| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: Statspack Report Help, Bursts of Activity, Amaze Your Friends ...
Thanks all, the issue turned out to be that the midtier has some sort of
throttle/sleep mechanism and it was in fact not doing anything. The java
threads seemed to be processing 1-2 connections at a time in a round robin
fashion and sleeping a lot. I don't think the developer has a good handle
on the innerworkings of the application.
-----Original Message-----
From: Jonathan Lewis [mailto:jonathan_at_jlcomp.demon.co.uk]
Sent: Saturday, February 14, 2004 3:42 AM
To: oracle-l_at_freelists.org
Subject: Re: Statspack Report Help, Bursts of Activity, Amaze Your
Friends...
I think John covered most of the points, but what you are saying, and what statspack is saying are totally consistent. Statspack reports about 3,200 seconds of CPU in half an hour, but you have 24 * 1800 CPU seconds available, which is about 43,000 seconds. Oracle is idle.
When Oracle is busy, it is thrashing the CPU
on logical I/O, tablescan row examination, and
execution pinned cursors. You may be able
to do something to reduce the CPU usage
(get rid of the tablescans for example), but
you won't make the 40 processes run faster
if they keep on stopping.
I think your problem is in the test harness that is running the front-end: maybe it just runs out of data; maybe the connections to the database break and it takes a long time for the front-end to notice.
One detail to note: the sql*net message from client totals about 11 * 30 minutes: this means your 40 sessions are not active for the whole time, or that wait time would be closer to 40 * 30 minutes if they were "active but doing nothing". They must be disconnecting.
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
The educated person is not the person
who can answer the questions, but the
person who can question the answers -- T. Schick Jr
Next public appearances:
March 2004 Hotsos Symposium - The Burden of Proof
March 2004 Charlotte NC OUG - CBO Tutorial
April 2004 Iceland
One-day tutorials:
http://www.jlcomp.demon.co.uk/tutorial.html
Three-day seminar:
see http://www.jlcomp.demon.co.uk/seminar.html
____UK___February
____UK___June
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
http://www.geocities.com/epost1/sp_1805_1808.txt
I am stumped here. I am benchmarking some software and this is what I am seeing. When running nmon on AIX 24 CPU box, all CPU's are idle, then we see bursts of activity on all CPU's to 100%. In the database I see 40 sessions coming in via JDBC Thin Client processing transactions, all idle, then all have serious bursts of activity. As you can see the most significant wait it CPU, but CPU's are Idle!!
Stumpppeddd...
nmon also shows that the disks are all idle, then experience bursts of activity. The disk system is an EMC, older model with not enough channels or controllers (not a disk guy), but I don't see any IO waits and I don't see anything in nmon.
If you can spot something obvious in my Statspack report that I am missing I would appreciate it.
Event 10046 trace on a session also doesn't show it having any significant waits.
My fear is we have a disk, OS or hardware configuration issue. I don't have a lot of access to the admins of this box.
Thanks!
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.htmlput 'unsubscribe' in the subject line.
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request_at_freelists.org
 |
 |