OLTP SGA Optimum Size

Hi Everybody,
I have a Database running Oracle 9i, release 9.2.0.6.0 64bit. It is running on an IBM PSeries with AIX 5L, 8gb RAM, but only 4 scsi disks of 136gb each. My current DB configuration is set as a General Database. Numbers are as follows in bytes:
buffer_cache=2684354560
free memory=536870912
library cache=354736912
sql area=323717552
free memory=310946880
miscellaneous=75226840
db_block_hash_buckets=22162544
free memory=16777216
FileOpenBlock=7812264
PL/SQL MPCODE=6413728
XDB Schema Cac=5794944
sim memory hea=5162712
event statistics per sess=5033840
KQR L PO=4770048
log_buffer=4457472
dictionary cache=3229952
Checkpoint queue=2624000
PL/SQL DIANA=2604688
KGLS heap=2016608
KQR M PO=1751744
sessions=1203280
1M buffer=1056768
KSXR receive buffers=1034000
VIRTUAL CIRCUITS=911480
KSXR pending messages que=853952
fixed_sga=749240
parameters=707896
MTTR advisory =292144
PL/SQL PPCODE=199560
type object de=174736
errors=115104
subheap=102312
trigger defini=63000
PX subheap=38784
trigger inform=21008
table definiti=19208
KGK heap=16984
KQR S SO=13624
pl/sql source=9384
trigger source=4368
joxs heap init=4240
PLS non-lib hp=3000
fixed allocation callback=600
Total SGA = 4116.715mb
SGA Max Size = 4452.976mb
The SGA Summary is as follows in bytes:
Database Buffers=2684354560
Fixed Size=749240
Redo Buffers=4468736
Variable Size=1979711488
I'm having big headaches with locks. I understand my Database is a OLTP but never configured it like that. I missed the opportunity at the time of creation. What would be a goos start un numbers for a SGA settings?
Please help...

Kevin Tyson wrote:
I don't know why people are talking about RAID and other hardware issues when the question was about the SGA. To determine whether your SGA is the right size, you need to calculate the buffer cache hit ratio, the library cache hit ratio, and other SGA stats. Your library cache hit ratio should be over 90 percent unless your developers didn't use bind variables. Your buffer cache hit ratio depends on your application, but if it's below 80 percent I would look at the top ten offending SQL statements by disk reads, buffer hits, and CPU. That's just a start. I have scripts at home for doing these things in Oracle9i. I might respond with them this weekend.
Kevin TysonThis is so completely, utterly wrong, I can't even begin to describe why. See [the faq|http://www.orafaq.com/wiki/Improving_Buffer_Cache_Hit_Ratio]. It was wrong even in the days when tfsmstat.sql was written, around 1993.
The title of the thread is about SGA, but the question is about locks. The question is making a questionable assumption that the SGA size has something to do with locks.
Original Poster: Please post the locking evidence that has you thinking whatever it is you are thinking. It is conceivable that having a large SGA may cause latching issues in some odd circumstances, but really, there are several methodologies that can lead quite directly to the actual problem. Hit ratio tuning is not one of them. Here is one way. It may help us help you if run a statspack for 15 minutes during a bad time and post the results.
Most performance issues are due to code, the first thing you need to do is figure out if you have a code problem or a system tuning problem. Raid 0 is not a good sign. Nothing wrong with asking about a recommended SGA, but as you can see, it doesn't get you far if no one sees an obvious wrong.

Similar Messages

What's the sga max size for a 32bit windows server?

Hi,
I have a 32bit Windows Server with 16G of RAM... What is the SGA max size that I can setup?

Depending on your version of Windows and your version of Oracle there are methods to increase the usuable memory available to Oracle including the MS \3GB and \PAE switches and the AWE memory extention feature.
Oracle has notes on the subject:
Windows Memory Configuration: 32-bit and 64-bit 873752.1 plus notes 46053.1 and 46001.1
How to use Very Large Memory, higher than 4Gb on Windows 2003 32 bit #342080.1
HTH -- Mark D Powell --

Optimum size for a full browser width image?

Hi all,
I'm pretty new to Muse and these forums (hello everyone!) and have been tinkering around with some features.
I'm currently creating a website for my band that integrate a full width browser image at the top of each page. I was wondering what the optimal settings are for the image size to make this work? It all currently works fine, but on larger screens the image is so zoomed in it's impossible to make out what it's of! The site is designed for smaller screens so this works fine, I was mainly wondering if there's a recommended size to optimised the images too for the width.
Any help would be greatly appreciated!
Kind regards, Ekke

This must be a hard question to answer...for anyone...because I have tried many times before to get this question answered. I have looked at other examples of Muse sites and I have decided to go with a 2000 width for my images when I want to go full screen. I have no idea if this is correct but I get good results pertaining to how it looks. Now, I don't know if it is the optimum size in terms of download speed. No one will touch this question thus far. My research came from looking at what others did that built Muse sites as a profession.
The Muse team should address this topic of optimal images for all platforms and how images can be used throughout the build not only for the above stated but for efficiency of site building.

SGA Max Size limit?

Hi,
I have Fujitsu mid range Server with 16gb RAM and 64 bit Windows Server 2003,10g R2 db installed, current i have SGA size 4gb..
What is SGA max size limit????
One of my report runs in 24 seconds...*will this issue b solved increasing the SGA size upto 10,12 gb?*

Yes,
You can also go for a 10046 event tracing...
ACCEPT sid PROMPT 'Enter SID: '
ACCEPT serial PROMPT 'Enter SERIAL#: '
ACCEPT action PROMPT 'Enter TRUE or FALSE: '
EXEC sys.DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(&sid,&serial,&action);
prompt Trace &action for &sid,&serial
exec DBMS_SYSTEM.SET_EV(10,20,10046,12,”);
Then you can check your dump file and see whcih events are higher......
For Eg. content could be like:
=====================
PARSING IN CURSOR #6 len=107 dep=1 uid=44 oct=6 lid=44 tim=1621758552415 hv=3988607735 ad='902c07a8'
UPDATE rn_lu_lastname_loca set entr_loca_id_plz14 = translate(entr_loca_id_plz14,'_','-') where rowid = :b1
END OF STMT
PARSE #6:c=0,e=981,p=0,cr=0,cu=0,mis=1,r=0,dep=1,og=0,tim=1621758552403
BINDS #6:
bind 0: dty=1 mxl=32(18) mal=00 scl=00 pre=00 oacflg=13 oacfl2=1 size=32 offset=0
bfp=10331d748 bln=32 avl=18 flg=09
value="AAAHINAATAAAwTTABV"
WAIT #6: nam='db file sequential read' ela= 12170 p1=6 p2=197843 p3=1
WAIT #6: nam='db file sequential read' ela= 8051 p1=14 p2=261084 p3=1
WAIT #6: nam='db file sequential read' ela= 7165 p1=19 p2=147722 p3=1
WAIT #6: nam='db file sequential read' ela= 9604 p1=19 p2=133999 p3=1
WAIT #6: nam='db file sequential read' ela= 6381 p1=19 p2=133801 p3=1
EXEC #6:c=10000,e=45750,p=5,cr=1,cu=10,mis=0,r=1,dep=1,og=4,tim=1621758598343
FETCH #5:c=0,e=357,p=0,cr=5,cu=0,mis=0,r=0,dep=1,og=4,tim=1621758598896
EXEC #1:c=30000,e=116691,p=36,cr=35,cu=10,mis=0,r=1,dep=0,og=4,tim=1621758599043
WAIT #1: nam='SQL*Net message to client' ela= 5 p1=1413697536 p2=1 p3=0
WAIT #1: nam='SQL*Net message from client' ela= 2283 p1=1413697536 p2=1 p3=0
Lines that start with WAIT
len Length of SQL statement.
dep Recursive depth of the cursor.
uid Schema user id of parsing user.
oct Oracle command type.
lid Privilege user id.
ela Elapsed time. 8i: in 1/1000th of a second, 9i: 1/1'000'000th of a second
tim Timestamp. Pre-Oracle9i, the times recorded by Oracle only have a resolution of 1/100th of a second (10mS). As of Oracle9i some times are available to microsecond accuracy (1/1,000,000th of a second). The timestamp can be used to determine times between points in the trace file. The value is the value in v$timer when the line was written. If there are TIMESTAMPS in the file you can use the difference between 'tim' values to determine an absolute time.
hv Hash id.
ad SQLTEXT address (see v$sqlarea and v$sqltext).
Lines that start with PARSE, EXEC or FETCH
#n n = number of cursor
c cpu time
e elapsed time
p physical reads
cr consistant reads
cu current mode reads
mis miss in cache (?)
r rows processed
dep recursive depth
og optimizer goal
tim time Content

Set SGA maximum size larget than it's component

Dear all,
I want to activate the DISM in Oracle 10g on Solaris container, therefore I need to set the parameter sga_max_size larger than it's component.
From my understanding, below are the sga components parameter and the current value on my system::
SHARED_POOL_SIZE (560M)
LARGE_POOL_SIZE (0)
JAVA_POOL_SIZE (32M)
DB_CACHE_SIZE (560M)
STREAMS_POOL_SIZE (0)
The thing is, my sga max size parameter (1168M) is already larger than above parameters.
Is there another sga component I've missed ? Plz advice. Thanks.

Hi,
SGA: The size is determined indirectly from the size of the contained memory areas.
1) Buffer pool: DB_BLOCK_BUFFERS (unit: blocks) or DB_CACHE_SIZE when you use the dynamic SGA
2) Shared pool: SHARED_POOL_SIZE
3) Java pool: JAVA_POOL_SIZE
4) Large pool: LARGE_POOL_SIZE
5) Streams pool (Oracle 10g or later): STREAMS_POOL_SIZE
6) Redo buffer: LOG_BUFFER
In addition, in the context of the dynamic SGA , you can define parameter
SGA_MAX_SIZE, which sets an upper limit for the total size of the SGA. In
general, you can only increase the size of parameters, such as DB_CACHE_SIZE
or SHARED_POOL_SIZE, up to the size defined by SGA_MAX_SIZE.
Thanks
Sunny

Sga max size is not changed.

We use Oracle 11gr2 on win2008R2.
Our member said that he set sga max size by
alter system set sga_max_size scope=spfile; max size to 6GB.
And he restared that OS and DB instance , then
show parameters sga_max_size was 7.6GB.
we use dynamic SGA and automatic memory management.
Is it natural things ? If so, I would like to explaint about to our customer.
Why the show parameters sga_max_size was 7.6GB ?

DUPLICATE POST. please continue with our original post. Please mark this question answered and check the following link:
sga max size is different

Sga max size is different

We use Oracle 11gr2 on win2008R2.
Our member set sga max size by
alter system set sga_max_size scope=spfile; max size to 6GB.
And we restared that OS and DB instance , then
show parameters sga_max_size was 7.6GB
Why is that ?

If MEMORY_TARGET is set to a non-zero value:
If SGA_TARGET and PGA_AGGREGATE_TARGET are set, they will be considered the minimum values for the sizes of SGA and the PGA respectively. MEMORY_TARGET values can range from SGA_TARGET + PGA_AGGREGATE_TARGET to MEMORY_MAX_TARGET.
Now check if your memory_target is set.
Read this metalink note as well:
Automatic Memory Management (AMM) on 11g (Doc ID 443746.1)

What is the optimum size for a WIFI subnet ?

For LAN design fine, keep the broadcast domain to an optimum size. What about WIFI design though ? I want to use just 1 SSID and can have multiple vlans if I use AP groups on the controller..But how big is optimum size for a WIFI subnet ? I guess it depends on required speed, number of AP ,vlans and traffic profiles, but are there any recommendations ?
Stu

Hello,
Great question
First, if you are using a WLC, the WLC will proxy the broadcast for the clien so that it wont be sent over the wireless. With that being said, you will typically see larger subnets on the WLC solution. For exmaple, we have /21 here and they work fine for our voice, data and video SSIDs. Although, some people like smaller subnets and deploy ap groups accordoningly.
Although, I have seen subnets on WiFi voice as a /16 with 5,000 phones in the subnet which to my surprise worked well. Of course, I wouldnt have designed it that large.
If you are autonmous same holds truse, but if you configure the AP to send over the broadcast to the wifi clients you should consider a smaller subnet like a /23 or /24.
As to client mix. Most folks will have a Voice SSID, Data SSID, Guest SSID so you can shape these accordiingly. Again, typically all my designs lead with a /21 or /22.

Optimum size of Contacts photos?

Anyone know what the default / optimum size would be for the photo on a contact to be sync'ed onto an iPhone? As in: 150x150 pixels or something? Address Book help doesn't mention it.

If you set the contact using Entourage, Outlook, or even MobileMe it'll be a tiny little image...
Here's what you need to do instead!
In iPhoto set up a new album. I named mine "Contacts". In this album put all the photos you'd like to use for your contacts. I suggest 1200x1600 sized photos because that's what the iPhone photos are (so you'll get a consistent experience).
In iTunes, sync this album to your iPhone.
On the iPhone, go to Photos, and select a photo and push the button on the bottom left of the screen. Select "Assign to Contact".
Once you've found the appropriate contact it will now appear full screen when you receive a call.
Hope that helps!

SGA+PGA Size allocation

Hi all
I Have a server System Configuration with
RAM size: 16GB Ram & Hard disk Size of 1TB.
I Installed Oracle 11g R2 in the server System
I want to Allocate Dedicated Memory for oracle..
How Should i calculate the Memory size of SGA+PGA for good Performance..

Pl do not post duplicates - Reg:-SGA & PGA Memory Allocation Size

SGA/PGA Sizes ???

I need help!
I have a Windows 2003 Server (w/sp2) with 4GB of RAM & 2 CPUs. Oracle 10.2.0.2 is installed.
ASMM is ENABLED by default (and it still is)
These are my sga/pga settings:
pga_aggregate_target = 1G
sga_max_size = 1504M
sga_target = 1000M
shared_pool_size = 504M
sort_area_size = 262144000 (250M)
large_pool_size = 104M
db_cache_size = 104M
I believe I should DISABLED ASMM, correct? Otherwise, my settings aren't used.
Based on my 4GB, I think Oracle can use 2GB of that.
What should I set the settings above too?
...thanks

First if you use Oracle's automatic memory management feature and provide spfile parameter values for parameters that are automatically managed by Oracle then the provided values are taken as minimum values for these areas. There are other situations where values for parameters like sort_area_size which is limited to 5% of pga_aggregate_target are in fact still referenced.
With a pga_aggregate_target of 1G I do not see any valid reason why you would have a sort_area_size set to 250M. What is the reason for this?
What kind of environment is this: OLTP, DSS, OLAP?
How many concurrent sessions?
Dedicated vs shared server?
Why do you not want to use automatic memory management? What problem have you encountered?
HTH -- Mark D Powell --

SGA granule size

If SGA_MAX_SIZE <= 128 mb, granule size will be 4 mb
If SGA_MAX_SIZE > 128 mb, granule size will be 16 mbMy SGA size is 360M
SQL> show parameter SGA_MAX_SIZE
NAME                                 TYPE        VALUE
sga_max_size                         big integer 360M
SQL> Then why the size of a granule is 4MB?
SQL> SELECT COMPONENT, GRANULE_SIZE/(1024*1024) "GRANULE_SIZE(Mb)" FROM V$SGA_DYNAMIC_COMPONENTS;
COMPONENT                                                        GRANULE_SIZE(Mb)
shared pool                                                                     4
large pool                                                                      4
java pool                                                                       4
streams pool                                                                    4
DEFAULT buffer cache                                                            4
KEEP buffer cache                                                               4
RECYCLE buffer cache                                                            4
DEFAULT 2K buffer cache                                                         4
DEFAULT 4K buffer cache                                                         4
DEFAULT 8K buffer cache                                                         4
DEFAULT 16K buffer cache                                                        4
DEFAULT 32K buffer cache                                                        4
Shared IO Pool                                                                  4
ASM Buffer Cache                                                                4
14 rows selected
SQL> Shouldn't have been 16?...

user12045475 wrote:
Hi,
your doc. is for 9i
for 10g and forward
If SGA_MAX_SIZE <= 1GB, granule size will be 4 mb
If SGA_MAX_SIZE > 1GB, granule size will be 16 mbHTHThis is not correct. You need to test for your release and platform.orcl> SELECT COMPONENT, GRANULE_SIZE/(1024*1024) "GRANULE_SIZE(Mb)" FROM V$SGA_DYNAMIC_COMPONENTS;
COMPONENT                                                        GRANULE_SIZE(Mb)
shared pool                                                                     8
large pool                                                                      8
java pool                                                                       8
streams pool                                                                    8
DEFAULT buffer cache                                                            8
KEEP buffer cache                                                               8
RECYCLE buffer cache                                                            8
DEFAULT 2K buffer cache                                                         8
DEFAULT 4K buffer cache                                                         8
DEFAULT 8K buffer cache                                                         8
DEFAULT 16K buffer cache                                                        8
DEFAULT 32K buffer cache                                                        8
Shared IO Pool                                                                  8
ASM Buffer Cache                                                                8
14 rows selected.
orcl> sho parameter sga_max_size
NAME                                 TYPE        VALUE
sga_max_size                         big integer 1104M
orcl> select * from v$version;
BANNER
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - Production
PL/SQL Release 11.2.0.3.0 - Production
CORE    11.2.0.3.0      Production
TNS for 32-bit Windows: Version 11.2.0.3.0 - Production
NLSRTL Version 11.2.0.3.0 - Production
orcl>

ORA-04031 on 10g - should I just adjust my SGA POOL SIZE?

Has anyone gotten this message frequently:
ORA-04031: unable to allocate 37536 bytes of shared memory ("shared pool","unknown object","sga heap(1,0)","session parame")
We are a business intelligence application that issues lots of large queries. We just migrated to 10g and we are seeing this every 2-3 days on our testing machine.
In particular, I am not sure about "sga heap"... I would just set my Shared Pool Size higher - currently 144 MB but will this help here? Thoughts?

In Oracle 10g a new feature called "automatic memory management" allows the dba to reserve a pool of shared memory that is used to allocate the shared pool, the buffer cache, the java pool and the large pool.
In general, when the database needs to allocate a large object into the shared pool and cannot find contiguous space available, it will automatically increase the shared pool size using free space from other SGA structure.
Since the space allocation is automatically managed by Oracle, the probability of getting ora-4031 errors may be greatly reduced. Automatic Memory Management is enabled when the parameter SGA_TARGET is greater than zero and the current setting can be obtained quering the v$sga_dynamic_components view.
Please refer to the 10g Administration Manual for further reference
http://download-west.oracle.com/docs/cd/B19306_01/server.102/b14231/toc.htm

Estimate cache displacement - optimum size of CACHE_SIZE

Is there a way to determine out of the statistics data in the dbanalyzer what would be the optimum cache size for the database?
I want to know that because we plan to buy a new box and I want to size it so the actual "daily used data" is in the cache -> no (or few) User-UKTs doing reading I/O. So basically building something like an "in memory database".
Markus

> interesting are the physical I/O. This tells us that not all pages which have to be processed can be found in the data Cache.
Yes - and there's a lot of it
> The parameter LRU_FOR_SCAN restricts the size of the data cache which is used for Scans. More Info about the Data Cache configuration you will find in the MaxDb Internals slides: [Parameters|http://maxdb.sap.com/training/internals_7.6/kernel_parameters_EN_76.pdf]
LRU_FOR_SCAN is set to NO. If we would set that to YES and one person would start a big job (MRP-runs for PP, CO precaluations etc.) this would "trash" the full cache and all other UKTs would need to do I/O to get their current data again.
> Could it be that we have a lot of scans in the system on large tables?
> Did you get any more detailrd information with the Command Monitor which commands (Reports) have to read from disk?
No full table scans but "just" longrunning jobs (tablename, runtime, Cache I/O, Disk I/O)
Table     Runtime          P accesses     P Cache I/O     # Disk I/O
===================================================================
COVP     61.866,92     6.289.191     196.602          6.092.589
ACCTIT     35.678,21     5.215.440     896.540          4.318.900
COSS     12.321,35     1.872.759     606.268          1.266.491
BALHDR     12.238,66     1.134.360     65.534          1.068.826
COSS     12.109,92     1.764.038     601.264          1.162.774
VBRP     11.909,13     1.850.768     660.289          1.190.479
COSS     11.188,26     1.872.726     601.976          1.270.750
BKPF     8.734,46     916.325          65.534          850.791
There's not a single table scan done, all are done via IX_RG or PK_RG or a combination of them. Those are just the "top 10", there are much more of those.
What I now want to archive is the Disk-I/O becoming 0 due to a big CACHE_SIZE.
>
> May be we should analyze this directly on your system via OSS call to get more ideas.
340157/2008 - same question there too
Markus

Optimum Size for Web Gallery

In trying to set up galleries that have quick view time, I'm trying to follow iPhoto advise and use a Medium sized photo for Web Galleries. iPhoto help sez:+When you send, print, or publish a photo from your iPhoto library, you need to choose the right size+. For me that would be Medium size.
However I can find no place to determine size. How can I choose the size I want to use?

smidi:
I believe that statement pertains to situation where you can control the size of the file to be used. However, in publishing a Web Gallery iPhoto determines the size. If you don't select the option to have viewers download the photos then the image size is 800 x 600 pixels. If you selected the download option then the full sized file is uploaded for viewer to download.
For emailing, if you use Mail, you will get a window which lets you select the size of image/file to send. For printing you want to use the largest size you can so you get the max resolution available.
Do you Twango?
TIP: For insurance against the iPhoto database corruption that many users have experienced I recommend making a backup copy of the Library6.iPhoto database file and keep it current. If problems crop up where iPhoto suddenly can't see any photos or thinks there are no photos in the library, replacing the working Library6.iPhoto file with the backup will often get the library back. By keeping it current I mean backup after each import and/or any serious editing or work on books, slideshows, calendars, cards, etc. That insures that if a problem pops up and you do need to replace the database file, you'll retain all those efforts. It doesn't take long to make the backup and it's good insurance.
I've written an Automator workflow application (requires Tiger), iPhoto dB File Backup, that will copy the selected Library6.iPhoto file from your iPhoto Library folder to the Pictures folder, replacing any previous version of it. It's compatible with iPhoto 08 libraries. You can download it at Toad's Cellar. Be sure to read the Read Me pdf file.

OLTP SGA Optimum Size

Similar Messages

Maybe you are looking for