Kbase P102421: When converting a database to UTF-8 how much larger should the -B parameter be?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
6/3/2010 |
|
Status: Verified
GOAL:
When converting a database to UTF-8 how much larger should the -B parameter be?
GOAL:
How to determine the -B for a UTF-8 database.
FACT(s) (Environment):
All Supported Operating Systems
Progress 9.1E
OpenEdge 10.x
FIX:
The amount of space required for shared memory for a UTF-8 database compared to a standard single byte database can not easily be predicted. When a database is converted to UTF-8, all characters currently in the database are only taking up one byte, however, once in UTF-8 mode, some single-byte characters take 2 bytes and some double-byte characters take 3 bytes in comparing with native ANSI codepages.
Based on that the -B parameter should be increased to ensure good performance, but, by how much can only be determined by monitoring the database via promon. Check to see what the buffer hit ratio is and set the -B high enough that the buffer hit ratio falls between 95 % and 98 %, if there is ample available memory space to do so.