HTTP Server for i

Secure, powerful, and complete

Overview

Getting Started

Documentation

Support

iSeries Webserver Search Engine - Trouble Shooting

This page contains the documentation that is specific to trouble shooting. Be sure to read iSeries Webserver Search Engine - Getting Started.

	Technical Details and Troubleshooting
	Tips and Techniques: Sample HMTL, Net.Data macro, and PTFs
	Traces for search and the web crawler

Technical Details and Troubleshooting

This section contains an assortment of errors that can occur.

Any errors with NET.DATA - See Net.Data Tracing or Traces for search and the web crawler.

Error: NET.DATA Error: File /search/sample_search.ndm not found. 

If you entered , for example, the URL  http://rchas123:2626/cgi-bin/db2www/search/sample_search.ndm/simple, then the sample_search.ndm should be in directory /search. If it is not, the error will occur. Change the directory on the URL or move the sample_search.ndm to the /search directory. 

Error: NET.DATA Error: File /search not found. 

If you entered , for example, URL http://rchas123:2626/cgi-bin/db2www/search/sample_search.ndm, you need to add /simple to point to the HTML block you want to run so that the complete URL is http://rchas123:2626/cgi-bin/db2www/search/sample_search.ndm/simple

Error: Error 403 when attempting to display a document found on a search 

If you receive Error 403, the browser is unable to serve (display) the file because of some problem associated with the directives in the configuration file. If you are running on an Apache server, make sure the Alias directive correctly associates the URL path to the actual document. If you are running on an Original server, check the Pass directives. This error usually occurs because there is no Alias or Pass directive for the directory specified in the URL. 

Here is an example Pass directive. For this example, we will assume it is the only Pass directive in the configuration file. 
                             Pass /shoes/*     /html/products/consumer/shoes/*
The second part of the Pass directive ( /html/products/consumer/shoes/*) indicates the parent directory for all of your documents. The first part of the Pass directive ( /shoes/*)  is used to map the path seen in the URL to the actual document path indicated in the second part of the URL.  If the URL  for the document displayed on the search results form is not URL:/shoes/...... followed by the rest of the path to the document, then the error will occur. Refer to URL Mapping Rules File for more details about the Mapping Rules file and your configuration. 

The index directory must be an IFS directory  such as /mydir. It cannot be a QSYS.LIB file member such as / QSYS.LIB/MYLIB.LIB/... Use the default index directory and try to create the index again. 

Error:  An error occurred in memory allocation when creating an index. 
If this error appears in the file errors when you View status of search index, you have attempted to index a document that requires more memory than the system has available. 

Error:  The index is currently busy. Please try your request later when updating an index. 
This message will be sent if someone else is updating, merging, or deleting an index when your request is made. Only one process can write to the index at a time. However, multiple requests for searches are allowed. 

Files in QSYS.LIB found on a search do not display correctly or display on one continuous line. 
In order for HTML or TEXT files to display correctly, set the file type to HTML. This can be set using STRSEU and bringing up the list of file members.   Change the TYPE to HTML as  below.    If you are using the ADDPFM command to add members to a file, you can set the parameter SRCTYPE to HTML  for each member you are adding to the file. 

                                      WORK WITH MEMBERS USING SEU 

The search engine administration is designed so that an index can be searched while the same index is being updated or merged. 

The CL commands such as CFGHTTPSCH send search engine and webcrawler errors within various HTPxxxx messages.

Example: You use CFGHTTPSCH to create your index but get error HTP1608 Request to create an index failed. Reason GEN5074.

To determine the meaning of GEN5074, press F1 on the message to see the Cause section in the second-level text. The meaning of reason GEN5074 is inserted there:
"The document list file cannot be found. Please check the name, the directory path, and the document list type and try your request again."

Does the Search Engine have any limits or restrictions? New addition!

The search engine has the following limits:

• The maximum number of documents allowed in a document list is 330,000.
• Very large files such as a pdf file can cause an error during indexing. Remove that file from the document list.
• Wildcard characters in search strings are not allowed for double byte languages. A wildcard search is implied for double byte languages. Strings within double byte words can be found without the need for wildcard characters. For example, "AccessLog" in Chinese can be found by searching for the string "Access".

How do I find out the status of my index or document list without using the browser? New addition!

Select options *PRTIDXSTS or *PRTDOCLSTS on the CFGHTTPSCH command. The output is sent to a spoolfile.

Authorization concerns 

The search engine administration will create the index directory for you with PUBLIC *RWX (read-write-execute) authority. If you use an existing index, make sure the authority on the parent directories is at least *PUBLIC *RX to allow users to traverse the path. Several files are created by the search administration in this directory.  If the authorization is incorrect, errors using the following functions will occur: 

• Searching the index
• Viewing the status of the index
• Creating a document list or mapping rules file in the index directory

Editing errors using Client Access

If you create a document list into an IFS file using the browser forms or the CFGHTTPSCH command into the IFS file system, the file will be tagged with your job CCSID. This is an EBCDIC CCSID. If you want to edit the document list through Client Access, you will discover that the file is unreadable. In order to edit the file, you must FTP the file from the iSeries to your PC, delete the file from the iSeries, edit the file, and then FTP it back to the iSeries. This will cause an ASCII file to be created. To check the codepage of an IFS file, enter an 8 next to the file and look for the codepage. For Latin 1 languages (English, Italian, French, German, etc.) the ASCII codepage is 819. If you do not see 819, the file probably has an EBCDIC codepage. A document list can be tagged with either an ASCII or EBCDIC CCSID. 

It is possible to use the same index name more than once as long as you are using a different index directory each time you create a new index. 

The search administration will create an index directory for you if one does not exist. When selecting to have the search administration create an index for you, be sure that all directories in the index directory path have enough authority to create new sub-directories as well as create files in the new directory. For example, if you specify an index directory called /footwear/shoes/tennis/index, directory /footwear and  /footwear/shoes must have *PUBLIC *RX authority. Directory /footwear/shoes/tennis must have *PUBLIC *RWX authority for the index administration to create the new directory /footwear/shoes/tennis/index. 

Immediate vs Background Indexing 

Select to create, update, or merge an index immediately when just a few documents are indexed at once. The sample above for indexing the sample recipe file can be created immediately. There are about 100 HTML documents that are indexed. The documents average between 500 and 2000 bytes each. 

Select to create, update, or merge an index as a background task when many documents are indexed at once to avoid the possibility of the browser timing out during the request. 

Files created during indexing 

The following files are created in the index directory during index processing. If the character string, xxxx is specified as the index name, the following  files will be created in the index directory. The size of this set of index files will be significantly larger than the total size of the documents that are indexed. 

      xxxx.KEY  xxxx.POS  xxxx.LEY  xxxx.QOS  Created when you create a new index (main index) 
      xxxx.K_Y  xxxx.P_S   xxxx.L_Y  xxxx.Q_S     Created on update index  (supplemental index) 

  The following files will be temporarily created in a work directory called /indexdirectory/ WRKGTRx where x is a number .

      xxxx.K!1  xxxx.K!2            xxxx.P!1  xxxx.P!2 
      xxxx.Knn (where n is digit)  xxxx.Pnn (where n is digit) 
      xxxx.L!1  xxxx.L!2              xxxx.Q!1  xxxx.Q!2 
      xxxx.Lnn (where n is digit)   xxxx.Qnn (where n is digit) 

The total amount of space needed for the temporary files above is one to four times larger than the primary index when creating a new index or than the supplemental index when adding documents to an existing index. 

Files errors during indexing 

If file errors occur when documents are being indexed, the errors will be listed at the end of the View status of search index form. Only the first 100 errors are listed. Correct the errors or remove the files in error from the directory and then update the index using the same document list. 

Deleting Documents from an Index 

Deleting documents from an index forces the whole index to be rewritten. This will take some time depending on the size of index. It is best to withhold deleting documents until you have a large set to delete. 

Incorrect national characters in the document title

If documents are not tagged with the correct CCSID, national characters might display incorrectly on the search results, although not necessarily when you display the document itself. For example, if your document is tagged as 850 but the encoding of the contents is 819, the title of the document might contain incorrect characters.

To trouble shoot this, do the following:

• Display the document directly from the browser. If, for example, the document is tagged as 819 but the contents are in EBCDIC or CCSID 37, you will see garbage displayed.
• If the document is tagged 850, for example, but the title on search results shows wrong national characters, you might have documents encoded in 819. The hex values for characters are different for each of these encodings. If you are FTPing the files from your PC, make sure the CCSID is set correctly. See also National Language Support Considerations and specifically More about CCSIDs and character encoding. If you encounter this type of problem, be sure to create a new index once you have correctly tagged documents.

Back to top