RE: FW: Strange Elapsed In Raw Trace

From: Joze Senegacnik <JozeS_at_hermes-plus.si>
Date: Thu, 14 Oct 2004 23:49:44 +0200
Message-ID: <5844F9B4711770478E44A075199030950A5162@simail01.adriatic.snt.eu>

Hi!

Some days ago I got a strange number for ela in one trace file. Look the = line after the time stamp.=20
.....
WAIT #4: nam=3D'db file sequential read' ela=3D 0 p1=3D44 p2=3D86649 = p3=3D1

WAIT #4: nam=3D'db file parallel read' ela=3D 3 p1=3D5 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 2 p1=3D6 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 2 p1=3D4 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 1 p1=3D3 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 2 p1=3D5 p2=3D11 p3=3D11

*** 2004-10-06 14:20:52.293
WAIT #4: nam=3D'db file parallel read' ela=3D -1962624467 p1=3D7 p2=3D11 = p3=3D11

WAIT #4: nam=3D'db file parallel read' ela=3D 3 p1=3D6 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 1 p1=3D3 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 2 p1=3D5 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 2 p1=3D4 p2=3D11 p3=3D11
WAIT #4: nam=3D'db file parallel read' ela=3D 1 p1=3D4 p2=3D7 p3=3D7
WAIT #4: nam=3D'db file sequential read' ela=3D 0 p1=3D93 p2=3D122235 =

p3=3D1
WAIT #4: nam=3D'db file parallel read' ela=3D 4 p1=3D6 p2=3D11 p3=3D11 ....
FETCH =
#4:c=3D513,e=3D16813,p=3D64290,cr=3D65074,cu=3D0,mis=3D0,r=3D1,dep=3D0,og= =3D4,tim=3D2332345878=20
...

Interesting is that the e of the first FETCH following the big number of = wait events is "normal". I am getting huge numbers for time_waited in = v$session_event and v$system_event that can't be anyhow justified.

The system is AIX 32 bit with 32 bit version of Oracle 8.1.7.

Regards, Joze

-----Original Message-----
From: oracle-l-bounce_at_freelists.org =
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of rjamya Sent: Thursday, October 14, 2004 9:43 PM To: cary.millsap_at_hotsos.com
Cc: Oracle-L (E-mail)
Subject: Re: FW: Strange Elapsed In Raw Trace

Cary,

you might also be interested to know that similar problem exists in AIX, it is a OS level problem, IBM doesn't want to (or can't) fix it. Oracle is dragging its feet in implementing workaround.

I logged the bug many moons ago and still am following it.

Raj

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

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

Received on Thu Oct 14 2004 - 16:45:20 CDT