After looking into disk I/O on this machine and doing more performance benchmarking my current thinking is that the problem isn't in the disk array afterall.
Initially -- not knowing how the array is configured or it's state -- I thought that slow disk writes and contention might be contributing to the problem. RAID 5 arrays are slow on writes and degraded arrays (those with failed or missing drives) are too, but neither look to be a factor here.
It's a perplexing problem -- we're running several busy Oracle sites, one with a database in excess of 10GB and over 15 million hits per month on less hardware with much better performance.
