Discussion:
Stack trace interpretation
(too old to reply)
l***@bluewin.ch
2018-08-01 12:11:38 UTC
Permalink
Hi,
I like to know what the numbers marked in red mean.I guess it is something like wait time in milliseconds.
sgipcwWait()+880<-gipcWaitOsd()+396<-gipcInternalWait()+20104<-gipcWaitF()
I am also wondering about the moduls called.
E.g. gipcWaitF.
My guess is that g somewhat stands for cluster and ipc might refer to
network communication. These are waits related to the ocssd.
Regards
Lothar
Harel Safra
2018-08-01 12:16:11 UTC
Permalink
Hi,
I *believe *that the numbers are the code line offset from the beginning of
the function.
For the meaning of the function names, I highly recommend Frits Hoogland's
script here:
https://fritshoogland.wordpress.com/2017/10/16/oracle-c-functions-annotations/

Harel Safra
Post by l***@bluewin.ch
Hi,
I like to know what the numbers marked in red mean.I guess it is something
like wait time in milliseconds.
sgipcwWait()+880<-gipcWaitOsd()+396<-gipcInternalWait()+20104<-gipcWaitF()
I am also wondering about the moduls called.
E.g. gipcWaitF. My guess is that g somewhat stands for cluster and ipc
might refer to network communication. These are waits related to the ocssd.
Regards
Lothar
Stefan Koehler
2018-08-01 13:06:47 UTC
Permalink
Hello Lothar,
the numbers are described on slide 10 (http://www.soocs.de/public/talk/160616_DOAG_Regio_NUE_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PPT.pdf) or on page 5 (http://www.soocs.de/public/talk/151118_DOAG2015_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PDF.pdf).

Best Regards
Stefan Koehler

Independent Oracle performance consultant and researcher
Website: http://www.soocs.de
Post by l***@bluewin.ch
Hi,
 
I like to know what the numbers marked in red mean.I guess it is something like wait time in milliseconds.
sgipcwWait()+880<-gipcWaitOsd()+396<-gipcInternalWait()+20104<-gipcWaitF()
 
I am also wondering about the moduls called.
E.g. gipcWaitF. My guess is that g somewhat stands for cluster and ipc might refer to network communication. These are waits related to the ocssd.
 
Regards
 
Lothar
--
http://www.freelists.org/webpage/oracle-l
Jared Still
2018-10-08 23:30:20 UTC
Permalink
I've just spent the day playing with stack traces.

Stefan, your paper and presentation are outstanding!

Of particular interest was the wait_even["event name"] syntax that allows
dumping stack traces in a 10046 trace file.

Somehow I had missed learning about that, very cool.

Should anyone be interested in this topic, here is my reading list for the
day.

Note: of course there is Perl :)

https://manybutfinite.com/post/journey-to-the-stack/
Journey to the Stack, Part I

https://manybutfinite.com/post/anatomy-of-a-program-in-memory/
Anatomy of a Program in Memory

https://manybutfinite.com/post/how-the-kernel-manages-your-memory/
How The Kernel Manages Your Memory

If you are interested in dumping the values from little endian words
http://perldoc.perl.org/perlpacktut.html
Perl pack tutor

Leading up to:

Paper:
http://www.soocs.de/public/talk/151118_DOAG2015_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PDF.pdf
Identifying performance issues beyond the Oracle wait interface

Presentation:
http://www.soocs.de/public/talk/160616_DOAG_Regio_NUE_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PPT.pdf
Troubleshooting Oracle performance issues beyond the Oracle wait interface


Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist
Principal Consultant at Pythian
Pythian Blog http://www.pythian.com/blog/author/still/
Github: https://github.com/jkstill
Post by Stefan Koehler
Hello Lothar,
the numbers are described on slide 10 (
http://www.soocs.de/public/talk/160616_DOAG_Regio_NUE_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PPT.pdf)
or on page 5 (
http://www.soocs.de/public/talk/151118_DOAG2015_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PDF.pdf
).
Best Regards
Stefan Koehler
Independent Oracle performance consultant and researcher
Website: http://www.soocs.de
Post by l***@bluewin.ch
Hi,
I like to know what the numbers marked in red mean.I guess it is
something like wait time in milliseconds.
sgipcwWait()+880<-gipcWaitOsd()+396<-gipcInternalWait()+20104<-gipcWaitF()
Post by l***@bluewin.ch
I am also wondering about the moduls called.
E.g. gipcWaitF. My guess is that g somewhat stands for cluster and ipc
might refer to network communication. These are waits related to the ocssd.
Post by l***@bluewin.ch
Regards
Lothar
--
http://www.freelists.org/webpage/oracle-l
l***@bluewin.ch
2018-10-09 08:44:12 UTC
Permalink
Thanks Jared,
and also thanks to Stefan. It turned out we did not need to analyze the trace as detailed after all.
Well, it was quite nice to do something outside my usual borders. I am a software guy and OS is not usually my playground.
Regards
Lothar
----UrsprÃŒngliche Nachricht----
Von : ***@gmail.com
Datum : 09/10/2018 - 01:30 (GMT)
An : ***@soocs.de
Cc : oracle-***@freelists.org, ***@bluewin.ch
Betreff : Re: Stack trace interpretation
I've just spent the day playing with stack traces.
Stefan, your paper and presentation are outstanding!
Of particular interest was the wait_even["event name"] syntax that allows dumping stack traces in a 10046 trace file.
Somehow I had missed learning about that, very cool.
Should anyone be interested in this topic, here is my reading list for the day.
Note: of course there is Perl :)
https://manybutfinite.com/post/journey-to-the-stack/
Journey to the Stack, Part I
https://manybutfinite.com/post/anatomy-of-a-program-in-memory/
Anatomy of a Program in Memory
https://manybutfinite.com/post/how-the-kernel-manages-your-memory/
How The Kernel Manages Your Memory
If you are interested in dumping the values from little endian words
http://perldoc.perl.org/perlpacktut.html
Perl pack tutor
Leading up to:
Paper:
http://www.soocs.de/public/talk/151118_DOAG2015_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PDF.pdf
Identifying performance issues beyond the Oracle wait interface
Presentation:
http://www.soocs.de/public/talk/160616_DOAG_Regio_NUE_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PPT.pdf
Troubleshooting Oracle performance issues beyond the Oracle wait interface
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist
Principal Consultant at Pythian
Pythian Blog http://www.pythian.com/blog/author/still/
Github: https://github.com/jkstill
On Wed, Aug 1, 2018 at 6:08 AM Stefan Koehler <***@soocs.de> wrote:
Hello Lothar,
the numbers are described on slide 10 (http://www.soocs.de/public/talk/160616_DOAG_Regio_NUE_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PPT.pdf) or on page 5 (http://www.soocs.de/public/talk/151118_DOAG2015_Identifying_Performance_Issues_Beyond_The_Oracle_Wait_Interface_PDF.pdf).
Best Regards
Stefan Koehler
Independent Oracle performance consultant and researcher
Website: http://www.soocs.de
Post by l***@bluewin.ch
Hi,
I like to know what the numbers marked in red mean.I guess it is something like wait time in milliseconds.
sgipcwWait()+880<-gipcWaitOsd()+396<-gipcInternalWait()+20104<-gipcWaitF()
I am also wondering about the moduls called.
E.g. gipcWaitF. My guess is that g somewhat stands for cluster and ipc might refer to network communication. These are waits related to the ocssd.
Regards
Lothar
--
http://www.freelists.org/webpage/oracle-l

Loading...