I'm sure I'm not the first person this has happened to. I have two servers both running 2008R2. OLDBOX is a standalone machine using its own internal RAID array. NEWBOX is our brand new system (faster, more RAM) and is connected to our brand new SAN, where I have a nice chunk of high-speed discs.
Every night, we run an ETL to get data into MAINDB, then users use a front-end that runs reports against MAINDB all day. I have identical copies of MAINDB on OLDBOX and NEWBOX, and NEWBOX beats OLDBOX's performance by a solid 10-20% for running reports.
Oddly, however, when we're doing the nightly load itself, NEWBOX has bizarre hangups on certain parts of the load that deal with a lot of data. OLDBOX, for example, processed the data in 2 minutes. NEWBOX took 30!
I examined the execution plans for the same query on both machines, and they seem more or less identical minus some parallism consolidations. One certainly isnt using an index the other isnt or anything. It's the same DB! Even after I created a covering index for the load on NEWBOX, the load still took 7 minutes.
Does anyone have any other ideas for where I could start as far as trying to get to the bottom of this? I don't see a reason for NEWBOX to load this data slower than OLDBOX.