Oracle-L: RE: Excessive library cache latch contention

From: Cary Millsap <cary.millsap_at_hotsos.com>
Date: Fri, 31 Jan 2003 14:35:08 -0800
Message-ID: <F001.00541241.20030131143508@fatcity.com>

Just to clarify...

If you upload something to our site, and you haven't purchased a license to run the Hotsos Profiler, then make sure you send a note to matt.seaton_at_hotsos.com, letting him know that you're interested in seeing a demo profile of your data.

We're glad to do demos of our software, especially if your trace file is particularly interesting. Bear with us, though: if everyone responds at once, then there'll be some queueing delay for assistance :).

Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com

Upcoming events:

- 2003 Hotsos Symposium, Feb 9-12 Dallas
- RMOUG Training Days 2003, Mar 5-6 Denver
- Hotsos Clinic 101, Mar 25-27 London

-----Original Message-----

Michael
Sent: Friday, January 31, 2003 3:13 PM
To: Multiple recipients of list ORACLE-L

Correction on this ... I should have said www.hotsos.com would have some additional insights on dealing with 10046 data for you if you are interested.

-----Original Message-----

Sent: Friday, January 31, 2003 10:57 AM
To: Multiple recipients of list ORACLE-L

Harvey,

Personally, I would go right to the user(s) who has complained about the slowdown and have them run the application at the peak hour/period where things seem very slow.

Set a 10046 event level 8 trace on that session(s) after they log on and then take a look at the trace file in the udump area after they are finished executing the queries.

There should be some strong clues in there about why the session is the waiting.

My bet is on some poorly written SQL and these session competing for blocks.

Check out Oracle Performance Tuning 101 Book and you can send your 10046 event data to www.hotsos.com for review if you are confused by it

fwiw. good luck. mike

-----Original Message-----

Sent: Friday, January 31, 2003 3:09 AM
To: Multiple recipients of list ORACLE-L

We've got about 30 sites all running the same application, and I'm consistently seeing large numbers of 106 (library cache) latch free waits.
They tend to happen at peak times during the day, and in the worst case I
saw 12 sessions all on a 106 latch free wait event, spread across 3 P1RAW
addresses.

Running Steve Adams latch_sleeps scripts, yields the following:

LATCH TYPE                                 IMPACT SLEEP RATE WAITS

HOLDING
LEVEL
------------------------------------- ----------- ----------

library cache                             1281502      0.11%

2399666
5

cache buffers chains                       273556      0.00%

23049
1

shared pool                                 73893      0.04%

91633
7

cache buffers lru chain                     12236      0.01%

70756
3

session allocation                          10639      0.06%

19969
5

row cache objects                            7835      0.00%

29816
4

cache buffer handles                         3646      0.00%

2575
3

transaction allocation                       2344      0.01%

4341
8

enqueue hash chains                          1831      0.01%

13722
4

redo writing                                  778      0.01%

17328
5

session idle bit                              714      0.00%

0
1

The results above are from an instance which has been up for 5 days

As you can see, library cache latch has a big impact (though I must admit,
I'm not sure what Steve's IMPACT formula actually tells me). When I check
across other sites, I see a similar pattern - large numbers of 106 latch misses and sleeps.

I guess what I'd like to know is where these latches are happening, which
objects / cursors etc are causing the contention. I've grappled with SQL against x$kglob, trying to join back to the P1RAW but am not getting very
far.

Any ideas?

TIA. Neil.

--

This correspondence is confidential and is solely for the intended recipient(s). If you are not the intended recipient, you must not use, disclose, copy, distribute or retain this message or any part of it. If you
are not the intended recipient please delete this correspondence from your
system and notify the sender immediately.

No warranty is given that this correspondence is free from any virus. In keeping with good computer practice, you should ensure that it is actually
virus free. E-mail messages may be subject to delays, non-delivery and unauthorised alterations therefore, information expressed in this message is
not given or endorsed by Sx3 unless otherwise notified by our duly authorised representative independent of this message.

Sx3 is a trading name of Service and Systems Solutions Limited, a limited
company registered in Northern Ireland under number NI 32979 whose registered office is at 120, Malone Road, Belfast, BT9 5HT.
--

Please see the official ORACLE-L FAQ: http://www.orafaq.net
--

Author: Harvey Neil
INET: Neil.Harvey_at_sx3.com

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services


---------------------------------------------------------------------

To REMOVE yourself from this mailing list, send an E-Mail message

to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
--

Please see the official ORACLE-L FAQ: http://www.orafaq.net
--

Author: Johnson, Michael
INET: Michael.Johnson_at_oln-afmc.af.mil

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services


---------------------------------------------------------------------

To REMOVE yourself from this mailing list, send an E-Mail message

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services


---------------------------------------------------------------------

To REMOVE yourself from this mailing list, send an E-Mail message

--

Please see the official ORACLE-L FAQ: http://www.orafaq.net
--

Author: Cary Millsap
INET: cary.millsap_at_hotsos.com

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services


---------------------------------------------------------------------

To REMOVE yourself from this mailing list, send an E-Mail message