The contents of your tracefile will become more meanigfull and not unimportant small …

Filter out ‘internal’ sql server proceses

Filter on ‘ApplicationName’

-- --
-- ApplicationName
 
-- SQL Profiler sessions
exec sp_trace_setfilter @TraceID, 10, 0, 7, N'SQL Profiler'
 
-- SQL executed by the SQL Agent
exec sp_trace_setfilter @TraceID, 10, 0, 7, N'SQLAgent%'
exec sp_trace_setfilter @TraceID, 10, 0, 7, N'SQLDMO_1%'
 

Filter on ‘LoginName’

-- --
-- LoginName
 
exec sp_trace_setfilter @TraceID, 11, 0, 7, N'remivisser'
exec sp_trace_setfilter @TraceID, 11, 0, 7, N'<username>'

Display running traces

When ‘c2 audit mode’ is enabled you will see a trace running at all times. Whenever the server can’t write to it’s trace file it will shut itself down. (You need to restart with the -f switch to fix your problem).

Example below first displays all running traces and next aborts the trace ‘facilitating’ the ‘c2 audit mode’ option. The following functions are used, ::fn_trace_getinfo (displays information about runing traces) and sp_trace_setstatus (used to alter the status of a trace).

1> exec sp_configure "c2 audit mode"
2> go
 name                                minimum     maximum     config_value run_value
 ----------------------------------- ----------- ----------- ------------ -----------
 c2 audit mode                                 0           1            1           1
 
1> select traceid, property, cast(value as char(64)) as value from ::fn_trace_getinfo(0)
2> go
 traceid     property    value
 ----------- ----------- ------------------------------------------------------------------------
           1           1 6
           1           2 C:\\Program Files\\Microsoft SQL Server\\MSSQL\\data\\audittrace_200609061836
           1           3 200
           1           4 NULL
           1           5 1
 
(5 rows affected)

Alter status of a running trace

To alter the status of a trace, eg starting/stopping it you use the sp_trace_setstatus function. This function takes two arguments; @traceid – the id of the trace and @status – the status of the trace, where status is one of the three options below;

• 0 – Stops the specified trace.
• 1 – Starts the specified trace.
• 2 – Closes the specified trace and deletes its definition from the server.

Now watch the results of aborting the trace in an MS SQLServer instance that is running in ‘c2 audit mode’;

1> exec sp_trace_setstatus 1, 0
2> go
[DBNETLIB]ConnectionCheckForData (CheckforData()).
[DBNETLIB]General network error. Check your network documentation.
$

Do not try this at home, the server shuts itself down.

The shutdown seems rather rude as I can not find any information in the mssqlserver logfile as to why it aborted.

See this article by Brian Moran where I got the inspiration to try this one out …

Handy Scripts

Some handy scripts :

Script to display running traces

select traceid,
       case property
         when 1 then 'Trace Options'
         when 2 then 'FileName'
         when 3 then 'MaxSize'
         when 4 then 'StopTime'
         when 5 then 'Current Trace status'
       end property,
       cast(value as char(64)) as value
  from ::fn_trace_getinfo( 0)
go
 

Script below generates commands to stop and close all runing traces

select 'exec sp_trace_setstatus ' + cast( traceid as varchar(8)) + ', 0' + char(10) +
       'exec sp_trace_setstatus ' + cast( traceid as varchar(8)) + ', 2'
  from ::fn_trace_getinfo(0)
 where property = 1
go
 

Retrieve trace file data using ::fn_trace_gettable function.

select * 
  from ::fn_trace_gettable( 'E:\\mssql\\admin\\iNYC01\\audit\\AUD_TSQLSTMT_2006-11-14.trc', default)
go
 
select * 
  into t_trace 
  from ::fn_trace_gettable( 'E:\\mssql\\admin\\iNYC01\\audit\\AUD_TSQLSTMT_2006-11-14.trc', default)
go
 

URI’s from the documentation

• Monitoring with SQL Profiler
• SQL Profiler Event Classes
• SQL Profiler Data Columns