You can utilize parallelizing export when using data pump,but for traditional export use the following
Export of Databases - reads data by running a select statement of the data and generating the DDL to perform the import process in future.
Fast Exports
1) Use direct=y 2) Until disk is not fully utilized try running exports in parallel. 3) Keep export file on different disk then the datafiles. 4) Run exports in two part rather than one i.e. 1st rows=n and second as indexes=n rows=y constraints=n.
Fast Imports
1) Use the first file out of two files created with the exports, It will insert all data but will not create any indexes and constraints. Once the data insertion is done run imp with indexfile option to extract script of index creation in text file. edit the file to include parallel clause and set parameters db_file_multiblock_read_count to 128 with sort_area_size to a higher value and workarea_size_policy=AUTO.
No comments:
Post a Comment