Euan Garden - Tour of SQL Server team (Part IV)

AWESOME!

Here is a link to Jamie's site:

http://www.sqlserverdatamining.com/DMCommunity/

What's the estimated cost for the whole machine (Superdome + hard drives)?

When we did the TPC benchmark with this machine it was $5.1 mill, you can get the disclosure doc here:

http://tpc.org/tpcc/results/tpcc_result_detail.asp?id=103082701

However this was 18 months ago, prices have come down for most stuff, but the box now handles 1TB of memory instead of 1/2 TB of memory. The procs are now 2GHz instead of 1.5Ghz but you get the idea

Thanks for the link! There's a lot of great info in the full disclosure report.

The Channel 9 Team wrote:

Soner Terek takes us behind an HP SuperDome computer. A terabyte of RAM. 2,500 hard drives. 64 64-bit processors.

Then Jamie Maclannan talks about data mining and gives us a demo.

This is the fourth, and final, part of the SQL Server team tour. Here's the other three parts:

Part I
Part II
Part III

Video length: 00:15:31.

I lOve gigabits, lots of storage and terrabytes!

TimP wrote:

What's the estimated cost for the whole machine (Superdome + hard drives)?

It costs, but its needed!

I get it: data mining is better distinguished from OLAP where I'm standing. AI... AI...

with a 2500 disk array, how often do you have to swap drives out?  I heard archive.org have aprox 6%pa disk failure and wondered if you'd found that consistent?

thx

Okay, that's the worst cable management I've seen in years...

Remember this box does not run all the time, when it does run its very busy but the rest of the time it is relaxing so I am not sure our workload is real. I'll check wih the engineers on the disk failure thing. I do know that last time we got all new disks for it they came on pallets that we had all over the building for storage.

pdrg wrote:

with a 2500 disk array, how often do you have to swap drives out?  I heard archive.org have aprox 6%pa disk failure and wondered if you'd found that consistent?

thx

That's actually quite high for a system of this size. We have a 1000 disk array, and we see about 5 failures per year, less than 1%pa.

I suspect that the Superdome system pushes their disk array a bit more than ours gets pushed though.

edit: It's an HP array running a 12-node VMS cluster. Probably the same drives.

Yeah, I know the archive.org folks can't run all theirs in RAID or anything fancy, and those disks are HAMMERRED and some are getting old now...

Maybe when it comes to throwing out time, you could chuck them your spares (well not the duff ones, obviously) .  Although they curently run on the L-operating system I reckon it has to be a worthy cause!

it's entirely FABULOUS!! so good to see what it done and the process used for testing. When i saw the amount of hdrives & RAM, i reminded me this:
http://salle.ccs.usherbrooke.ca/webcast/index.html 

After seeing that i have a question, for SQL server 2000 were such important tests done or it's a new philosophy to ask lots of real users to work in real conditions to see the weekness of SQL S?

We did a lot of testing for SQL Server 2000 and the Suprtdome machine was originally used for testing and benchmarkign SQL Server 2000. We have come up with many new an innovative ways to test SQL Server 2005, but it some ways what is most impressive is not the actual testing, but the technique we use to fix the bugs, Soner and Edmund talk about this.

As Robert said, these are the last in the current round of SQl Server videos, we did some very early in the life of Channel9, you can find them here;

http://channel9.msdn.com/ShowPost.aspx?PostID=6089
http://channel9.msdn.com/ShowPost.aspx?PostID=6187
http://channel9.msdn.com/ShowPost.aspx?PostID=6308
http://channel9.msdn.com/ShowPost.aspx?PostID=6469
http://channel9.msdn.com/ShowPost.aspx?PostID=6857
http://channel9.msdn.com/ShowPost.aspx?PostID=6998
http://channel9.msdn.com/ShowPost.aspx?PostID=7098

EuanG wrote:

When we did the TPC benchmark with this machine it was $5.1 mill, you can get the disclosure doc here:

http://tpc.org/tpcc/results/tpcc_result_detail.asp?id=103082701

However this was 18 months ago, prices have come down for most stuff, but the box now handles 1TB of memory instead of 1/2 TB of memory. The procs are now 2GHz instead of 1.5Ghz but you get the idea

I may get one for Christmas... If I win the lottery

I'm going to wait for ebay.