March 11, 2008 at 7:32 am
Seeing high level of page faults/sec with little or no pages/sec. Page faults averaging @ 2000/sec. This is a SQL 2005 Enterprise Clustered environment, w/ 16Gb RAM, 8 proc/dual 64 bit system.
We've seen similar behavior on another instance running on a separate node on this cluster as well.
Any suggestions on what I should be looking at/for?
Thanks!
- Tim Ford, SQL Server MVPhttp://www.sqlcruise.comhttp://www.thesqlagentman.com http://www.linkedin.com/in/timothyford
March 11, 2008 at 8:59 pm
High Page Fault/sec with no Pages/Sec just means that you are getting a lot of Soft Faults; that is, that the page faults are being satisfied with pages already in memory, without having to read (or write) them to or from the disk.
Generally, that is a good thing, so unless you are actually, experiencing slow response times or run times, I would not worry about it.
[font="Times New Roman"]-- RBarryYoung[/font], [font="Times New Roman"] (302)375-0451[/font] blog: MovingSQL.com, Twitter: @RBarryYoung[font="Arial Black"]
Proactive Performance Solutions, Inc. [/font][font="Verdana"] "Performance is our middle name."[/font]
March 12, 2008 at 10:01 am
have a read of my research and testing - it may help http://sqlblogcasts.com/blogs/grumpyolddba/archive/2008/01/06/configuring-windows-2003-x64-for-sql-server.aspx
I'm also currently working with a very busy sql2k setup and that's been interesting to say the least - I think what is obvious is that you probably need to allocate far more memory to o/s and application/binaries than is generally stated. My sql2k system has been absolutley hammered but I've had almost zero paging.
[font="Comic Sans MS"]The GrumpyOldDBA[/font]
www.grumpyolddba.co.uk
http://sqlblogcasts.com/blogs/grumpyolddba/
March 12, 2008 at 11:04 am
Interestingly enough I read the blog yesterday Colin. Very informative. Thank you very much.
My concerns with not allowing SQL to manage memory is that this is a clustered environment (multi-instance). If one instance was to fail over and have two instances then running on a single node there would be issues, correct?
The other issue is that one of the 3 nodes on this cluster is 8 procs, dual core, the other 2 nodes are 4 procs, dual core. I am not quite sure how to force memory so that any of these instances would play friendly if moved from node to node.
Suggestions?
- Tim Ford, SQL Server MVPhttp://www.sqlcruise.comhttp://www.thesqlagentman.com http://www.linkedin.com/in/timothyford
April 6, 2011 at 9:42 am
I know this is an old post but wanted to ask if any of you knew which counters should be added to perfmon to monitor a SQL 2008 SP2 Server that's experiencing slowness, and which counters are good to monitor performance of a SQL server's health overall.
Any takers?
______________________________________________________________________________Never argue with an idiot; Theyll drag you down to their level and beat you with experience
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply