Hi, We have a significant Process Chain that we run 3 times daily. 2 times occur during the day when users are actively reporting. Before each chain run we drop the indexes and then rebuild them at the end of the chain. This has worked fine, but I am concerned about reporting performance during the 45 minutes the chain is running.
I'm curious what practices people use or what the recommended best practice should be.
1. Drop all InfoCube Indexes at beginning of chain and rebuild 45 minutes later at end of chain?
2. Drop and rebuild each infocube index as it's being loaded?
3. Don't Drop and rebuild the InfoCube Indexes during the day when users are active.
I'm thinking this route because we are just loading deltas and it shouldn't take much more time to load with the infocube indexes in place.
4. Any other ideas?
Thanks for any opinions.