I have had TREMENDOUS success at clients (as in 4+ orders of magnitude improvement) using dynamic sql to handle search/paging needs. If you are careful about SQL Injection this...

It is doubtful that an index on a table with 10 rows will help anything. Focus your efforts on the 20M row table.

Replication is a very complex topic. There are numerous tuning options you have if it isn't keeping up and that is causing the tlog growth. Probably best to...

Yeah, but you write REALLY sucky queries Steve!! 😀

Came over the top on me, eh?? Oh, I yield to the master!...

GilaMonster (5/3/2011)

---

Eddie Wuerch (5/2/2011)

---

Looks like you're rebuilding an index, or performing other index-modification work. You'll get lots of page latching on metadata pages (mostly the GAM) as the process allocates...

Grant Fritchey (5/2/2011)

---

Yeah, the best performance tuning questions come up on here, all the time. But you can also look at my blog, Paul White's blog[/url], Gail Shaw's blog, and...

ChrisM@home (5/2/2011)

---

Hi Virgil

This link[/url] will show you an alternative set-based method of calculating running totals and...

An index can still be fragmented after a rebuild. The most common reason is because virtually every server in existence has autogrowth managing the database sizes. I advise...

I will give you the same info I did the first time I posted: get a professional on your system for a quick review. There are just too...

Nice catch Eddie.

Azhar, since you missed this, I will say that you REALLY need to get a professional on your system for a review and to help you out. ...

I'll make it plus 2 million to Ninja's + 1 million 😀

1) run the free sp_whoisactive to check for blocking, resource contention and lots of other stuff

2) there are LOTs of things that could be at fault here. Way too many...

1) how many spindles underly the RAID sets you mention?

2) what else is on those RAID sets?

3) Have you checked for WHERE the actual bottleneck is? Assuming a SAN,...

Maybe your CPU utilization sucked on the old machines because the IO wasn't feeding them data fast enough and now on the new server it is. Outside of that...