I have two databases in the same cluster that are almost identical. One is a copy of the other as we are developing some new features in the copy.
My problem is that the exact same simple query performs great in the original database ("ises") and dismally in the copy database ("ises_coelacanth"). The problem is that in ises, it uses an index scan, but in ises_coelacanth it uses a sequential scan:
The difference is that Postgres is coming to alternate conclusions as to what plan to use. Given that the copy is causing the 'problem', the question to ask is; did you run ANALYZE on the table once the data was copied in?
-- Moshe Jacobson Nead Werx, Inc. | Manager of Systems Engineering 2323 Cumberland Parkway, Suite 201 | Atlanta, GA 30339