Partitions size on EXADATA
Michael Nosovsky, July 25, 2016 - 5:05 pm UTC
The answer has helpful information, but it does not answer the main question: if the partitioning strategy is different on EXADATA versus not EXADATA.
July 25, 2016 - 7:02 pm UTC
I would not think of it in terms of exadata vs non-exadata, but more in the general sense of server firepower.
If I can gigabytes per second, then I *might* lean toward larger partition sizes to keep parse times and dictionary size smaller because I know that the reduced granularity will be made up for by my I/O bandwidth.
(I'm assuming here that we're well above any sizes that might impact things like storage indexes etc...ie, even our 'smallest' parttion sizes will be measured in megabytes not kilobytes)
A reader, July 26, 2016 - 2:18 am UTC
Thank you!