Urgent: Some workstations faster than others... 
Author Message
 Urgent: Some workstations faster than others...

   Now I know that this is not supposed to happen in a client
server environment, but believe me it is occurring.

  Running M/S SQL 6.50.201 in a Windows N/T environment...

   What could possibly cause half the workstations on a given
floor to obtain markedly faster response times for a given
massive SQL stored procedure calculation than for other workstations
with identical configurations?  It is always the SAME stations
that take 25-30 minutes and the other stations always take about
3 minutes.   We are running directly from ISQL to obtain these
comparison times.

              Exec XYZ_Proc_Name "parm1", "parm2"

Possible causes we are examining:

1) Local SQL Trace Logging set on some client stations?  Where do
we verify this setting?

2) Protocols: Named Pipes vs. TCP/IP on different stations?

3) Different ISQL installation/configuration parameters?

   This is now running in a very large recent conversion to an
Ethernet 100Mhz backbone that was on a Token Ring previously.
The clients range from Pentium 120's to Pentium II 350's.

   The slow running stations sometimes appear to fill either
tempdb or the log file, while the others, of course do not.
But always the same stations have quick response times and the
same other group have the slow response times by 10 to 1 ratios.

Help.

Jack Daniels



Mon, 28 Jan 2002 03:00:00 GMT
 Urgent: Some workstations faster than others...

Jim,

Only thing I can think about (but that doesn't explain filling of tempdb) is
if you are running Win95 (98?) clients and try to use Named Pipes over IPX.
I believe that this is the config where you can have netlib problems. I'm
afraid that I don't recall the details, though...

--
Tibor Karaszi
MCDBA, MCSE, MCSD, MCT, SQL Server MVP
Cornerstone Sweden AB
Please reply to the newsgroup only, not by email.


Quote:
>    Now I know that this is not supposed to happen in a client
> server environment, but believe me it is occurring.

>   Running M/S SQL 6.50.201 in a Windows N/T environment...

>    What could possibly cause half the workstations on a given
> floor to obtain markedly faster response times for a given
> massive SQL stored procedure calculation than for other workstations
> with identical configurations?  It is always the SAME stations
> that take 25-30 minutes and the other stations always take about
> 3 minutes.   We are running directly from ISQL to obtain these
> comparison times.

>               Exec XYZ_Proc_Name "parm1", "parm2"

> Possible causes we are examining:

> 1) Local SQL Trace Logging set on some client stations?  Where do
> we verify this setting?

> 2) Protocols: Named Pipes vs. TCP/IP on different stations?

> 3) Different ISQL installation/configuration parameters?

>    This is now running in a very large recent conversion to an
> Ethernet 100Mhz backbone that was on a Token Ring previously.
> The clients range from Pentium 120's to Pentium II 350's.

>    The slow running stations sometimes appear to fill either
> tempdb or the log file, while the others, of course do not.
> But always the same stations have quick response times and the
> same other group have the slow response times by 10 to 1 ratios.

> Help.

> Jack Daniels




Mon, 28 Jan 2002 03:00:00 GMT
 Urgent: Some workstations faster than others...
Jim-

On the machines that are running slow, open the properties for the DSN they
are using to connect with.  Look for an option entitled "Create temporary
stored procedures for prepared SQL statements".  Make sure this is turned
OFF.

--

|  Tracy McKibben, MCSE
|  Eden Prairie, MN


Quote:
>    Now I know that this is not supposed to happen in a client
> server environment, but believe me it is occurring.

>   Running M/S SQL 6.50.201 in a Windows N/T environment...

>    What could possibly cause half the workstations on a given
> floor to obtain markedly faster response times for a given
> massive SQL stored procedure calculation than for other workstations
> with identical configurations?  It is always the SAME stations
> that take 25-30 minutes and the other stations always take about
> 3 minutes.   We are running directly from ISQL to obtain these
> comparison times.

>               Exec XYZ_Proc_Name "parm1", "parm2"

> Possible causes we are examining:

> 1) Local SQL Trace Logging set on some client stations?  Where do
> we verify this setting?

> 2) Protocols: Named Pipes vs. TCP/IP on different stations?

> 3) Different ISQL installation/configuration parameters?

>    This is now running in a very large recent conversion to an
> Ethernet 100Mhz backbone that was on a Token Ring previously.
> The clients range from Pentium 120's to Pentium II 350's.

>    The slow running stations sometimes appear to fill either
> tempdb or the log file, while the others, of course do not.
> But always the same stations have quick response times and the
> same other group have the slow response times by 10 to 1 ratios.

> Help.

> Jack Daniels




Sat, 09 Feb 2002 03:00:00 GMT
 
 [ 3 post ] 

 Relevant Pages 

1. Slow connection on some clients, fast on others

2. Slow connection on some clients, fast on others

3. Slow connection on some clients, fast on others :(

4. NT Workstation connects faster than Win 95

5. Urgent Help - Update on a table very slow if select is being run on another workstation

6. Its urgent...who can make it fast??

7. URGENT, Faster Insert or Update

8. SQL runs slower after moving to a new ( and faster ) machine - urgent

9. URGENT: need cobol compiler + tutorial FAST!!

10. Copying a database from one workstation to another workstation

11. SQL 6.5 Workstation Installation Problem on NT 4.0 Workstation

12. problems with installation of SQL Workstation 6.0 on NT Workstation 4.0


 
Powered by phpBB® Forum Software