Kbase P137491: Are there ways to make the load of a .df faster?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
19/11/2008 |
|
Status: Unverified
GOAL:
Are there ways to make the load of a .df faster?
GOAL:
Can I speed up the load of a .df?
GOAL:
How can I make a .df apply faster?
FACT(s) (Environment):
All Supported Operating Systems
Progress/OpenEdge Product Family
FIX:
1. If you are renaming a lot of indexes:
This causes a rename of the current index to a temporary index name and the creation of a new index with the new name. This is likely causing the indexes to be built during the load of the .df file.To prevent this time consuming activity,add the indexes as inactive to speed up the .df load then you can either build the indexes with idxbuild offline or idxactivate or idxfix online. Idxbuild is multi-threaded.
2. If the databases are backed up you can use -i to reduce the volume of writes to the before image file, but that will most likely be insignificant in comparison to the number of reads to your data blocks, because of the need to read records to rebuild the indexes or the updates happening to the index blocks for the rebuild of the indexes defined in the .df file.
3. Putting the databases on multiple disks or RAID 1 or 10 sets will help decrease the time for the load.
4. You can also increase the -B to buffer more of the blocks read to optimize the reads of data for the building of the indexes if running online using idxactivate or idxfix.