Creating IVF text indexes occupies a large amount of memory which comes from OS, not DCCA? (DBMR2242)
Creating inverted-file text indexes requires a large amount of memory resource. The memory extra allocated for inverted-file text indexes by OS doesn’t occupy the shared memory – DCCA and it will be free after use. Users can manually specify the approximate upper bound of the memory by adding the keyword DB_IFMem in dmconfig.ini. If users have no plan to use IVF text indexes, there is no need to set the keyword DB_IFMem.
Version: DBMaker4.x/5.x
Product: Bundle/Normal
Platform: Windows/Linux