The LDAP server was saturated by the RCC logging of error messages.
Could we consider buffering error writing (every 1s for exemple)
