Xerox DocuShare Support & Software プリント

ページ / 10
Best Practices for Content Indexing
2
All DocuShare 6.6.1 servers 
Note: 
Perform these procedures before running dsindex index_all.
Indexing documents that contain many numbers, such as spreadsheets, consumes CPU, memory, and 
disk space resources. All numbers and unstemmed terms not found in a standard dictionary cause 
additional indexing overhead. To reduce the indexing overhead caused by these types of documents 
and unstemmed terms, follow the recommendation below on the MIME Types table, the idoltool 
UnstemmedTermTree, and the AutonomyIDOLServer IndexNumbers parameters. To improve content 
indexing speed, use the maxfilesize configuration to restrict the document file size.
Consider the site users, their needs and search patterns, then perform any or all of the following 
recommendations:
Disk Space and Memory Allocation
Size the server and the index resources. Know how much will be indexed. 
Gather metrics on projected document count, projected size of content store 
DSHOME\documents, and estimated index size. Refer to the "Estimating disk space" article 
in the DocuShare Knowledge Base.
Configure IDOL to match available resources. 
Set the IDOL configuration to match the available RAM resources on the DocuShare server. 
Increase performance by taking advantage of additional RAM when running IDOL on a 64-
bit Windows server. Run the appropriate idolsetup command to configure for the available 
RAM. 
Note: 
A minimum of 6 GB of RAM is recommended for DocuShare running on 64-bit Windows 
servers. 
For 4-8 GB RAM, run DSHOME\bin\idolsetup.bat..\config\idol_default.config
For 8-16 GB RAM, run DSHOME\bin\idolsetup.bat..\config\idol_medium.config
For 16-32 GB RAM, run DSHOME\bin\idolsetup.bat..\config\idol_large.config
For 32+ GB RAM, run DSHOME\bin\idolsetup.bat..\config\idol_unlimited.config
DocuShare Database Maintenance
Verify the existence of database indices 
Run resetIndexes to rebuild the default set of DocuShare database indices, if necessary.
Optimize the database 
Use either UPDATE STATISTICS (SQL Server), GATHER STATISTICS (Oracle), or VACUUM 
(PostgreSQL) to optimize the database. 
Schedule database optimization to run daily.