Logical connections very different from user connections

Hi,
We are administering a 64 bit sql 2005 instance that averages 400
transactions per second or so during business hours. The server is
performing quite well, but recently we've been seeing some performance
tick data that I don't understand.
In the SQLServer:General Statistics category, the server reports 270
logical connections. However, the User Connections perf ticks returns
491 connections. I am at a loss to explain the difference between the
two values. We consider 270 client connections to be an acceptable
number, but 491 is significantly higher than we would like to see.
I've searched for documentation explaining the difference between the
two perf ticks, but I haven't had any luck. I'm looking for a) some
information on what the difference between the two ticks is, and b)
some information that will help me determine if the discrepancy is
important.
Any advice or pointers as to where I could look next would be greatly
appreciated.
Sean Reilly
Software Architect
Point2 Technologies, Inc
(306) 955-1855Doesn't SQL Server group user connections based on certain conditions?
This would explain the difference.
Jim.
"Sean Reilly" wrote:

> Hi,
> We are administering a 64 bit sql 2005 instance that averages 400
> transactions per second or so during business hours. The server is
> performing quite well, but recently we've been seeing some performance
> tick data that I don't understand.
> In the SQLServer:General Statistics category, the server reports 270
> logical connections. However, the User Connections perf ticks returns
> 491 connections. I am at a loss to explain the difference between the
> two values. We consider 270 client connections to be an acceptable
> number, but 491 is significantly higher than we would like to see.
> I've searched for documentation explaining the difference between the
> two perf ticks, but I haven't had any luck. I'm looking for a) some
> information on what the difference between the two ticks is, and b)
> some information that will help me determine if the discrepancy is
> important.
> Any advice or pointers as to where I could look next would be greatly
> appreciated.
> Sean Reilly
> Software Architect
> Point2 Technologies, Inc
> (306) 955-1855
>|||Jim,
I hope that something like this does explain what I'm seeing. Do you
have any ideas as to what those conditions might be?
Thanks,
Sean|||From BOL:
"Connections can be shared among users. Users running OLE DB applications
need a connection for each open connection object, users running Open
Database Connectivity (ODBC) applications need a connection for each active
connection handle in the application, and users running DB-Library
applications need one connection for each process started that calls the
DB-Library dbopen function."
I'm not sure you have a lot of control over these features, but it would be
worth investigation "Connections" in any MS publications.
"Sean Reilly" wrote:

> Jim,
> I hope that something like this does explain what I'm seeing. Do you
> have any ideas as to what those conditions might be?
> Thanks,
> Sean
>|||Jim,
I'm pretty sure that the BOL entry you quoted refers to the client end
of the connection. I believe that Ole DB, ODBC and DB-Library
connections are all the same kind of connection from the server's
perspective.
A further data point for anyone who may be interested: when I use the
activity montior in SSMS, the number of connections reported matches
the smaller perf tick value (logical connections).