Re: Where are my trace files going?

Here's something that may or may not be happening. It's something I ran into when working with traces. Have you mv'd or rm'd any of your trace files? The reason I ask, is because while doing some tracing of a certain session on Solaris recently, I began the tracing, turned it off, and then mv'd the file somewhere's so I could analyze it. As this was a long-running session (an XML gateway process), I wanted to try and trace another run, so I turned tracing back on, expecting Oracle to generate a new trace file for me, but no such luck. After a few frustrating hours of investigation, I finally realized what was happening. The OS process keeps the filehandle of the trace file open, even after you tell Oracle to turn off tracing (or at least that's what I gathered), so once I mv'd the file, all bets were off as to where Oracle would write to once I started tracing the process again. A combination of truss'ing the process and inspecting /proc info told me that Oracle had the root!  dir of the filesystem on which the trace file had been now open as the trace filehandle, which was very confusing. It wasn't until I restarted the session, and got a new OS pid, that I got another trace file generated.

At any rate, this may or may not be what you're up against, but from what I found, if you mv/rm the trace file associated with a given PID, you'll probably need to get a new Oracle session / OS pid going to get a new tracefile.

HTH,

• Dan Hanks

On Mon, 1 Dec 2003, Thomas A. La Porte wrote:

> Environment: Oracle 8.1.7.4 on RedHat AS2.1
>
>
> I'm in the process of preparing to convert a 90M row table from
> heap-organized to index organized. I think I've pretty well got a
> formula for doing the actual table conversion. It's a very basic
> table, four NUMBER columns, with a PK on the first two. I'm
> planning to extract the data to a sorted, comma-delimited flat
> file, then using SQL Loader with direct path to do the import.
> On a 2.8GHz 2-way Linux box, that process is taking me about 30
> minutes.
>
> Building two additional indexes on that table after the data has
> been loaded is taking anywhere from 30 minutes up to three hours.
> I'm trying to determine why there is a wide variation in the
> performance of the index build operation by tracing the session,
> however, I can't seem to generate a trace file! I can run the
> 'alter session' to get a 100046 trace, but no file is created in
> my user_dump_dest (nor, for that matter, is it being created in
> my background_dump_dest, core_dump_dest, or any place else on the
> local machine as best as I can tell).
>
> I've run these traces before in other instances on the same
> machine, and I don't believe that there is a file ownership or
> permissions problem anywhere in the mix. Does anyone have any
> thoughts on where my trace file is going, if it is going
> anywhere? Or how to determine why I'm not generating a trace
> file?
>
> Any thoughts or pointers are greatly appreciated.
>
> -- Tom
>
> Thomas A. La Porte, DreamWorks SKG
> <mailto:tlaporte_at_anim.dreamworks.com>
>
>
>

-- 
========================================================================
   Daniel Hanks - Systems/Database Administrator
   About Inc., Web Services Division
========================================================================
-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.net
-- 
Author: Daniel Hanks
  INET: hanksdc_at_about-inc.com

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).