Hi everyone!

Generally speaking, is there any good pratices that i can try in order to
improve insert performances on a big table? I can't just throw the indices
away, since the table is very big and select performance would be
drastically reduced.

Some detais:

- Big table ( > 30 million rows)
- Almost continuous input of data: there are transactions with about 100
inserts almost every minute or so.
- Selects are made depending on the client demand

ONE possible solution that i'm looking at is to remove old data ( > 3
months old) and do a backup/restore after it, eventually transforming this
on a policy, removing old data on a daily basis.
Also, i disabled automatic sweep and i'm running it on a daily basis, at
the dawn.

I'm not looking for a magic bullet that will solve my problem at once, but
any tip that can give me a slightly performance improvement it will be very
welcome in the long run.

Tks in advance.


[Non-text portions of this message have been removed]