RFI eliminates most of the need for suiciding the log. The two most common reasons for suiciding the log in the past were:
No backups are available or the backups are too old.
Insufficient time to restore.
There should never, of course, be a situation where backups are unavailable or are too old. Unfortunately that is too often the case, either because the dumps are bad or due to poor planning. In such situations, suiciding the log may be the only recourse. Aside from those situations, however, you should never consider suiciding a viable option.
DBAs should test all backup and restore procedures before
relying on them. If you attempt to load a dump on the original database
and it completes only partially, you will have eliminated the possibility
of using that database again and may even have eliminated the final
chance to recover data by suiciding the log.
