> First thing to do, always, is state the openldap version you are using.
Sorry I forgot that, we are using OpenLDAP: slapd 2.4.23
> But one can say that the "no changes, no logs" properties holds for BDB also
I believe that myself too. Our logs were reasonable before, and with the change came the log overflow
I reply with our solution, for it seems we have solved the situation. After the installation of the lastbind overlay, we had an unrelated issue which forced us to perform a backup on one node (via slapcat) and restore into the other (slapadd). Since we have our main database for openldap and the accesslog's separated in two different directories, we should have performed the process (slapcat-slapadd) for each different database... but as it seems we did only for the main one. This situation (N1 with 2 databases, and N2 with only 1) triggered the massive log growth, forcing us to stop N2.
Therefore, the solution is: perform a cleanup of N2, and import both databases before starting the service. And log growth rate goes back to normal.
So I hope this helps if somebody else steps upon this same situation.
Thank you guys.

Oriol Rosa
Security Technical Consultant
SIA Spain, S.A.