September 22, 2005 at 12:43 pm
I am actually in a hardware testing group which is currently doing scalability testing for the SQL Servers used with our proprietary application. I am not looking at how well the application is written or how to optimize the databases on the servers. My question is strictly hardware related so I don't know if anyone can help me.
We have a pretty beefy server, for our purposes, it has a Pentium IV XEON processor that does hyperthreading (2.2Ghz), 1GB of RAM and three 36 GB drives in a RAID 5 configuration. When we track performance counters while running our application, we are consistently seeing Page Faults/sec over 20. However, our drive activity is low, our available MB of RAM is over 75% of what is available and the processor utilization is consistently lower than 40%.
I know there is a difference between hard page faults (i.e. going to disk) and soft page faults (i.e. located on physical RAM in a unidentified sector) but the articles and everything we see says that over 20 is a warning sign. I really don't see us having a problem but I wondered if anyone had examples of why we would see this value when hardware isn't being stressed.
Sorry for being so wordy.
September 23, 2005 at 5:16 am
I've seen issues with SQL Server max server memory setting being too low which can cause the behavior you describe. Look at memory properties in Enterprise Manager to determine max server mem properties.
September 23, 2005 at 7:03 am
Thanks for the answer. I had read that that setting could affect this. Unfortunately, we aren't using that setting. We are letting SQL grab as much memory as it needs since it is a dedicated SQL Server.
September 23, 2005 at 11:11 am
If this is a scalability test then should't you be using quad CPUs and 4 Gb of RAM ?
RegardsRudy KomacsarSenior Database Administrator"Ave Caesar! - Morituri te salutamus."
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply