1
Vote

Using existing TraceTables

description

Hi all,
I m planning to run multiple instances of AsTrace NT Service ( on more than 5-10 machines ) to collate the Traces from multiple SSAS. But everytime I install the service onto new machine, the TraceTable ( say name : TraceTable1 ) gets reset [ all the old entries are deleted ].
 
Is there a way, I can use the same TraceTable for output from multiple machines not affecting the entries present ?

comments

furmangg wrote Jan 12, 2012 at 3:50 PM

There's a setting in the latest build of ASTrace which is:
        <setting name="PreserveHistory" serializeAs="String">
            <value>True</value>
        </setting>
If PreserveHistory is set to true, then when each service starts, it will dump all the rows from the main trace table into an archive table.

If you run multiple services at the same time with PreserveHistory=true, this might work. I'd be concerned what would happen if two services pointing at the same table started at the same time, though.

I would just recommend each service point at a separate SQL destination table. Then put a view over the top of the 5-10 tables (plus the 5-10 archive tables) and UNION ALL them together.

furmangg wrote Jan 12, 2012 at 3:51 PM

I should say that ASTrace uses Profiler APIs, and as far as I now, Profiler when it writes to a table will clear out the table when it starts. That's why we added the PreserveHistory setting so we could save the rows before starting the Profiler trace in ASTrace. Make sense?

wrote Feb 21, 2013 at 10:34 PM