RE: "latch: cache buffers chains" wait in NON-RAC Benchmark Runs
Date: Sat, 16 Feb 2008 10:17:27 +0530
Thanks Jonathan for responding & the clarifications
U wrote -> "select obj, tch from sys.x$bh where hladdr = '6F9CBE00' -- adjust as necessary"
How is the Value of "hladdr" to be specified for our NON-RAC database?
From: oracle-l-bounce_at_freelists.org [oracle-l-bounce_at_freelists.org] On Behalf Of Jonathan Lewis [jonathan_at_jlcomp.demon.co.uk] Sent: Friday, February 15, 2008 9:39 PM
Subject: Re: "latch: cache buffers chains" wait in NON-RAC Benchmark Runs
The Burleson article and the Metalink note are both seriously defective.
"Cache buffers chains" latch activity is a natural consequence of accessing data buffers. Contention for "cache buffers chains" latches is more likely in a system with a high degree of concurrency where lots of small jobs are visiting the same (relatively small) number of data blocks.
The biggest problem with the Burleson article is that he seems to be confusing latch contention with buffer busy waits.
The biggest problem with the Metalink article is that it supplies a ridiculous
query to report the "problem object" from the child latch address. Do NOT run that
query on a real production system. See the following note that comments on the
http://www.jlcomp.demon.co.uk/kiddy_scripts.html#_Sometimes_the_best_thing_to_do_is_tIf you want an efficient query to get the relevant information in 10.2, then allyou need isa query like: select obj, tch from sys.x$bh where hladdr = '6F9CBE00' -- adjust as necessary ;The OBJ is the data_object_id of the object (or dataobj# from obj$,since you're prepared to poke around at the level of the x$).Another thing to consider is that the TCH can be relatively high for anobject that is not subject to much access. The TCH is only increased atmost once every three seconds. Take a look at the following
51813 1 50303 10 2 13 41383 1 2 13 9006 1 9076 1 2 13 237 438It's Object 237 has been touched at least once every threeseconds for the last 22 minutes (in fact it's job$). Object50303 has been hammered to death for the last 30 seconds.You have to temper the touch count with knowledge aboutwhat MIGHT have been happening to the objects.To help you address the problem, you may get more cluesfrom a simple statspack/awr report - find the SQL that doesmost gets, and check the segment statistics for the objectssubject to most buffer gets - cross reference to see if thetwo sets of information are consistent, then see if youcan reduce the work done by those statements. Regards
- CAUTION - Disclaimer ***************** This e-mail contains PRIVILEGED AND CONFIDENTIAL INFORMATION intended solely for the use of the addressee(s). If you are not the intended recipient, please notify the sender by e-mail and delete the original message. Further, you are not to copy, disclose, or distribute this e-mail or its contents to any other person and any such actions are unlawful. This e-mail may contain viruses. Infosys has taken every reasonable precaution to minimize this risk, but is not liable for any damage you may sustain as a result of any virus in this e-mail. You should carry out your own virus checks before opening the e-mail or attachment. Infosys reserves the right to monitor and review the content of all messages sent to or from this e-mail address. Messages sent to or from this e-mail address may be stored on the Infosys e-mail system. ***INFOSYS******** End of Disclaimer ********INFOSYS*** -- http://www.freelists.org/webpage/oracle-l