Error reindexing attachments in Drupal 7 using Apache Solr Attachments module

Question

I've installed Solr 3.4.0 and Tika .10 on my Windows 2003/IIS6 server. I have loaded the Apache Solr Attachments module and configured so I can search on documents attached to pages in my Drupal 7 web site. When I reindex on the Apache Solr search configuration page I'm getting 5 of these notices:

Notice: Undefined index: filepath in apachesolr_attachments_add_documents() (line 133 of W:\www\drupal-7.8\sites\all\modules\apachesolr_attachments\apachesolr_attachments.index.inc).

The search index page indicates the attachments have been submitted for reindexing (message on the page -> "There are 0 items remaining to be examined for attachments out of 23 total. The search index is generated by running cron. 100% of the site content has been sent to the server. There are 0 items left to send. " and "Number of documents in index: 23 (0 sent but not yet processed)"). On the Search settings page I see "100% of the site has been indexed. There are 0 items left to index." under Indexing Status.

Search is working for web site content but search for content in attachments is not working. If I got into Solr admin and query I can see web site content but searches for words in attachments return no results. So the attachments are not getting into Solr. Any ideas about what this error means or how I can debug it?

Answer 1

So I figured this out in case anybody else runs into this issue.

In the 7.x-1.x version of Apache Solr Attachments, there is a line of code in apachesolr_attachments.index.inc that uses the PHP function mb_detect_encoding. That command relies on the PHP EXIF extension being installed and it is not by default. Once I modified my PHP installation to include it, it was able to process the files. I ran into some other issues though but finally got it all working. Details are posted here -> http://drupal.org/node/1377416