Hopefully the "senior consultants" did more than just wave their hands
and mysteriously intone "add more partitions"? They should have
explained exactly why more partitions would help, and how to do
implement their suggestions, else they aren't who they claim to be.
Can you share the reasoning they provided?
It is a charlatan's trick to recommend a lengthy and complex course of
action, and then be long gone with the money before the recommendation
is attempted. In all likelihood, little or no improvement will be
realized. Then, of course, those who attempted to implement the vague
recommendation will be cited for having missed some important detail or
other.
Can you also provide more information about the 10 tables?
- what type of partitioning is used on these 10 tables (i.e. range,
hash, list, or if composite, range-hash, etc)?
- are all indexes local or are some global, particularly the unique
indexes enforcing the primary- or unique-key constraints?
- how much downtime is available to implement table and index rebuilds?
On 9/30/16 05:45, Balwanth B wrote:
Please need your suggestions for below
Hello everyone,
Need your suggestion on below
Version 11.2.0.4
We are having 10 production database tables with millions of records( biggest one is 450 million records... highly transaction databases).. some tables are already partitioned( now we have to increase the number of partitions up to 1024 partitions advice given by senior consultants based on same running at different environment)... All this tables have referential integrity between them.. can you please advice me what is best practice to increasing number of partitions and creating the partitions for existing tables...
Your suggestions will be much helpful.
Thanks,
Bobilli