Re: page_in values in vmstat output

Bob wrote:

> We are facing general dissatisfaction among the users about the system
> response time and overall performance of the application.
>
> When we ran the vmstat command on the Sun Solaris Server, we observed
> non-zero values in the page_in column of the output, and we suspect
> that this is one of the reasons for the poor performance.
>
> Availability of RAM.
>
> $ prtconf|grep Memory
> Memory size: 7168 Megabytes
>
> Time|Average runque_waits|page_ins from vmstat output.
>
> TIME Runque Page_in
> 1/22/03 9:45:14 AM 0 12830
> 1/22/03 9:40:13 AM 0 15163
> 1/22/03 9:35:13 AM 0 13938
> 1/22/03 9:30:13 AM 0 17692
> 1/22/03 9:25:12 AM 0 16095
> 1/22/03 9:20:12 AM 0 14986
> 1/22/03 9:15:12 AM 0 15324
> 1/22/03 9:10:11 AM 0 18675
> 1/22/03 9:05:11 AM 0 19822
> 1/22/03 9:00:11 AM 0 12061
> 1/22/03 8:55:10 AM 0 10130
> 1/22/03 8:50:10 AM 0 15357
>
> However the percentage of total sga's of the 4 dbs(1.3GB) running on
> this server is only 18% of the total RAM (7GB) available on this
> server. This server is not having any applications on it execept these
> 4 database servers. Users will connect thorough a CITRIX application
> server.
>
> 1st Database
> SQL> conn sys
> Connected.
> SQL> show sga;
>
> Total System Global Area 753217696 bytes
> Fixed Size 73888 bytes
> Variable Size 664915968 bytes
> Database Buffers 81920000 bytes
> Redo Buffers 6307840 bytes
>
> 2nd Database
> SQL> conn sys
> Connected.
> SQL> show sga;
>
> Total System Global Area 354767008 bytes
> Fixed Size 73888 bytes
> Variable Size 339472384 bytes
> Database Buffers 13107200 bytes
> Redo Buffers 2113536 bytes
> 3rd Database
> SQL> conn sys
> Connected.
> SQL> show sga;
>
> Total System Global Area 163115168 bytes
> Fixed Size 73888 bytes
> Variable Size 77828096 bytes
> Database Buffers 81920000 bytes
> Redo Buffers 3293184 bytes
>
> 4th database
> SQL> conn sys
> Connected.
> SQL> show sga;
>
> Total System Global Area 78991520 bytes
> Fixed Size 73888 bytes
> Variable Size 74551296 bytes
> Database Buffers 2252800 bytes
> Redo Buffers 2113536 bytes
>
> We are not able to understand why the server is paging when around 80%
> of the RAM is not reserved for any other application.
>
> Our server is heavily IO bound and IO is not distributed equally among
> the 16 disks available in this server. Does this fact has any remote
> connection with our above problem.
>
> Is this a common phenomenon or am I missing some thing here?.

Are the kernel parameters /etc/system configured for a single Oracle instance or for all four?
32bit or 64bit?

What version and edition of Oracle?
Are statistics kept current with DBMS_STATS? Have you isolated the problem to the database and not the middle-ware?

Daniel Morgan Received on Sun Feb 09 2003 - 12:14:44 CST