Frequently Asked Questions

• How can I contribute to the project?

  There are several ways you can help. Please visit the Help Us page on the mnoGoSearch website.

• If I have pages with no keywords in the META description, will the document get indexed?

  Yes. By default mnoGoSearch indexes the entire document. By changing the default values you can give different weights to different parts of the document, like body, title, keywords, description.

• I'm wondering if mnoGoSearch supports multi-byte character sets. I'm especially interested in EUC-JP.

  Multi-byte character sets (including EUC-JP) are supported starting from the version 3.2.x.

• Why does indexer fail with a message like can't load library 'libmysqlclient.so.6'?

  You probably don't have MySQL libraries properly installed. On Linux, just find out where your libmysqlclient.so.6 is, then add the path to this file into /etc/ld.so.conf. Then run ldconfig as root.

  You can also set the value of the environment variable LD_LIBRARY_PATH to the path to the required dynamic libraries. This works on any OS.

  Note that this is a general UNIX configuration issue, not only mnoGoSearch specific.

• What is the maximum data size mnoGoSearch can can handle?

  When using the blob storage mode, introduced in the version 3.2.8, mnoGoSearch is able to index and search through about one or two million documents on a single computer. The mnoGoSearch cluster introduced in the version 3.3.x can distribute data over multiple servers and is able to index up to about 500 million documents when using 256 machines.

• I want to index multiple virtual domains on a single box. What settings do I need to change for that?

  Virtual domains are fully supported. No additional settings are required.

• Is it possible to use different templates for searches, i.e. view search results in the design of different sites without having to recompile search.cgi?

  There are several ways to do it. They are given here in the order how search.cgi detects template name:

  1. search.cgi checks the environment variable UDMSEARCH_TEMPLATE. Thus you can put the desired search template file name into this variable.

  2. search.cgi also supports Apache internal redirect. It checks the REDIRECT_STATUS and REDIRECT_URL environment variables. To activate this template option you may add these lines to Apache srm.conf:

     AddType text/html .zhtml
     AddHandler zhtml .zhtml
     Action zhtml /cgi-bin/search.cgi
     

     Put search.cgi into your /cgi-bin/ directory. Then put the HTML template into your site directory structure under any name with the .zhtml extension, for example template.zhtml. Now you can open the search page: http://www.site.com/path/to/template.zhtml. You can also use any available extension instead of .zhtml.

  3. If the above two methods fail, search.cgi will open the template having the same name with the script being executed, using the SCRIPT_NAME environment variable. search.cgi will open the file ETC/search.htm, search1.cgi will open ETC/search1.htm and so on, where ETC is mnoGoSearch /etc directory (usually /usr/local/mnogosearch/etc). So, you can use the same search.cgi with different templates without having to recompile it. Just create one or multiple hard or symbolic links to search.cgi or copy it under different names then put the corresponding search templates into the /etc directory of your mnoGoSearch installation.

• indexer doesn't store some links in the database although they are expected to be stored. indexer doesn't grab all the links from my web site, as a result only half the site is indexed.

  There are two possible reasons:

  1. indexer finds the links, but rejects to put them into the database. This most likely means that your indexer.conf configured not to allow these links.

  2. indexer doesn't find the links on the page because of a bug in the HTML parser.

  Imaging that http://localhost/links.html is a page that contains ignored links.

  To discover the reason, start indexer with these command line arguments:

  indexer -am -u http://localhost/links.html

  It will display various debug information, including:

  • every link found on this page found

  • information why indexer rejects or accepts every link found

  With this information provided it is easy to detect what is the reason of unexpected indexer behavior: either 1 or 2. When you know the reason, you can either change your indexer.conf and check again, or file a bug if you're sure that indexer.conf is OK.

  When reporting this sort of bugs, please include indexer -am -u http://localhost/links.html output, as well as the Server/Realm and the Allow/Disallow commands from your indexer.conf.