I have been attempting to develop a useful and functional template for
database tracing/profiling that will enable me to collect metrics for
performance tuning. The database is used as an OLTP database as well as
running reports. Below is a list of my trace properties and data columns.
I would be interested to see other examples and strategies for the Profiler.
thanx
I am going over the output of a Profiler trace and I've found that the duration for many occurrences of EventClass 15 (Logout) is several seconds, up to a maximum of 20 seconds. That seems excessive just to complete a logout, so my question is, does the duration figure reflect only the time to complete the logout operation or does it include the total time that the connection has been active for?
Anyone can tell me how I can take out the events which produce SSMS? (open query windows create three rows). I don't see anythinq filter for SSMS in filter events of profiler. I'm doing at trace and this is not confortable.
im trying to understand how the profiler works. so i started 2profilers,one listen to another and I saw the profiler is running:exec sp_trace_create @P1 output, 1, NULL, NULL, NULLwhich means @tracefile = NULLso where from the profiler read the results?!
HiI want read a trace file generated by SQL Server 2005 througr SQLServer 2000.But fn_trace_gettable function in SQL 2000 does not recognize the fileas of proper format.If there is some other tool or utility available through which i canread the file generated by SQL Server 2005.Or if I can get the file format of the file then I will write my owntool.ThanksPushkar
If a table has a trigger on it, and I am profiling.. on StmtCompleted... no filters... all teh stored proc code comes up, but, is there any way at all to see the same for trigger statements? I want to trace thru the proc and thru all trigger code also. Any ideas on work-around to trace trigger code, if Profiler can't do it? Thanks, Bruce
We have two servers each running SQL7. I cannot run a trace on one server from the other. Whatever server name I enter in the drop down box, the trace only records activity on the server that profiler is running on. Even if I put a non-existent server name in the box (!), the trace accepts the name but still only runs on the host server.
Can anyone provide with an example of how to script a profiler trace to have the data wind up in a SQL Table. The scripting mechnism that comes with SQL Server will not allow you to put the results in a table.
I've set the Duration of my trace to "Greater than or Equal to: 1000". However when I start my trace the Duration column is now empty. Prior to the setting, there were values showing in this column. Any ideas on how to fix this?
Is there a way to setup a trace to show only direct TSQL statements triggered on my server? note I don't want to capture Procedure calls or the statements called within the procs.
Actually many people are firing direct SQL statements on server. And some are coming from entity framework as well. I just want to capture those.
I have discovered trace output in MSSQLDATAMSSQL.1MSSQLLOG that I have not kicked off. It is at various times and limited to 20MB. So that tells me a server event is kicking off a pre-defined trace. The trace contains mostly hash warnings and sort warnings. I have looked through my Agent Jobs, Agent Alerts, and perfmon and don't find anything that is set up to kick off a trace under a specified condition. I have checked the job activity, SQL error logs, SQL server logs, and the server's event viewer for any odd events or event times that correlate with the times of the traces. I have checked each database's sys.sql_modules for a definition containing '%sp_trace%'. Where else can I check to find what would be triggering these traces?
Our app logins don't have permissions high enough to run traces, I verified:
You do not have permission to run 'SP_TRACE_CREATE'
I am the DBA, not a .NET programmer -- so I am lacking experience if there's anything on the .NET side.
This is SQL 2005 64-bit running active/passive on a Win2003 clustered pair.
With SQL Server 2005 there is an option to grant a person access to Profiler for tracing SQL. This is done with the "GRANT ALTER TRACE" statement. The statement has to be executed at server level i.e. the master database.
The user in question only has access to certain databases on that server. The security problem that arises is that with the Profiler rights active, he can see the sql commands that are executed on the databases he has no rights for. Those SQL commands are executed by others users.
How do I configure security rules so that the person in question can use Profiler, but can only see the SQL statements that are executed on the databases he has the rights for? TIA!
How on earth can a Profile Trace be run where SSE 2005 is installed??? In the past, with MSDE, we always installed the 'tools' on a local workstation, so that we had EnterPrise Manager and its suite of tools...no problem. Yet, with Management Studio Express (err...Distress?), there's no way to do this! I've scoured the net, and I see threads where people have done it, yet, no one seems to be clear...including Microsoft...on how to obtain this MOST IMPORTANT of all tools for an SQL deployment.
I am attempting to create a new trace but I get the following error message: "failed to start a new trace".
I have been doing some digging and as I understand it, I had to find the directory Profiler uses for temporary files. So, I typed the following in the command window "SET TMP" and I received the following reply:
C:UsersRossAppDataLocalTemp
Now, according to the forum: [URL] ...
I am supposed to check that the system folder pointed to by the TMP environment variable exists and is not crammed with files.
Well, when I went to the directory C:UsersRossAppDataLocalTemp, it is indeed full of both files and directories. The size is 16.3 MB and has 133 files and 63 folders.
When I had a look at the Environment Variables window and chose TMP the value is "%USERPROFILE%AppDataLocalTemp" which according to my limited understanding is the equivalent to C:UsersRossAppDataLocalTemp.
So, what I am wondering is am I supposed to totally clear out this directory? I am not too keen on doing this because I don't want to stuff my PC up.
This is for SQL Server 2005 SP4 Build 5266. We have been having performance issues in production. There are tight deadlines to be met and it is important that they are solved promptly.
Yesterday we replicated the situation in the acceptance testing environment. The jobs take 8 hours to run and we started at 2:00 PM.
Just before the jobs ran I set up an SQL Server Profiler trace to catch processes with a duration of longer then 12 seconds. I set it to save the results to a database table.
Last night I checked the table at 5:00 PM and there were entries in the table. However, I could be mistaken.
At 9:00 PM I checked the table and it was empty.
This morning I arrived at work and checked SQL Server Profiler. The trace was running and within SQL Server Profiler, there are 100s of results. I stopped the trace. However, checking the table, it is empty.
I thought I would be able to save the trace results to a file. When I chose "Save As" from the file menu, all the options are greyed out (trace file, trace template, trace table, etc).
The results are there but there is no way of saving them and no way of exporting them. How could this have happened?
Is there a location, where SQL Server Profiler saves the results in a temporary space. I may be able to open them and retrieve them. How can I save the results? Why are all my options greyed out?
Set up a trace with the events RPC:Completed, SQL:BatchCompleted, SQL:BatchStarting, and SQL:StmtCompleted.
When I issue the statement: SELECT * FROM XyzView there is nothing captured in Profiler. If I script out the view and then execute the select statement that defines the view, it does show up in Profiler.
I've tried adding a lot of the other events, i.e. SP:StmtCompleted and the various other StmtStarting events and the trace still does not capture anything.
Am I capturing the wrong events or is this known behavior? My goal is to see what the overhead is for using a view versus persisting the results of the view as a table and referencing that instead. The view in question is against static data, joins 9 tables, and is referenced a lot.
I can use the stats generated when I execute the select that defines the view but I still find this to be curious behavior so I assume I'm doing something wrong.
I am trying to load all the MDX queries that run on a Analysis Server instance into a database for further analysis. A SQL Profiler is setup which captures the MDX queries, and when I am loading the Profiler info to database, some of the queries are not coming up in full length.The TextData field doestn't show full MDX query. When loading to the database, the field is next data type. Is there any workaround to get the complete MDX query?Â
Hi there - can anyone advise on the following issue. We have recently performed some server side tracing on a particular SQL instance over 24hr period. We are now attempting to load these into a database for analysis. Here lies the problem.
When we are loading the profiler trace files (one at a time) into the database the transaction log is growing at an excessive rate. Even though the database is in SIMPLE mode.
We are loading the traces using the command:
INSERT INTO sqlTableToLoad SELECT * FROM ::fn_trace_gettable('MytraceFileName', DEFAULT)
Can anyone advise how we could possibly get round this issue as we're running out of space due to the transaction log.
I've just joined a project for improving performance issue on MS-SqlServer(6.5). I was DBA Sybase a couple years ago I am unsure that MSSQL works in a manner as Sybase. I want to change some configuration options (sp_compile) and modify the table and indexes definition for improving the performance and avoid having deadlocks.
Is true that : When field, which has a default value assigned to NULL, is equal at a variable-length field and when you execute an update on it you will get 2 transactions : first a delete and then insert ? (On Sybase there is the concept of Index Covering to avoid that overhead what about Ms-Sqlserver 6.5).
To avoid having deadlock I want to use the FILLFACTOR I think this option is usefull only on cluster indexes, that's right ?
For huge tables the order of fields is important : first position put the primary key field after put the FK, secondary key, short type field (tinyint,int,datetime...) go on with the alpha fields (char) NOT NULL at last the NULL fields Is that design efficient for Ms-SQLserver ?
As a developer I have worked with all the tools provided by Sql server package (from the former 7 to the new 2005) and never felt that I shouldn´t use them because of performance issues! Now my systems collegue says to me that running a sql trace is very bad for the database. I have checked and for me doesnt seem that way.. Is this true? Running a trace can bring me a performance problem? When do you ususally use traces?
Hi, We have a poorly performing SQL 2000 db. i have just defragged ( the HD, not indexes, these are done daily via SQL Agent) the data files of our server and have not found any improvement in response. I have now got into using SQL profiler to analyse the server performance. in the results that the trace is returning there are some huge (REALLY BIG) values for the duration and cpu values but these rows have no textdata value returned (ie it is null)
why is this? for these rows, the reads and writes columns are also high.
if these rows are what is taking the cpu's time then how can i identify what the server is doing to make any changes?
any thoughts on what other values i might trace or what action i can take to find the slow down cause?
in performance manager the processors (dual Xeons) are rarely dropping below 60%.
Everywhere I read, it states that running SQL Profiler can affect performance of your SQL Server. My question is - how much of an impact will it really make? Will I see a 1% degredation in peformance? 5%? 50%? I haven't been able to find a good answer. We currently have SQL Profiler running all day long for almost 3 years, and the databases are still humming.
Is it the amount of data you are requesting from the trace that affects performance? There are some compliance tools out there (Idera Compliance Manager, IPLocks, etc) that run a profiler trace to get data. There are other DBAs in my organization who don't want to use them because "profiler traces will degrade my SQL Server performance". How true is this really.
Any help I can get would be extremely appreciated.
I am running a trace and one of the columns is start time I want to import corresponding performance log data in the profiler It is a new feature in 2005 however this option is disabled in the profiler this option is at File -> Import performance Data
I need to import data from a CSV file into a db I'm designing. I figured I'd use DTS (which I understand now uses VS in the form of a 'Business Intelligence' (BI) project). My problem is that my only choices for BI projects are: Analysis Service Project, Import Analysis Services 9.0 Database, Integration Services Project, Report Server Project Wizard, Report Model Project, Report Server Project. No "Data Transformation Project". I have SQL Server 2005 Developer's Edition. Might it be that DTS is not included in that version? If I can't use DTS, what choices do I have?
Hi! I'm new to SQL Server 7 and I hope I could get some help. I'd like to duplicate an empty database and use the duplicate for the transactions. I want the empty database to somehow act like a template so, whenever I want to, I could create a new instance of it. Problem is I don't know if it is possible to do that on SQL server (it's possible in Access). If it is possible, I hope any of you could tell me how. Thanks.
I'm having a hard time trying to figure out how to run the logtemp.sql script that is included with IIS4 for ODBC logging. Any help would be appreciated.
In Enterprise Manager when I bring up Trigger Properties there is an active button titled "Save as Template". But when I switch to an existing trigger the button is disabled. Oddly, the Help file for the dialog box doesn't even mention this button.
Anybody know what this does, how it works, or why it would be usefull? Any references to The Holy Book would be helpfull.
how can I specify the number of records that have to be sent back dynamically in an Xml Template?
I am currently busy developing a client-server application that pulls data from an Sql Server via Xml. If the client application has a bad connection to the server, the program has to get the data in small portions.
When I use:
SELECT TOP @amount * FROM employees
I get the error: syntax near '@amount'.
The query fails as wel when I tried to use:
SET ROWCOUNT @amount
SELECT * FROM employees
SET ROWCOUNT 0
This works fine in Sql analyser, but sql doesn't allow the integers to be parameters and casting them to int does not work either.
