I am from Germany and am planning to migrate my business to XWiki as a new internal knowledge base. During my Proof of Concept, I came across the following limitations regarding the search functionalities in XWiki.
(1) Solr Indexing Cycle not configurable
Under Settings → Administer Wiki → Search → Search, I found no configuration regarding the Indexing Cycle job. Does the wiki automatically perform an index delta-update after every new change in the wiki (changes on a wiki page or new attachments)?
My understanding is that under Settings → Administer Wiki → Search → Search → Indexing, it is possible to manually trigger an index delta update with “Add to index,” correct? If so, I wasn’t able to find the commands to instruct the system to “Add to index” all the new PDF attachments added to the wiki since the last index sync. This has led to more open questions regarding my observation (2).
For my observation (1) I red already the following forum articles
Solr advanced searches - is there any documentation?
(2) Not all PDF attachments or large text content is getting indexed - size limitations defined by Xwiki
I performed a test upload of several public government PDF documents, sometimes up to 500 pages per PDF document. Is there a limitation regarding the number of indexed words or PDF pages per PDF attachment? It seems that for smaller PDF documents, XWiki has no indexing problems, but larger PDF files are not shown in the search. The PDF documents had sizes ranging from 1 to 13 Megabytes. I even took the 500-page document, which wasn’t shown in XWiki search, and deleted/reduced its page count to 70, and everything worked fine for the search engine test.
#inside text test
For the “inside text test” I attempted to copy the 500 pages of text into an XWiki page as site text content. This load test showed that the text amount for an XWiki page is also limited. After approximately 200 PDF pages of text, I received this message: “An authorization error occurred when performing this action. You might have been logged out since you started to edit this page.” During this load test, the text editor was very buggy, even when loading 50 pages into the site. After saving the content of 50 pages, I was immediately able to search and find this newly added content via the global wiki search.
For my observation (2) I red already the following forum articles
Attachment upload with bigger sizes sometimes not working
No attachments appear in search results
Problem with max attachement size XWiki 14.10.4
Improved Quick Search
Attachement are not recorded
Search suggestion only showing attachments, search only documents
Xwiki page text not included in solr search results
Content from new documents cannot be searched (Solr)
Note: I am new to Docker and XWiki. Please provide hints, advice, or existing forum content to help me delve deeper into possible configuration activities on the file level in Docker or in the local XWiki folder.
I hope that by solving these questions, even more people can be motivated to migrate to the wonderful XWiki.
XWiki automatically react to any page modification to trigger the indexing of the page. Most of the time it feels immediate but what happen in practice is that the order to index of the page is put in a low priority queue so if there is really a lot of activity you might need to wait a little bit before the new content is fully indexed (you can see the size of the queue in the administration).
