Re: Fragmentation with restore multiple channels

It would be most effective to just ignore nonsense like "fragmentation" and "channels" and "data bits fluxing through the multiplexer" and instead work from facts.

This is a tuning exercise, just like any other. If you have a particular program/session which is running more slowly and you can initiate a 10046 extended SQL trace, then please do so, and then please aggregate the trace file using TKPROF, TRCANLZR, or the Hotsos/Method-R Profiler to see *exactly* what is happening. If you're permitted, feel free to share small portions of that report with the ORACLE-L list to help obtain an interpretation. If you know a specific program/session you'd like to trace but have never done it before or just don't know how, ask the list.

If you have a more general problem and cannot pinpoint problems with a single process immediately, then let's instead start from a higher level with an AWR report covering the time period when the problem was observed. Although the HTML version of the standard AWR report is easier to use, if you would like to share it with the ORACLE-L list, please obtain a text version of AWR and share three things: 1) the very top section of the report showing "snapshot" times along with "Elapsed" and "DB Time" minutes, plus 2) the "Top Five Timed Events" section, and then 3) the "Top SQL Ordered by Elapsed Time" section. These will make for a very good starting point for analysis.

Don't let anyone bamboozle with BS and wild guesses. Even someone who really does know what they're talking about can be wrong if they're just guessing. Use the tools that everyone discusses here on ORACLE-L everyday, and use them to counter guesses with facts.

Hope this helps...

Good luck!

-Tim

On 9/3/2012 8:34 AM, Alex Cwb wrote:
> Hi people,
>
> Recently I am facing a problem with fragmentation. A supplier is
> arguing that the way we did a restore using RMAN with multiple
> channels is causing many fragmentation on the database, and
> consequently their application on my company is running slowly.
>
> The restore has been done using 6 channels, and I guess that is not
> true. The database is being used for testing purposes only, and during
> this phase it's not happening updates.
>
> Anyways, I did some research and I couldnt find a resonable answer. So
> I am in doubt if restore in multiple channels really causes
> fragmentation.
> Does someone have any experience on that??
>
> Thank you for sharing.
> Alex
> --
> http://www.freelists.org/webpage/oracle-l
>

--
http://www.freelists.org/webpage/oracle-l