Difference between revisions of "Search Indexes"

[checked revision][checked revision]
 
(4 intermediate revisions by the same user not shown)
Line 12: Line 12:
 
* Text files (TXT)
 
* Text files (TXT)
 
* HTML files (HTM and HTML)
 
* HTML files (HTM and HTML)
 +
* PDF files (PDF)
  
 
Typical tasks regarding indexes are described in the following sections.
 
Typical tasks regarding indexes are described in the following sections.
Line 21: Line 22:
  
 
* '''Plain Text Files (TXT, CSV)'''<br/>The IFilter responsible for these file types is shipped with Windows by default. In case this option is disabled, a registry values might be wrong. Open the registry editor and verify that the ''Default'' value of the key ''HKEY_LOCAL_MACHINE\SOFTWARE\Classes\.csv\PersistentHandler'' is set to ''{5e941d80-bf96-11cd-b579-08002b30bfeb}'' and correct it if necessary.
 
* '''Plain Text Files (TXT, CSV)'''<br/>The IFilter responsible for these file types is shipped with Windows by default. In case this option is disabled, a registry values might be wrong. Open the registry editor and verify that the ''Default'' value of the key ''HKEY_LOCAL_MACHINE\SOFTWARE\Classes\.csv\PersistentHandler'' is set to ''{5e941d80-bf96-11cd-b579-08002b30bfeb}'' and correct it if necessary.
* '''Portable Document Format (PDF)'''<br/>For indexing PDF attachments, install [ftp://ftp.adobe.com/pub/adobe/acrobat/win/11.x/PDFFilter64Setup.msi Adobe PDF iFilter] on 64bit operating systems, or [ftp://ftp.adobe.com/pub/adobe/reader/win/11.x/11.0.10/en_US/AdbeRdr11010_en_US.exe Adobe Acrobat Reader 11] on 32bit operating systems (Internet Explorer or FTP client necessary for download).
 
*:'''Important Notices:'''
 
** Take note of the [https://www.adobe.com/devnet-docs/etk_deprecated/tools/AdminGuide/Acrobat_Reader_IFilter_configuration.pdf installation instructions] of the 64bit IFilter. Especially adding the ''bin'' folder of the IFilter installation folder to the ''PATH'' system variable increases indexing speed a lot.  The ''PATH'' system variable can be adjusted via an administrative command prompt (cmd) and then rundll32 sysdm.cpl,EditEnvironmentVariables
 
** Newer versions of Adobe Acrobat Reader do not contain an IFilter. Thus, please deactivate the automatic update function of Adobe Acrobat Reader 11.
 
 
* '''Microsoft Office (97-2003), Microsoft Office (2007 and later),'''<br/>All these file types are supported by the [https://www.microsoft.com/en-us/download/details.aspx?id=17062 Microsoft Office 2010 Filter Pack].
 
* '''Microsoft Office (97-2003), Microsoft Office (2007 and later),'''<br/>All these file types are supported by the [https://www.microsoft.com/en-us/download/details.aspx?id=17062 Microsoft Office 2010 Filter Pack].
 
*  '''Open Document Format (Libre Office/Open Office)'''<br/>These file types require a working installation of OpenOffice or [https://www.libreoffice.org/ LibreOffice]. Latest version tested was ''LibreOffice 5.1''.
 
*  '''Open Document Format (Libre Office/Open Office)'''<br/>These file types require a working installation of OpenOffice or [https://www.libreoffice.org/ LibreOffice]. Latest version tested was ''LibreOffice 5.1''.
Line 54: Line 51:
 
* when restoring restoring backups without search indexes.
 
* when restoring restoring backups without search indexes.
  
In these cases it may occur that archive, index and settings are no longer synchronous and that individual indexes must be rebuilt. Generally, a corresponding error message will be displayed. In case multiple indexes should be rebuilt, make sure other processes such as archiving profiles, exporting profiles or jobs are paused, or else the indexing process may get interrupted.
+
In these cases it may occur that archive, index and settings are no longer synchronous and that individual indexes must be rebuilt. Generally, a corresponding error message will be displayed. In case multiple indexes should be rebuilt, we recommend to start MailStore Server in [[MailStore_Server_Service_Configuration#Controlling_the_Service|Safe Mode]] to avoid interference by background tasks such as archiving profiles, exporting profiles, jobs or user interaction.
  
 
To rebuild search indexes please proceed as follows:
 
To rebuild search indexes please proceed as follows:

Latest revision as of 10:24, 25 August 2021

MailStore Server offers users an extremely fast full-text search. All emails a user has read-access to are searched, in most cases in only fractions of a second. To ensure this remarkable speed, MailStore Server sets up so-called search indexes during archiving. They work in a way similar to the indexes often found in the back of books: looking up something in an index gets results significantly faster than searching each single page.

MailStore Server maintains one index file each

  • per file group and
  • per user.

MailStore Server can index all file types for which a so-called IFilter is installed on the MailStore Server computer. Typically, IFilters exist at least for all applications which are installed on the respective machines.

For reasons of stability and performance, MailStore Server processes the following file types directly, regardless of the IFilters that are installed:

  • Text files (TXT)
  • HTML files (HTM and HTML)
  • PDF files (PDF)

Typical tasks regarding indexes are described in the following sections.

Install Missing IFilters

Typically, IFilters exist at least for all applications which are installed on the respective machines. If, for example, Microsoft Office 2013 is installed, the corresponding IFilter for Microsoft Office documents are installed as well.

Install the following IFilters on the MailStore Server computer to index the corresponding file types. Restart the MailStore Server service after the IFilter installation to let MailStore detect the newly installed IFilters.

  • Plain Text Files (TXT, CSV)
    The IFilter responsible for these file types is shipped with Windows by default. In case this option is disabled, a registry values might be wrong. Open the registry editor and verify that the Default value of the key HKEY_LOCAL_MACHINE\SOFTWARE\Classes\.csv\PersistentHandler is set to {5e941d80-bf96-11cd-b579-08002b30bfeb} and correct it if necessary.
  • Microsoft Office (97-2003), Microsoft Office (2007 and later),
    All these file types are supported by the Microsoft Office 2010 Filter Pack.
  • Open Document Format (Libre Office/Open Office)
    These file types require a working installation of OpenOffice or LibreOffice. Latest version tested was LibreOffice 5.1.
    Notice: Though Microsoft's Office 2010 Filter Pack registers support for Open Document Format files, indexing does not work with that IFilters.

Additional information about IFilters can be found in the corresponding Wikipedia article.

Setting Up Indexing for the Contents of File Attachments

In the standard configuration, MailStore Server includes the file names of file attachments in the search indexes but not their contents. To enable MailStore Server to search the contents of file attachments, it has to be configured accordingly. Please proceed as follows:

  • Start MailStore Client and log on as administrator (admin).
  • Click on Administrative Tools > Storage and then on Search Indexes.
  • In the area Attachments click on Change...
  • Select the file type groups or enter a space separated list of file extensions of attachment types you would like archive in the text field below Other File Extensions.
    Tech index 02-9 7.png
  • Click on OK to save the settings.

MailStore displays a notice, if file extensions were added for which no IFilter is installed on the MailStore Server machine.

The new settings apply to all emails that are archived as of now. To apply the settings to already archived emails, rebuild the appropriate search indexes as described below.

Rebuilding Search Indexes

Normally, the creation, maintenance and usage of search indexes is completely transparent meaning that neither administrators nor MailStore users need to know of their existence or their internal workings. In some cases, however, some maintenance may become necessary. For example:

  • because of power outage,
  • because of unexpected termination of the MailStore Server service
  • because of missing network connectivity (only with storage on an NAS),
  • because of changes to the index configuration or
  • when restoring restoring backups without search indexes.

In these cases it may occur that archive, index and settings are no longer synchronous and that individual indexes must be rebuilt. Generally, a corresponding error message will be displayed. In case multiple indexes should be rebuilt, we recommend to start MailStore Server in Safe Mode to avoid interference by background tasks such as archiving profiles, exporting profiles, jobs or user interaction.

To rebuild search indexes please proceed as follows:

  • Start MailStore Client and log on as administrator (admin).
  • Click on Administrative Tools > Storage and then on Search Indexes.
  • Check all search indexes to be rebuilt. Indexes with the status Please Rebuild are already checked for your convenience.
  • Click on Rebuild Search Indexes.
    Tech index 01.png