Greg Moore wrote:
>
> Cary,
>
> In a TKPROF report, there is a small table at the top. It includes a column
> for CPU time and another for Elapsed Time.
>
> Suppose the total line shows CPU = 3.00 and elapsed = 5.00. You would think
> that the two seconds difference would be for waits, but th
Cary,
In a TKPROF report, there is a small table at the top. It includes a column
for CPU time and another for Elapsed Time.
Suppose the total line shows CPU = 3.00 and elapsed = 5.00. You would think
that the two seconds difference would be for waits, but that doesn't seem to
be true.
Since
0.010340 second ~ 10/1,000 second = 10ms (= 1/100 second)
If the number had been on the order of 0.001 (1ms) or less, it would've
looked more IPC-like.
Cary Millsap
Hotsos Enterprises, Ltd.
[EMAIL PROTECTED]
http://www.hotsos.com
-Original Message-
Sent: Wednesday, May 22, 2002 7:38
> the latency per call on the "SQL*Net message
> from client" event looked suspiciously LAN-like
> (order of 10ms), not IPC-like (order of 1ms or less)
> Oracle Kernel EventDurationCallsAvg
> -- -- --
> SQL
Hi Cary,
I also have a same problem but could not solve it yet
I have two m/cs and with identical DB layout etc and
running the
same pro*C code under same data volume.The top waits
are as follows.
1.oracle-8.0.5 ( 2 processor/RAM 512 MB)
Total run time 45 mins
Event
llsap"
otsos.com>cc:
Sent by: Subject: RE: I/O EVENTS
[EM
Cary,
This is greatness.
Jack
--- Cary Millsap <[EMAIL PROTECTED]> wrote:
> I have an example for you (Anjo, I hope you won't
> mind). A prospect we
> visited once upon a time had been fighting a
> performance problem with an
> Oracle Payroll program. They "knew" what their
> problem was: very
I have an example for you (Anjo, I hope you won't mind). A prospect we
visited once upon a time had been fighting a performance problem with an
Oracle Payroll program. They "knew" what their problem was: very
clearly, v$system_event was telling them that their overwhelmingly
dominant system perfor
oore <[EMAIL PROTECTED]>
Sent by: [EMAIL PROTECTED]
05/21/2002 12:13 AM
Please respond to ORACLE-L
To: Multiple recipients of list ORACLE-L <[EMAIL PROTECTED]>
cc:
Subject:Re: I/O EVENTS
> So just looking at v$system_event is dangerous.
> Looking
> So just looking at v$system_event is dangerous.
> Looking v$system_event and v$sysstat
> is much better but still not perfect. The third
> way is ... (mail me ;-))
Anjo,
Jared has an even hand on the tiller. Go ahead an post an informative
example of how the v$ views don't allow you to di
In short:
Oracle is either using CPU or is waiting on a resource.
If oracle is waiting on a resource, it could be an Oracle event
(v$system_event/v$session_event), BUT what about waits that are not registered there
. like
waiting for CPU, waiting for a memory page to be paged in ..
So j
Download the YAPP paper ;-)
Anjo.
John Kanagaraj wrote:
> Greg,
>
> > > Can I assume a i/o bottleneck from the following
> > >
> > > select * from v$system_event
> > > order by TIME_WAITED;
> >
> > No. Wait events may only make up a small amount of
> > processing that Oracle
> > is doing for
Hi Greg,
> Maybe not one, but what about two? At the same time v$system_event is
> checked a couple of times, so you can see a time slice,
> v$sysstat can be
> checked, focusing on CPU used by this session, parse time cpu
> and recursive
> cpu usage. One view gives wait time, one gives CPU ti
John,
> Hmm I wouldn't think so. If there were just _one_ overall view that I
> could check to determine an Oracle bottleneck, it would be this view.
Maybe not one, but what about two? At the same time v$system_event is
checked a couple of times, so you can see a time slice, v$sysstat can b
Greg,
> > Can I assume a i/o bottleneck from the following
> >
> > select * from v$system_event
> > order by TIME_WAITED;
>
> No. Wait events may only make up a small amount of
> processing that Oracle
> is doing for you.
Hmm I wouldn't think so. If there were just _one_ overall view that
Hi,
far better than getting individual descriptions is to research for yourself
(this isn't an RTFM message honestly).
Good sources of information for this subject are Anjo Kolk's excellent
document on wait events and enqueues
(http://www.dbatoolbox.com/WP2001/dbamisc/events.pdf) and the Oracle
do
> Can I assume a i/o bottleneck from the following
>
> select * from v$system_event
> order by TIME_WAITED;
No. Wait events may only make up a small amount of processing that Oracle
is doing for you.
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Greg Moore
INET:
17 matches
Mail list logo