search feature and architecture questions

[Sree]

Hi,

Iam very new to this product. Can you please clarify me couple of doubts:

1) Is OpenKM using Lucene for text search in documents and text files?
2) Is OpenKM supports text search in image documents? If so is it uses Lucene for text search in image documents?
3) In the architecture mentioned about EJB module. What is the use of EJB? Is it used for storing security information of documents?
4) If I use repository as a file system then where the security information will be stored for documents? Is information stored in database?
5) If I use repository as a database for storing both documents as blobs and security information then is it possible to access the documents using folder type browsing?

[jllort]

1- OpenKM core bases on jackrabbit that uses Lucene -> in fact lucene is used as search engine by OpenKM.

2- OpenKM professional comes with an open source OCR configuration to do it, on lucene goes text file from tif image - for example.

3- OpenKM can run with EJB and Without it, we made it implementation for some client security needs, but really we\'ve not needed to use it. Althought it we\'ll mantain for it reason.

4- Default configuration stores on file system. But you can store documents on a database. Really the good configuration, as we said in other post is a mix between structure nodes on a database an files ( blobs ) directly on hard disk. Doing it you\'ll get one of the best perfomance.

5- The hard disk structure file to save is not browsable directly because structure is obfuscated as names. Althought it the problem is the same between database or file system. The admin user of database or system always can acces to documents -> really it\'s the same secutiry problem but from diferent point of view. The security server restrictions or database ones are some different level than end user across OpenKM ( it\'s other security role ).

[Sree]

Thank you for clarifying my doubts.

I have only one doubt. Where exactly the file permissions will be stored?

[jllort]

Permissions that you see as OpenKM users, are stored on OpenKM internal node structure. The node structure could be stored on database or on file system as a data.

[phillman5]

Jllort,

>4- Default configuration stores on file system. But you can store documents on a database. Really the good configuration, as
>we said in other post is a mix between structure nodes on a database an files ( blobs ) directly on hard disk. Doing it you\'ll
>get one of the best perfomance.

>\"Files on OpenKM can be stored on several ways, dbms ( postgre, mysql, oracle ) or file system.\"

I am a little confused by what is meant by \'blobs\'? Is there a way to have OpenKM\'s folders and files replicate folders and files on the Hard Drive?

[jllort]

Blob is a database type, used to save binary. Forget this word !!!

Folders and documents are stored in a particular structure, you can not reply on hard disk as you see when navigate using OpenKM UI, but with webdav you can browser repository as a normal explorer OS window.

[phillman5]

Josep (Jllort)

Please see the discussion at TOPIC: Re:Having OpenKM store files using file system

http://www.openkm.com/Usage/1767-ReHavi ... .html#1767

Wiseman stated: I am sure there are many out there who would like the backend to be a filesystem instead of a database, and detailed instructions would be great.

For many reasons, mainly for recovery if things go wrong, I think we would adopt the Professional version of OpenKM for our project(s) if we could have the backend to be a filesystem (MacOS-X) instead of a database. If we (Wiseman and myself) could get more explicit directions how to do this, it would be very much be appreciated. We have gotten replies such as \"it is not simple to manage this\" to \"It is simple, but require some knowledge bout JDBC and Jackrabbit\", so I am left a little confused how involved this request really is.

[pavila]

If you want to know more about backend storage, you should read these pages:

* Jackrabbit Configuration
* Persistence Manager (PM) FAQ
* DataStore