in reply to
Combining Ultra-Dynamic Files to Avoid Clustering (Ideas?)
I would just use BerkeleyDB. It solves this exact problem and can handle terabyte-size databases.
Comment on
Re: Combining Ultra-Dynamic Files to Avoid Clustering (Ideas?)
In Section
Seekers of Perl Wisdom