Hi all,
On SQL Server 7 SP4.
3 tables with about 100 000 rows each one, frequently updated.
1 clustered index on primary key, and 2 non clustered index on columns used by JOIN and by queries.
Fillfactor for index : 90.
PAD_INDEX : not activated.
I have some deadlocks problems, probably due to index pages and key locking (as precised in error log). Is it possible that the cause of that is splitting index pages while UPDATE ?
May I activate PAD_INDEX to solve part of this problem ?
Thanks for any helpIt's a performance issue thing, I don't think that's your problem
http://www.sql-server-performance.com/oltp_performance.asp|||Well, the nature of deadlocks has NOTHING to do with performance, thus cannot be fixed by recommendations given in the link. The only applicable advice which is not even in the section implied is to keep your transactions short. This will minimize the possibility of deadlocks, but will not eliminate them. As you all know, it deals with object access order, and should be addressed accordingly, by examining all possible transactions involving the same set of objects in reverse order.|||Thanks all, I'm going to have a look at it.|||pad index is really not neccesary. i have chased this guy down many roads
i even had extended conversations with kimberly tripp and kalen delaney on this and everyone i have talked to has told me that pad index is just not gonna provide any real advantage.
Showing posts with label clustered. Show all posts
Showing posts with label clustered. Show all posts
Wednesday, March 28, 2012
Monday, March 26, 2012
PackageForTheWEB error
I'm having a package for the WEB error when installing
MSSQL 2000 in a clustered windows 2000 advance server.
I have observe the task manager what process was invoke
when the error appeared and it is MSSEARCH.EXE.
What can I do to complete my installation?
Thanks
Janz,
More info is necessary in order to troubleshoot this issue... Could you post
the full output of -- SELECT @.@.version -- Also, attach the file
SearchSetup.log that is usually located at C:\WINNT\ and as well as
sqlstp.log usually located at C:\WINNT\Temp\ and post any errors or issues
related to MSSearch in either of these logs. You can also attach these
files to your post and I'll review them as well. Is your clustered
environment an active/active or passive/active? Which node is this error
occurring on? If you fallover, does the error re-occur?
Note, you should also review your server's Application event log for any
"Microsoft Search" or MssCi source events, errors or warnings as this too
would be helpful info.
Regards,
John
"Janz" <anonymous@.discussions.microsoft.com> wrote in message
news:0ff101c48bd8$577838f0$a601280a@.phx.gbl...
> I'm having a package for the WEB error when installing
> MSSQL 2000 in a clustered windows 2000 advance server.
> I have observe the task manager what process was invoke
> when the error appeared and it is MSSEARCH.EXE.
> What can I do to complete my installation?
>
> Thanks
MSSQL 2000 in a clustered windows 2000 advance server.
I have observe the task manager what process was invoke
when the error appeared and it is MSSEARCH.EXE.
What can I do to complete my installation?
Thanks
Janz,
More info is necessary in order to troubleshoot this issue... Could you post
the full output of -- SELECT @.@.version -- Also, attach the file
SearchSetup.log that is usually located at C:\WINNT\ and as well as
sqlstp.log usually located at C:\WINNT\Temp\ and post any errors or issues
related to MSSearch in either of these logs. You can also attach these
files to your post and I'll review them as well. Is your clustered
environment an active/active or passive/active? Which node is this error
occurring on? If you fallover, does the error re-occur?
Note, you should also review your server's Application event log for any
"Microsoft Search" or MssCi source events, errors or warnings as this too
would be helpful info.
Regards,
John
"Janz" <anonymous@.discussions.microsoft.com> wrote in message
news:0ff101c48bd8$577838f0$a601280a@.phx.gbl...
> I'm having a package for the WEB error when installing
> MSSQL 2000 in a clustered windows 2000 advance server.
> I have observe the task manager what process was invoke
> when the error appeared and it is MSSEARCH.EXE.
> What can I do to complete my installation?
>
> Thanks
Friday, March 9, 2012
Ownership chain question
HI,
MS SQL 7.0 on NT 4 (clustered)
Our developers use Infomaker to create many end-user's reports. They are
used to creating many Stored Proc and many "intermediate" tables instead of
creating "pure" Infomaker Reports. Report's components are often modified
directly on Production database during office hours by thoses developpers.
All of our end-user's reports are going to be rebuilt with Crystal Report.
Is it a good idea to ask our developpers to locate thoses end-user's reports
components into another databases instead of mixing DATA and REPORTS
component into the same Production Database ? (I know that we must take care
of Ownership chain)
Thank you
Danny"Danny Presse" <dpresse@.congresmtl.com> wrote in message
news:eA6M9CfXDHA.2476@.tk2msftngp13.phx.gbl...
> HI,
> MS SQL 7.0 on NT 4 (clustered)
> Our developers use Infomaker to create many end-user's reports. They are
> used to creating many Stored Proc and many "intermediate" tables instead
of
> creating "pure" Infomaker Reports. Report's components are often modified
> directly on Production database during office hours by thoses developpers.
> All of our end-user's reports are going to be rebuilt with Crystal Report.
> Is it a good idea to ask our developpers to locate thoses end-user's
reports
> components into another databases instead of mixing DATA and REPORTS
> component into the same Production Database ? (I know that we must take
care
> of Ownership chain)
Yes. Definitely.
Give them guest access only to the production database (you need this for
cross-database views to work).
Create views in their database that proxy over to the production database.
This keeps them from having to bind to the database name of the production
database.
David
MS SQL 7.0 on NT 4 (clustered)
Our developers use Infomaker to create many end-user's reports. They are
used to creating many Stored Proc and many "intermediate" tables instead of
creating "pure" Infomaker Reports. Report's components are often modified
directly on Production database during office hours by thoses developpers.
All of our end-user's reports are going to be rebuilt with Crystal Report.
Is it a good idea to ask our developpers to locate thoses end-user's reports
components into another databases instead of mixing DATA and REPORTS
component into the same Production Database ? (I know that we must take care
of Ownership chain)
Thank you
Danny"Danny Presse" <dpresse@.congresmtl.com> wrote in message
news:eA6M9CfXDHA.2476@.tk2msftngp13.phx.gbl...
> HI,
> MS SQL 7.0 on NT 4 (clustered)
> Our developers use Infomaker to create many end-user's reports. They are
> used to creating many Stored Proc and many "intermediate" tables instead
of
> creating "pure" Infomaker Reports. Report's components are often modified
> directly on Production database during office hours by thoses developpers.
> All of our end-user's reports are going to be rebuilt with Crystal Report.
> Is it a good idea to ask our developpers to locate thoses end-user's
reports
> components into another databases instead of mixing DATA and REPORTS
> component into the same Production Database ? (I know that we must take
care
> of Ownership chain)
Yes. Definitely.
Give them guest access only to the production database (you need this for
cross-database views to work).
Create views in their database that proxy over to the production database.
This keeps them from having to bind to the database name of the production
database.
David
Subscribe to:
Posts (Atom)