>> Note that if you are using an HP EVA SAN, then it's not really necessary anyway.

Vincent, can you please offer proof of that assertion?

To OP, Jimmy also has a...

1) this is the first of 3 blog posts that are the Bible for troubleshooting deadlocks. http://blogs.msdn.com/bartd/archive/2006/09/09/Deadlock-Troubleshooting_2C00_-Part-1.aspx

2) If you are getting that much frag in a day then you...

Fraggle-805517 (9/23/2009)

---

Something I haven't seen anyone discuss is do you need to put each of these on a separate drive or do you still get the performance increase if you...

WayneS (9/23/2009)

---

Jeff - excellent advice. My question...

That's just SICK!! Blasphemy I tell you!!! 😀

Pretty impressive Marco!! I thought though that Jeff had said 10 was the sweet spot for the number of values. Looks like you are using 16? Course...

Gotta love this stuff!! Nicely done Lynn and follow-on posters too. Very self-aware of you Lynn to acknowledge that you lost sight of the big picture and forgot...

Sorry, but you have not given us enough information to advise you. How many spindles underly the proposed tempdb drive? Are these spindles shared for any other IO...

1) can't a RAID50 have half + 1 drives fail??

2) make sure you do things like sector alignment and 64K cluster size formatting.

3) I believe RAID 50 can be significantly...

1) get a professional tuner to help/mentor you and your system. You have admitted to spending weeks on this. A competent tuner could have given you a slew...

1) I believe you can actually force 2 indexes with the INDEX ( index_val [ ,...n ] ) table hint. Then you can see what the query plan is...

1) maybe it is slow because there are a kajillion joins? 🙂 This can lead to a) excessive IO (ESPECIALLY if indexes aren't available/used) and b) blocking. ...

>> "apples to cannonballs"

I like that one! 🙂 Curious why the OP would even imply these are equivalent...

How, exactly, are you doing the shrink? Are you using the gui or the tsql directly?

1) the log shrink was a bad recommendation

2) 50mb growth is too low

3) you mentioned no jobs were running. how did you verify this? did you run sp_who2...