Glenn Berry's SQL Server Performance - All Comments

re: How to Easily Log SQL Server 2008 Instance Level Metrics

Glenn Berry — Wed, 10 Mar 2010 18:20:12 GMT

Well, we add one row per minute. Each row is a bigint, a datetime, and five ints. I am also using Page data compression on the table.

The amount of size used by the table is pretty trival. I have never needed to trim it down so far.

re: How to Easily Log SQL Server 2008 Instance Level Metrics

Steve Jones — Wed, 10 Mar 2010 17:35:55 GMT

How much data is once a minute? Do you trim this down over time?

SQL Server Central Debt on Me

SQL Server Central Debt on Me — Tue, 09 Mar 2010 16:59:17 GMT

Pingback from SQL Server Central Debt on Me

re: Easy Ways to Detect I/O Pressure in SQL Server 2008

Jason Brimhall — Tue, 09 Mar 2010 16:53:54 GMT

Nice post Glenn.

re: Weekly Electrical Usage from Google PowerMeter

Glenn Berry — Mon, 01 Mar 2010 18:08:11 GMT

I have read that an oversized furnance cycles too often (which wastes energy) in a few places:

www.furnacecompare.com/buying_a_furnace.html

"An oversized furnace (one whose capacity is too large) costs substantially more to operate than a correctly sized furnace. This is because oversized furnaces typically cycle on and off more often than correctly sized furnaces, and cycling on and off is more expensive than running at steady state."

re: Weekly Electrical Usage from Google PowerMeter

Steve Jones — Mon, 01 Mar 2010 17:44:39 GMT

Interesting, why does the sawtooth show that the furnace is too large?

re: Rocky Mountain Tech Trifecta Impressions

Steve Jones — Mon, 01 Mar 2010 17:39:33 GMT

Sorry I missed dinner with everyone, but great to see you as well.

re: Dr. DMV Script for Rocky Mountain Tech Trifecta

Jason Brimhall — Sat, 27 Feb 2010 05:46:05 GMT

Thanks for sharing

???????????? ???????????????????? ???????????? » ?????????? ???????? » ???????????? ???????????????? «Power teister

Sat, 27 Feb 2010 03:45:11 GMT

Pingback from ???????????? ???????????????????? ???????????? » ?????????? ???????? » ???????????? ???????????????? «Power teister

re: Experiments with Data Compression in SQL Server 2008

Nicholas Cain — Fri, 26 Feb 2010 17:26:42 GMT

Talltop: there are algorithms that look for patterns in the data that they can use to perform the compression. When you enable TDE and encrypt the data you eliminate any patterns and so you won't get backup compression.

re: Experiments with Data Compression in SQL Server 2008

Glenn Berry — Fri, 26 Feb 2010 13:25:00 GMT

In these experiments, the query time was reduced in Query 2 (when the clustered index was compressed) compared to Query 1. CPU utilization did not go up by much, but I dis not measure it precisely, I was just watching in Task Manager.

I will re-run the experiment and measure it in PerfMon this time.

re: Experiments with Data Compression in SQL Server 2008

talltop — Fri, 26 Feb 2010 11:10:08 GMT

Good read. Also, has anyone noticed that TDE db encryption almost nullifies the compression on backups?

re: Micron Real SSD C300

Old Dirty Bastard -- *ODB* — Thu, 25 Feb 2010 15:18:21 GMT

Guys, I have found one single whitepaper on SSD's & OLTP systems. Our infrastructure guys want to try them. Can you list some real-life SSD performance tests using SSD's on OLTP systems? A comparison to 15k SCSI drives would be most useful as we are a dell shop. thx.

re: Experiments with Data Compression in SQL Server 2008

Steve Jones — Mon, 22 Feb 2010 15:40:09 GMT

That's interesting. Did it change query time substantially? Did CPU raise up?

cpanvote: a Perl Mini-Project | The Pythian Blog | Perl WebDev Insider

cpanvote: a Perl Mini-Project | The Pythian Blog | Perl WebDev Insider — Sat, 20 Feb 2010 03:50:24 GMT

Pingback from cpanvote: a Perl Mini-Project | The Pythian Blog | Perl WebDev Insider