Re: Query ASH for Undo Blocks

Stefan,

Following some of the examples of the newer trace mechanism that you and Tanel have published - can you think up a specification for tracing that would allow the process stats to be dumped to trace for a given SQL_ID at the start of call and end of fetch? I think there's a PROCSTAT call from oradebug, but would have to do some research to see if I could get the effect from "set events".

Regards
Jonathan Lewis

On Mon, 13 Jun 2022 at 13:22, Stefan Koehler <contact_at_soocs.de> wrote:

> Hello Lothar,
> you can trace logical I/Os (events 10200 and 10201) or use already
> available OS tracing scripts for that, e.g. ...
>
> * SystemTap:
> https://db-blog.web.cern.ch/blog/luca-canali/2014-12-life-oracle-io-tracing-logical-and-physical-io-systemtap
> * DTrace:
> https://alexanderanokhin.wordpress.com/2011/11/13/dynamic-tracing-of-oracle-logical-io/)
>
>
> ... for the latter you need to use web.archive.org.
>
> Best Regards
> Stefan Koehler
>
> Independent Oracle performance consultant and researcher
> Website: http://www.soocs.de
> Twitter: _at_OracleSK
>
> > Lothar Flatz <l.flatz_at_bluewin.ch> hat am 13.06.2022 14:04 geschrieben:
> >
> > Hi Guys,
> >
> > thanks a lot Lok, Mohamed and Nenad. Unfortunately that index is almost
> all cached.
> > Thus I have only 2 physical read both from undo....
> > I can see some evidence that we have a undo issue, but can not proof it.
> >
> > Thanks
> >
> > Lothar
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

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