger the set of index files to more likely you are to have this problem.

toby · Oct-08-2009, 09:06 AM

Hi again!

I deinstalled my Zoom Search Engine (+ Plugins) from our computer. After that, I cleaned both directories: The installation directory and the "Wrensoft" folder inside the Own Docs/All Users/App data/.
Then, I downloaded the the current version and installed it. Zoom's "Help/About" confirms that I'm now on Build 1018. After installing the plugin-set and configuring my Zoom, I started indexing again... Same problem.

Then, I started to change some settings... Depending on the massive files (mainly PDF) we have, I set up Zoom's limits/configs to this:

Skip options: Skip words less than 3 chars
Indexing options: Indexing word rules: all character to join words are checked
Limits: max files: 20,000 // max unique words: 1,250,000 // max file size: 153,600 KB (150.00 MB) // max desc length: 150 chars // none of the four checkboxes are checked
Languages: UTF-8 // German.zlang // Stemming enabled for German // International search options: Enable accents/diacritic insensitivity for: accents, umlauts, ligature are checked // Use digraphs for umlauts is checked // substring match is checked

I saved this config file and started indexing, but there was no change in the result...

Then, I started indexing again (same config), but I stopped it after Zoom has approx. 3000 files indexed (before, when Zoom searches the whole directory, it has approx. 9500 files indexed). This does work!

So I guess it is because of the amount of unique words (~980,000) in ~9500 PDF documents. I think it has something to do with Zoom's database files (*.zdat) or its filesizes. Here is a comparison between the not working files (indexed ~9500 files with ~980,000 unique words) and the working one (indexed ~3000 file with ~500,000 unique words)

Code:

SUMMARY
==============================================================
                               |    WORKING    |  NOT WORKING
-------------------------------+---------------+--------------
files indexed                  |          3000 |          9570
unique words                   |       497,905 |       977,395
total words                    |    11,242,464 |    37,412,160
total bytes scanned/downloaded | 3,486,371,083 | 1,912,882,797


FILE SIZE
==============================================================
                               |    WORKING    |  NOT WORKING
-------------------------------+---------------+--------------
dictionary.zdat                |     14,714 KB |     31,105 KB
pagedata.zdat                  |        557 KB |      1,742 KB
pageinfo.zdat                  |         50 KB |        159 KB
pagetext.zdat                  |     54,732 KB |    187,657 KB
wordmap.zdat                   |     21,253 KB |     55,249 KB


---
(WORKING = stopped indexing after 3000 files)
(NOT WORKING = indexed full directory; 9500 files)