My stance on clustered databases is that you replace hardware risk with software risk. So they make sense if they get you better throughput or better uptime sta

My stance on clustered databases is that you replace hardware risk with software risk. So they make sense if they get you better throughput or better uptime stats, but you've still got to plan for a software bug bringing down the whole thing.