Dear all,
Please humour me if this is a silly comment , or has been made before -
but....
when Monty posted the original message I had a few thoughts - ........
I understand that there is a issue with
loop many times
perform a set a updates/inserts
run vauuum
end loop
I would not peform a test this way.... let Postgresql MVCC sort out the
relevant rows to manage... so run vacuum once only....after all on a real
system you try to minimize the number of times vacuum ( or its related
equivalents in other dbms's - e.g : analyze ) are run ( as they are typically
expensive ).
So I guess I am saying that if the Benchmark requires multiple runs of vacuum
/ analyze type functions then maybe the design of the benchmark should be
examined a bit...
regards
Mark