Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: Global Cache Converts & V$SYSSTAT "Time"

RE: Global Cache Converts & V$SYSSTAT "Time"

From: Kevin Closson <kevinc_at_polyserve.com>
Date: Thu, 20 Oct 2005 20:11:35 -0700
Message-ID: <B9782AD410794F4687F2B5B4A6FF3501021D4E36@ex1.ms.polyserve.com>


remember I'm old and crusty. please explain how anything that doesn't appear on the radar in timed events can be of concern ?  

BTW, please don't get me started about how many points there are in a GCS/GES remote event that involve process context switches....    

I admit I don't see any such timings on any of my (many may many 16 node RAC clusters), but I don't have RAC instances sitting on the periphery idle either...but most importantly, I may be calloused is saying so, if it isn't an event the makes the top 5% I wouldn't loose sleep...it must just be "the way it works" and after paying ~60K per CPU for RAC licensing, that must be sufficient.  


	From: Marquez, Chris [mailto:cmarquez_at_collegeboard.org] 
	Sent: Thursday, October 20, 2005 7:43 PM
	To: Kevin Closson; oracle-l_at_freelists.org
	Subject: RE: Global Cache Converts & V$SYSSTAT "Time"
	
	


	This can not be good!?
	
	Wait Events fine, but the gc times are sick.
	
	This was a rather shot inverval between statspack.snap's  (taken
later at night less busy).                  
	Top 5 Timed Events
	~~~~~~~~~~~~~~~~~~
% Total
	Event                                               Waits
Time (s) Ela Time
	-------------------------------------------- ------------
----------- --------
	CPU time
123    43.92
	db file sequential read                            10,004
89    31.82
	db file scattered read                              5,103
52    18.66
	log file sync                                         867
6     2.08
	control file sequential read                          449
2      .86
	
-------------------------------------------------------------
	Cluster Statistics for DB:   Instance:   Snaps: 11 -12
	
	Global Cache Service - Workload Characteristics
	-----------------------------------------------
	Ave global cache get time (ms):                          422.0
	Ave global cache convert time (ms):                    2,121.0
	
	
	
	Chris Marquez
	Oracle DBA
	
	
	-----Original Message-----
	From: oracle-l-bounce_at_freelists.org on behalf of Marquez, Chris
	Sent: Thu 10/20/2005 6:24 PM
	To: Kevin Closson; oracle-l_at_freelists.org
	Subject: RE: Global Cache Converts & V$SYSSTAT "Time"
	
	

>>is this one of your top 5 wait events ?
I have not run a statspack recently and regarding sessions I have been seeing all day, "No". Very few RAC related waits today if any, but I have seen them inexplicably in the past. Node2 for us is always idle, no db sessions....so RAC waits
should be NULL or minimal always...but that is not always the case.         

        Also, things (load, waits, etc.) were much less in general with node 2 instance down for the past few days.         

	Chris Marquez
	Oracle DBA
	
	
	
	-----Original Message-----
	From: Kevin Closson [mailto:kevinc_at_polyserve.com]
	Sent: Thu 10/20/2005 6:17 PM
	To: Marquez, Chris; oracle-l_at_freelists.org
	Subject: RE: Global Cache Converts & V$SYSSTAT "Time"
	
	is this one of your top 5 wait events ?
	
	        col "Average Time" format 999.999
	        SELECT V$SYSSTAT.VALUE*10 / (select value from v$sysstat
where
	name = 'global cache gets')  "Average Time"
	        FROM V$SYSSTAT
	        WHERE V$SYSSTAT.NAME = 'global cache get time'
	        /
	        Average Time
	        ------------
	              19.883
	       
	
	
	
	
	


--
http://www.freelists.org/webpage/oracle-l
Received on Thu Oct 20 2005 - 22:14:08 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US