I'll be curious to see if the new drive does away with some or all of the delays I'd gotten used to seeing between about 10:30 and midnight Pacific time (1:30 - 3 AM Eastern). I assume that's when backups and other system maintenance would be scheduled, and a failing drive might sometimes have been too busy doing retries to keep up with user requests for pages.

Thank you for the fast fix and especially for managing to preserve so much, maybe all, of the database.