I am still working on the search. I am configuring it to do an "index" search. This will be a separate Indexed file when a search string is entered. This dramatically reduces server strain.
Even a simple search like "2001" would crash the server. The way SQL works is that it would search all one million posts for that search string for the term 2001. Then, it would display the results for you.
An index search means that it would search only that index file (one file versus one million) and display the results for you. Theoretically, this would reduce server strain by about 90%. "Theoretically" being the key word there.
The downside to index only searches is that it wont be current until I create a new index. It may be one of those things I only do about every 6 months or so. I've been creating an index for 3 days now, and it is only 22% completed.
Google search is a great work around, and if this new search method that I'm working on doesn't bode well, google could just be the way we do it moving forward. Time will tell.
Stay tuned, and patient.