Db_CACHE_SIZE
I have it set to 0 which just show that I use it.
This is the output I have now:
select id,name,block_size, current_size,buffers from v$buffer_pool;
3,DEFAULT,8192,4656,578508;
My SGA is 6G.
what will be the optimum size I could resize this parameter to?
Thanks,
Hi
USE v$db_cahce_advice for setting db_cache_size parameter
but this view is available on 9.2 onwards.
Thanks and Regards
Kuljeet Pal Singh
Similar Messages
-
Buffer Qua is 60 % , incre the sga_max_size and db_cache_size on wht basis
Dear All,
In our XI Production system, data buffer quality is 60%, as i found out that sga_max_size(1.1 GB) and db_cache_size(570 MB) not set as per the standard, and also they are less when compared with development(8 GB RAM).
Our RAM size in PRD is 10 GB,and when i see in topas 90% of RAM is filled, if we increase these parameter by 1GB, i think System's overall perform will be effected as the total RAM will be filled. Please suggest me how to solve the problem and on what basis these parameters should be set(what percent of RAM).
OS : AIX 5.3
SAP XI 7.00
DB : Oracle
Please inform whether i have to include any other info
Regards,
Balaji Vedagiria) Why was it changed to less than default? Was the purpose to give memory to Java stacks?
Ans: These parameters are set long back
tell me how much RAM memory does the ABAP+JAVA stack occupy(minimum), as 90% of RAM is filled now.
b) Is there a performance issue now, XI is not a database heavy applicaton? If not why bother?
Ans: Currently there are not performance issues, but in EWA report it is RED and my customer wants it to increase the quality to atleast by 90%.
c) Why not change it back to default or to the same as QA?
Ans: As my analysis we need to increae the parameters by 1GB both means, the remaining RAM(1GB) will be filled means, my total RAM will filled this will lead to overall performance problem.
Regards
Balaji Vedagiri -
Hi,
Oracle 9.2.0.5.0 ( solaris )
I've got a query which when run on a production machine runs very slow ( 10 hours ), but on a preproduction machine ( with same data ) takes about a 10th of the time. I have confirmed that on both machines we are getting the same plan.
The only thing I can nail it down to, is that in production I'm seeing lots more "db file sequential read" wait events. Can I assume this is due to the blocks not being in/staying in the cache?
When running on preprod, the hit ratio for the query is .90 + , on production it drops down to .70 - .80 ( as per query below )
I have plenty of memory available on the machine, would it be wise to size up the caches? db_cache_size, db_keep_cache_size, db_recycle_cache_size ?
SELECT (P1.value + P2.value - P3.value) / (P1.value + P2.value)
FROM v$sesstat P1, v$statname N1, v$sesstat P2, v$statname N2,
v$sesstat P3, v$statname N3
WHERE N1.name = 'db block gets'
AND P1.statistic# = N1.statistic#
AND P1.sid = &sid
AND N2.name = 'consistent gets'
AND P2.statistic# = N2.statistic#
AND P2.sid = P1.sid
AND N3.name = 'physical reads'
AND P3.statistic# = N3.statistic#
AND P3.sid = P1.sid
PRE-PRODUCTION
call count cpu elapsed disk query current rows
Parse 1 0.64 0.64 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 2 186.92 329.88 162174 5144281 5 1
total 4 187.56 330.53 162174 5144281 5 1
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
SQL*Net message to client 2 0.00 0.00
db file sequential read 160098 1.44 162.52
db file scattered read 1 0.00 0.00
direct path write 27 0.66 3.36
direct path read 97 0.00 0.02
SQL*Net message from client 2 985.79 985.79
PRODUCTION
call count cpu elapsed disk query current rows
Parse 1 2.41 2.34 79 16 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 2 844.76 12305.06 1507519 5226663 0 1
total 4 847.17 12307.41 1507598 5226679 0 1
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
SQL*Net message to client 2 0.00 0.00
db file sequential read 1502104 4.40 11849.13
direct path write 361 0.57 3.06
direct path read 361 0.05 0.88
buffer busy waits 36 0.02 0.17
latch free 5 0.01 0.01
log buffer space 2 1.00 1.37
SQL*Net message from client 2 687.95 687.95
Suggestions for further investigation more than welcome.user12044475 wrote:
Hi,
Oracle 9.2.0.5.0 ( solaris )
I've got a query which when run on a production machine runs very slow ( 10 hours ), but on a preproduction machine ( with same data ) takes about a 10th of the time. I have confirmed that on both machines we are getting the same plan.
The only thing I can nail it down to, is that in production I'm seeing lots more "db file sequential read" wait events. Can I assume this is due to the blocks not being in/staying in the cache?
There are more physical reads, and the average read time is longer. This may simply be a reflection of the fact that other people are working on the production database at the same time and (a) kicking your data out of the cache and (b) causing you to queue at the disc as they do their I/O. A larger cache MIGHT protect your data a little longer, and MAY reduce their I/O at the same time so that the I/Os are faster - but we have no idea what side effects might then appear.
It's also worth considering whether you did something as you tranferred the data from production to pre-production that helped to improve query performance. (As a simple example, an export/import could have eliminated a lot of row migration - and the nature of your plan means you MIGHT be suffering a lot of excess I/O from "table fetch continued row"). So, how did you get the data from production to test, how long ago, what's happened to it since, and do you have any session statistics taken as you ran the two queries ?
Since your execution plan (prediction) is a long way off the actual run time, though, (even on the pre-production system), it's probably more important to work out whether you can make your query much more efficient before you make any dramatic changes to the system. I notice that you have three existences subqueries that appear at the end of the plan - such subqueries wreck the optimizer's arithmetic in your version of Oracle and can make it do very silly things. (See for example this blog note: http://jonathanlewis.wordpress.com/2006/11/08/subquery-selectivity )
The effect of the subqueries may (for example) be why you have a full tablescan on the second table in a nested loop join at one point in your query. The expectation of a vastly reduced number of rows may be why you are seeing nested loops all over the place when (possibly) a couple of hash joins would be more appropriate.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}{noformat} (lowercase, curly brackets, no spaces) so that the text appears in fixed format.
There is a +"Preview"+ tab at the top of the text entry panel. Use this to check what your message will look like before you post the message. If it looks a complete mess you're unlikely to get a response. (Click on the +"Plain text"+ tab if you want to edit the text to tidy it up.)
+"I believe in evidence. I believe in observation, measurement, and reasoning, confirmed by independent observers. I'll believe anything, no matter how wild and ridiculous, if there is evidence for it. The wilder and more ridiculous something is, however, the firmer and more solid the evidence will have to be."+
Isaac Asimov -
Oracle 9.2.0.8 , decrease db_cache_size ,but OS hasn't freed those memory
hi,
anyone ever had the same expierence as me , succeeded in alter system set db_cache_size to a smaller value , and query from v$buffer_pool shows clearly it's in effect now ,
but , somehow , OS memory utiliztion are still as high as before .
anyway to release those memory ?
is it a must to restart the instance ?Are you not using oradism utility ? What is your platform ?
If its sun solaris + 9i, then i have experienced this before if you set SGA_MAX_SIZE parameter. What happens is say your SGA_MAX_SIZE is 4GB and your DB_CACHE_SIZE is 512MB initially. Oracle locks the 4GB into physical memory initially. The oradism utility locks only whats "currently" allocated into physical memory and allocates the rest from swap. -
Sizing DB_CACHE_SIZE and SHARED_POOL_SIZE
Hello, is the a good rule to use when sizing DB_CACHE_SIZE and SHARED_POOL_SIZE is terms of a percentage of the total memory you are allocating to Oracle?
For example I have used the following:
I take 70% of the total memory on the machine and allocate it to Oracle.
70% to DB_CACHE_SIZE
30% to SHARED_POOL_SIZE
I am looking for the best way to do this as a starting point. Thank you.
DavidHi,
Database concepts
Database Administrator guide
Oracle Reference
are some of the online books.
You may refer to
http://www.oracle.com/pls/db92/db92.docindex?remark=homepage
Also check the documentation for statspack and try installing statspack on your system and see the reports, to identify the bottelnecks and resolve the problems.
Performance tuning will be a continuous activity till you have a tuned stable system. As other member said, there are no golden rules for the pool sizes.
Regards,
Badri. -
(9I) DYNAMIC SGA : SGA_MAX_SIZE, DB_CACHE_SIZE, DB_KEEP_CACHE_SIZE
제품 : ORACLE SERVER
작성날짜 : 2005-01-05
(9I) DYNAMIC SGA : SGA_MAX_SIZE, DB_CACHE_SIZE, DB_KEEP_CACHE_SIZE
==================================================================
PURPOSE
Oracle 9i의 새 기능인 동적으로 SGA 파라미터들을 변경하는 방법에
대하여 알아보기로 한다.
Explanation
Oracle 8i까지는 Buffer Cache, Shared Pool, Large Pool 등과 같은 SGA
파라미터들에 대해 그 크기를 동적으로, db가 운영 중인 상태에서는 변경할
수가 없었다.
즉, 이러한 파라미터들을 변경하려면 db를 shutdown하고 initSID.ora 화일에
서 그 크기를 다시 설정하고, 이 파라미터를 이용해서 db 인스턴스를 restart
해야만 했었다.
Oracle 9i에서는 DBA가 ALTER SYSTEM 명령을 이용해서 SGA 파라미터의 크기
를 동적으로 변경할 수 있게 되었다. 이 특정을 'Dynamic SGA'라고 부른다.
SGA 전체의 최대 크기(SGA_MAX_SIZE)를 정의하고 그 한도 내에서 파라미터의
크기를 변경할 수 있는 것이다. 데이타베이스를 shutdown/startup 없이 작업
이 가능하기 때문에 'Planned Downtime'을 줄이는 한 방법으로도 이해할 수
있다.
이 글에서는 SGA에 할당할 수 있는 최소 단위인 'Granule'의 개념을 살펴보
고, 이 granule이 어떠한 방법에 의해 동적으로 할당되는지에 대해 알아보고
자 한다.
또한 Buffer Cache 파라미터 중 새로운 것과 이전 버전에 비해 달라진 내용
을 소개하기로 한다.
1. Granule
Granule은 가상 메모리 상의 연속된 공간으로, dynamic SGA 모델에서 할당할
수 있는 최소 단위이다. 이 granule의 크기는 SGA 전체의 추정값
(SGA_MAX_SIZE)에 따라 다음과 같이 구분된다.
4MB if estimated SGA size is < 128M
16MB otherwise
SGA의 Buffer Cache, Shared Pool, Large Pool 등의 파라미터는 이 granule
단위로 늘어나거나 줄어들 수 있다. (현재 dynamic SGA를 사용할 수 있는
SGA 관련 파라미터는 Buffer Cache, Shared Pool, Large Pool 세 가지이다.)
2. Dynamic SGA(DB_CACHE_SIZE, SHARED_POOL_SIZE)
DBA는 ALTER SYSTEM 명령을 통해 initSID.ora 화일에 정의된 SGA 관련 파라미
터 값을 동적으로 변경할 수 있다. SGA 파라미터의 크기를 늘려주기 위해서
는 필요한 만큼의 free granule이 존재해야만 하며, 현재 사용하고 있는 SGA
의 크기가 SGA_MAX_SIZE보다 작아야 한다. Free granule이 없다고 해서 다른
파라미터로부터 granule을 free시켜서 그 granule을 이용할 수 있는 것은 아
니다.
반드시 DBA가 명시적으로 free/allocate해야 한다.
다음의 예를 살펴보자. 설명을 단순화하기 위해 이 경우는 SGA가 Buffer
Cache와 Shared Pool로만 구성되었다고만 하자.
예) initSID.ora
SGA_MAX_SIZE = 128M
DB_CACHE_SIZE = 96M
SHARED_POOL_SIZE = 32M
Note : DB_CACHE_SIZE는 Oracle 9i에 새롭게 도입된 파라미터이다.
위와 같은 상태일 때 동적으로 SHARED_POOL_SIZE를 64M로 늘리면 에러가 발생
한다.
SQL> ALTER SYSTEM SET SHARED_POOL_SIZE=64M;
(insufficient memory error message)
이 에러는 SHARED_POOL_SIZE를 늘림으로써 전체 SGA의 크기가 SGA_MAX_SIZE
보다 커지기 때문에 발생한다. (96M + 64M > 128M)
이를 해결하기 위해서는 DB_CACHE_SIZE를 줄인 후, SHARED_POOL_SIZE를 늘린다.
SQL> ALTER SYSTEM SET DB_CACHE_SIZE=64M;
SQL> ALTER SYSTEM SET SHARED_POOL_SIZE=64M;
Note : DB_CACHE_SIZE가 shrink되는 동안에
ALTER SYSTEM SET SHARED_POOL_SIZE=64M;
를 하면 insufficient error가 발생할 수도 있다.
이 경우는 DB_CACHE_SIZE가 shrink된 후 다시 수행하면 정상적으로
수행이 된다.
Note : 위 예제의 경우 estimated SGA 크기가 128M 이상이므로, granule의
단위는 16M이다. 따라서 SGA 파라미터의 크기를 16M의 정수배로 했다.
16M의 정수배가 아닌 경우는 지정한 값보다 큰 값에 대해 16M의
정수배 중 가장 가까운 값을 택하게 된다.
즉, 아래 두 문장의 결과는 똑같다.
SQL> ALTER SYSTEM SET SHARED_POOL_SIZE=64M;
SQL> ALTER SYSTEM SET SHARED_POOL_SIZE=49M;
Note : LARGE_POOL_SIZE 와 JAVA_POOL_SIZE 파라미터는 동적으로 변경하는
것이 불가능하다.
1) Dynamic Shared Pool
인스턴스 start 후, Shared Pool의 크기는 다음과 같은 명령에 의해 동적으
로 변경(grow or shrink)될 수 있다.
ALTER SYSTEM SET SHARED_POOL_SIZE=64M;
다음과 같은 제약 사항이 있다.
- 실제 할당되는 크기는 16M의 정수배가 된다.
- 전체 SGA의 크기는 SGA_MAX_SIZE를 초과할 수는 없다.
2) Dynamic Buffer Cache
인스턴스 start 후, Buffer Cache의 크기는 다음과 같은 명령에 의해 동적으
로 변경(grow or shrink)될 수 있다.
ALTER SYSTEM SET DB_CACHE_SIZE=96M;
다음과 같은 제약 사항이 있다.
- 실제 할당되는 크기는 16M의 정수배가 된다.
- 전체 SGA의 크기는 SGA_MAX_SIZE를 초과할 수는 없다.
- DB_CACHE_SIZE는 0이 될 수 없다.
3. Buffer Cache 파라미터의 변경된 내용
여기서는 Buffer Cache 파라미터와 관련하여 Oracle 9i에 의미가 없어진 파라
미터와 새롭게 추가된 파라미터, 그리고 dynamic SGA 중 Buffer Cache와 관련
이 있는 부분에 대해 기술하고자 한다.
1) Deprecated Buffer Cache Parameters
다음의 세 가지 파라미터는 backward compatibility를 위해 존재하는 것으
로, 차후 의미가 없어진다.
- DB_BLOCK_BUFFERS
- BUFFER_POOL_KEEP
- BUFFER_POOL_RECYCLE
위의 파라미터들이 정의되어 있으면 이 값들을 사용하게 될 것이다. 하지만,
다음에 나올 새로운 파라미터들을 사용하는 것이 좋으며, 만일 위 파라미터
(DB_BLOCK_BUFFERS, BUFFER_POOL_KEEP, BUFFER_POOL_RECYCLE) 값들을 사용
한다면 이 글에서 설명한 dynamic SGA 특징을 사용할 수는 없다. 또한
initSID.ora 화일에 위 파라미터들과 새로운 파라미터를 동시에 기술한다면
에러가 발생한다.
2) New Buffer Cache Sizing Parameters
다음의 세 파라미터가 추가되었다. 이 파라미터들은 primary block size에
대한 buffer cache 정보를 다루고 있다.
- DB_CACHE_SIZE
- DB_KEEP_CACHE_SIZE
- DB_RECYCLE_CACHE_SIZE
DB_CACHE_SIZE 파라미터에 지정된 값은 primary block size에 대한 default
Buffer Pool의 크기를 의미한다. 또한 이전 버전과 마찬가지로 KEEP과
RECYCLE buffer pool을 둘 수 있는데, 이는 DB_KEEP_CACHE_SIZE,
DB_RECYCLE_CACHE_SIZE 라는 파라미터를 이용한다.
이전 버전과 다른 점은 이전 버전의 경우 각각의 파라미터
(DB_BLOCK_BUFFERS, BUFFER_POOL_KEEP,BUFFER_POOL_RECYCLE)에 정의된 값들
이 buffer 갯수(즉, 실제 메모리 크기를 구하려면 db_block_size를 곱했어야
했다. )였는데 반해 이제는 구체적인 메모리 크기이다.
또한 이전에는 DB_BLOCK_BUFFERS가 BUFFER_POOL_KEEP, BUFFER_POOL_RECYCLE
의 값을 포함하고 있었지만, 이제는 DB_CACHE_SIZE가 DB_KEEP_CACHE_SIZE,
DB_RECYCLE_CACHE_SIZE를 포함하고 있지 않다.
즉, 각각의 파라미터들은 독립적이다.
Note : Oracle 9i부터는 multiple block size(2K, 4K, 8K, 16K, 32K)를 지원한다.
위에서 언급한 primary block size는 DB_BLOCK_SIZE에 의해 정해진 block
size를 의미한다. (SYSTEM tablespace는 이 block size를 이용한다.)
3) Dynamic Buffer Cache Size Parameters
바로 위에서 언급한 세 파라미터는 아래와 같이 ALTER SYSTEM 명령에 의해
동적으로 변경 가능하다.
SQL> ALTER SYSTEM SET DB_CACHE_SIZE=96M;
SQL> ALTER SYSTEM SET DB_KEEP_CACHE_SIZE=16M;
SQL> ALTER SYSTEM SET DB_RECYCLE_CACHE_SIZE=16M;
Example
none
Reference Documents
<Note:148495.1>Hello Martin,
Can I execute by ORA-27102 the following commandos in order to set and use the values of sga_max_size immediately?
Variant 1)
>sqlplus /nolog
>connect / as sysdba
> startup nomout
> ALTER SYSTEM SET SGA_MAX_SIZE= value SCOPE=pfile;
> shutdown immediate
> startup
Variant 2)
Changing the values of sga_max_size, etc. in init<DBSID>.ora
>sqlplus /nolog
>connect / as sysdba
> startup from pfile = /oracle/<SID>/dbs/ init<DBSID>.ora;
Thank you very much!
regards
Thom -
Need to increase db_cache_size
Hi Forum Friends,
I have one concern, On one production server as per ADDM analysis report we need to increase db_cache_size at db level .
According to my knowledge, SGA_MAX_SIZE= db_cache_size+shared_pool size + large pool size . Pls correct if i am wrong ?
As per my DB settings:
============================================================
sga_max_size = 70G
db_cache_size = 60928M
shared_pool_size = 8G
large_pool_size= 0
sga_target= 0
As per ADDM report db_cache_size should be 64,768M
================================================================================================================
But If my above statement is correct then we cannot increase db_cache_size untill we will increase sga_max_size so that all parameter get involved in it ?
Regards,
Vishal RainaThe SGA = Fixed size + Variable size + DB Cache Buffers + Online Redo Buffers. Try querying v$sga and v$stastat to see how your space is allocated.
Also Oracle has a strong tendency to over-identify space needs, that is, just because one of Oracle's features suggests more space is needed does not mean it really is. You may be able to relieve some of the demand on the buffer cache by tuning SQL to be more efficient. Then see if Oracle still wants more space.
HTH -- Mark D Powell -- -
How can i do it?
I've looked at the V$DB_CACHE_ADVICE but don't know how to interpret it (what kind of information is needed to determine db_cache_size)
Please help!
Thanks,OK.
Take the contents, and graph them as shown in that section of the doc. The curve has a 'if you double the cache, you get huge benefits' area, and it has a 'if you double the cache, you get almost no added benefits' area.
Stick to the 'huge benefits' portion, and stop increasing the cache when you get to the 'almost no added benefits' portion. -
Resizing resize db_cache_size parameter
Hi
i'm facing problem of IO and taking statspack report and analyzing it from statspackanalyzer they are advising me to increase db_cache_size to avoid many full table scans. could you help me on how to check an appropriate value of my db_cache_size as my configuration is below:
OS RHEL 4
RAM: 4 Gb
CPU: 8 X Intel(R) Xeon(TM) CPU 3.20GHz
Oracle dabatsae server 9.2.0.4
My actual SGA is
SQL> sho sga
Total System Global Area 957945792 bytes
Fixed Size 452544 bytes
Variable Size 150994944 bytes
Database Buffers 805306368 bytes
Redo Buffers 1191936 bytes
db_cache_size = 805306368
shared_pool_size = 50331648
log_buffer = 1048576
And the v$db_cache_size_advice show me the below output.
80 10010 1.3576 1442537193
160 20020 1.3063 1387963597
240 30030 1.2798 1359825236
320 40040 1.2475 1325527521
400 50050 1.2078 1283381091
480 60060 1.164 1236834685
560 70070 1.115 1184781753
640 80080 1.0636 1130102023
720 90090 1.0224 1086346464
768 96096 1 1062540512 (Current)
800 100100 .9814 1042794822
880 110110 .9547 1014390821
960 120120 .9343 992724772
1,040 130130 .8642 918284631
1,120 140140 .7784 827123440
1,200 150150 .767 814920392
1,280 160160 .7597 807265044
1,360 170170 .7539 801045972
1,440 180180 .7493 796205111
1,520 190190 .7456 792210522
1,600 200200 .7414 787812228
Thanks for your help.Background Wait Events for DB: GABON Instance: gabon Snaps: 12 -13
-> ordered by wait time desc, waits desc (idle events last)
Avg
Total Wait wait Waits
Event Waits Timeouts Time (s) (ms) /txn
control file parallel write 1,155 0 434 375 0.5
db file scattered read 222 0 47 212 0.1
control file sequential read 773 0 25 33 0.4
db file sequential read 77 0 3 40 0.0
log file parallel write 3,246 2,812 0 0 1.5
db file parallel write 3,104 0 0 0 1.4
rdbms ipc reply 2 0 0 21 0.0
db file single write 4 0 0 11 0.0
log buffer space 6 0 0 4 0.0
log file single write 2 0 0 6 0.0
log file sequential read 2 0 0 4 0.0
direct path read 1,692 0 0 0 0.8
direct path write 440 0 0 0 0.2
latch free 6 1 0 0 0.0
LGWR wait for redo copy 38 0 0 0 0.0
rdbms ipc message 16,736 5,965 19,462 1163 7.6
pmon timer 1,432 1,432 3,562 2488 0.6
smon timer 12 11 3,330 ###### 0.0
SQL ordered by Gets for DB: GABON Instance: gabon Snaps: 12 -13
-> End Buffer Gets Threshold: 10000
-> Note that resources reported for PL/SQL includes the resources used by
all SQL statements called within the PL/SQL code. As individual SQL
statements are also reported, it is possible and valid for the summed
total % to exceed 100
CPU Elapsd
Buffer Gets Executions Gets per Exec %Total Time (s) Time (s) Hash Value
16,712,841 1 16,712,841.0 37.0 771.56 2430.22 2643858346
Module: sqlservr.exe
SELECT max(startdate) maxstartdate FROM in_cdr
1,628,702 1 1,628,702.0 3.6 308.91 429.15 3952595216
Module: TOAD 8.6.0.38
select trunc(startdate) reportdate, count(*) calls, sum(duration
) actualsecs, sum(cost) revenue, sum(balancebefore-balanceafter)
calculated_revenue from in_cdr partition(incdr_2010_0401) grou
p by trunc(startdate) union select trunc(startdate) reportdate,
count(*) calls, sum(duration) actualsecs, sum(cost) revenue, sum
1,529,024 1 1,529,024.0 3.4 93.13 567.55 1639719688
Module: sqlservr.exe
SELECT max(startdate) maxstartdate FROM voucher_cdr
952,565 303 3,143.8 2.1 31.82 264.22 485362562
Module: sqlldr@cdranalysis (TNS V1-V3)
INSERT INTO IN_CDR (ANUMBER,BNUMBER,STARTDATE,TIMEZONE,DURATIO
N,CURRENCYTYPE,COST,DEDICATEDACCUSED,BALANCEAFTER,TRAFFICCASE,TE
LESERVICECODE,LOCATION,DATAVOLUME,NUMBEROFEVENTS,FAFINDICATOR,NE
TWORKID,SERVICEPROVIDERID,SERVICECLASS,ACCOUNTGROUPID,SERVICEOFF
ERINGS,SELECTEDCOMMUNITYID,BALANCEBEFORE,BALANCEDEDICATEDAFTER,B
773,681 3 257,893.7 1.7 30.16 477.10 1041979289
Module: SQL*Plus
begin processairdata; end;
679,361 1 679,361.0 1.5 80.05 124.50 4241315052
Module: sqlservr.exe
Select * from SYSADM.subscriberdump where trunc(startdate)= t
runc(sysdate)-1
612,387 1 612,387.0 1.4 36.62 91.01 3058615587
Module: sqlservr.exe
select startdate,ServiceclassID,subscriberidmsisdn from Subscrib
erDump where startdate = to_date('2010-06-07','YYYY-MM-DD') and
subscriberimsi is not null and accountactivatedflag=1
292,140 32,640 9.0 0.6 11.84 419.64 1609964435
Module: SQL*Plus
INSERT INTO ADJUSTMENT_CDR ( NOMINALAMOUNT, MSISDN, S
ERVICECLASSID, ORIGINNODEID, ORIGINOPERATORID, DESCRIPTION
, DESCRIPTION2, AMOUNTDA1, AMOUNTDA2, AMOUNTDA3, AMOUNT
DA4, AMOUNTDA5, startdate, sequencenumber, FILEID )
VALUES ( :b15 , :b14 , :b13 , :b12 , :b11 , :b10 ,
194,749 3 64,916.3 0.4 5.11 5.42 2839087957
Module: SQL*Plus
SELECT ACDR.transactionamount , ACDR.accountnumber ,
SQL ordered by Gets for DB: GABON Instance: gabon Snaps: 12 -13
-> End Buffer Gets Threshold: 10000
-> Note that resources reported for PL/SQL includes the resources used by
all SQL statements called within the PL/SQL code. As individual SQL
statements are also reported, it is possible and valid for the summed
total % to exceed 100
CPU Elapsd
Buffer Gets Executions Gets per Exec %Total Time (s) Time (s) Hash Value
ACDR.serviceclasscurrent , ACDR.originhostname ,
ACDR.originoperatorid , ACDR.externaldata1, ACDR.ex
ternaldata2, ACDR.origintimestamp , ACDR.localsequ
encenumber, ACDR.ADJUSTMENTAMOUNTDEDICATED1, ACDR.
184,394 34,508 5.3 0.4 5.96 5.96 2453141568
Module: SQL*Plus
DELETE from AIR_CDR where ROWID = :b1
141,583 39 3,630.3 0.3 3.03 3.93 169629744
Module: sqlldr@cdranalysis (TNS V1-V3)
INSERT INTO AIR_CDR (ORIGINNODETYPE,ORIGINHOSTNAME,ORIGINFILEI
D,ORIGINTRANSACTIONID,ORIGINOPERATORID,ORIGINTIMESTAMP,HOSTNAME,
LOCALSEQUENCENUMBER,TIMESTAMP,SERVICECLASSCURRENT,TRANSACTIONTYP
E,TRANSACTIONCODE,TRANSACTIONAMOUNT,TRANSACTIONAMOUNTDEDICATED1,
TRANSACTIONAMOUNTDEDICATED2,TRANSACTIONAMOUNTDEDICATED3,TRANSACT
111,639 1 111,639.0 0.2 4.89 121.06 1973241403
Module: SQL*Plus
select * from msi_space
84,898 3 28,299.3 0.2 2.47 2.57 1562134355
Module: SQL*Plus
SELECT ACDR.subscribernumber , ACDR.origintimestamp ,
ACDR.VOUCHERSERIALNUMBER , ACDR.transactionamount,
ACDR.RECHARGEAMOUNTMARKET, ACDR.VOUCHERGROUP, ACDR
.localsequencenumber, ACDR.FILEID from AIR_CDR ACDR where
ACDR.INSERTED = 'N' and ACDR.originnodetype IN ( 'IVR', 'UGW'
63,675 8,205 7.8 0.1 2.49 2.19 4143084494
select privilege#,level from sysauth$ connect by grantee#=prior
privilege# and privilege#>0 start with grantee#=:1 and privilege
#>0
50,394 1 50,394.0 0.1 3.07 9.26 4130194716
Module: sqlservr.exe
select trunc(startdate)Days,serviceclass,datavolume,nano,bnumber
,cost from in_cdr partition (incdr_2010_0607) where teleservicec
ode in ('5','6')
40,714 1 40,714.0 0.1 1.40 5.42 5349178
Module: SQL*Plus
update infiles set records_loaded = 1072 where filename = '09101
6-0234-Ccn1CDRFile-100607-1951-503848-7395-08062010130806-MINSAT
.txt20100608'
40,714 1 40,714.0 0.1 1.55 2.94 203676107
Module: SQL*Plus
update infiles set records_loaded = 984 where filename = '091016
-0234-Ccn1CDRFile-100607-1917-503784-7274-08062010123948-MINSAT. -
The oracle version is 9.2.0.7
I want to increase db_cache_size dynamically to 6gb but i cannot increase it to even 2 gb. It is 1.5gb right now. The overall system memory is 16gb.
Can anyone tell what is the reason
sys@kishore> alter system set DB_CACHE_SIZE=3000m;
alter system set DB_CACHE_SIZE=3000m
ERROR at line 1:
ORA-02097: parameter cannot be modified because specified value is invalid
ORA-00384: Insufficient memory to grow cacheYou might have mistaken,
if you have a 32-bit Oracle instance, the address space is only 2 to the power of 32, or 4 GB, on all platforms (except on Linux/Windows, the address space is actually 2 to the power of 31). So regardless how big your physical RAM is, the max addressable memory to 32-bit Oracle is 4G.
and check here
http://www-128.ibm.com/developerworks/db2/library/techarticle/dm-0410evans/#PartI
In general, 32-bit operating system kernels can exploit four gigabytes of real memory (the physical RAM shared by the operating system and the running applications), whereas 64-bit operating system kernels can exploit more than this. Of course, some 32-bit operating system kernels can exploit more than four gigabytes of memory, but don’t do this as well as 64-bit kernels. -
Set optimal db_cache_size param
I read this article http://articles.techrepublic.com.com/5100-10878_11-5031958.html because I want to set optimal db_cache_size parameter on my database. But I still don't know which oracle metric should check to draw such chart and choose best value for my database. Could anybody give me some advice?
The article you read is 8 years old, and doesn't apply to Oracle 10g and beyond, as Oracle 10g and beyond are self-tuning.
Setting db_cache_size in 10g and beyond will disable self-tuning.
So if you are running 10g and beyond (yet again you didn't include your version) the best thing you can do is
- not to read 8 year old articles
- not to set db_cache_size
and better still
be very cautious about anything Donald K Burleson writes, as his tuning approach often includes generic measures, and most other experts like Jonathan Lewis, Tom Kyte, Richard Foote, Cary Millsap and lot of others
fundamentally disagree with him.
Finally: how to determine db_cache_size is in the article, so I suggest you re-read it. Hint: especially read the bits on the cache advisor. Try to be not distracted by the impressive multi-color graphs.
Sybrand Bakker
Senior Oracle DBA -
I am asked by oracle to provide some info using the below query.
select sum(sharable_mem) from v$db_cache_size where sharable_mem>4100 and <=10000
Does this view exist? This is on 9.2.0.7.0
Thanks
/SKHthis view doesn't exist.
db_cache_size itself is a initial parameter
maybe they refer to v$db_object_cache
SQL> desc v$db_object_cache
Name Null? Type
OWNER VARCHAR2(64)
NAME VARCHAR2(1000)
DB_LINK VARCHAR2(64)
NAMESPACE VARCHAR2(28)
TYPE VARCHAR2(28)
SHARABLE_MEM NUMBER
LOADS NUMBER
EXECUTIONS NUMBER
LOCKS NUMBER
PINS NUMBER
KEPT VARCHAR2(3)
CHILD_LATCH NUMBER -
Pga_aggregate_target & db_cache_size for an Oracle 10g Datawarehouse env
We have an Oracle 10g Datawarehousing environment , with roughly 180 GB odd data running on 3 - node RAC
with 16 GB RAM & 4 CPUs each and roughly we have 200 users and night jobs running on this D/W .
We find that query performance of all ETL Processes & joins are quite slow .Certain packages take 35 min . or so to get executed.
How much should we modify the value of pga_aggregate_target & db_cache_size parameter for this Datawarehouse environment ? This is a Production database, with Oracle Database 10g Enterprise Edition Release 10.1.0.5.0.
We use OWB 10g Tool for this D/W.
Current PGA_AGGREGATE_TARGET is 8589934592, whereas db_cache_size is
1073741824
Please suggest ,
Thanks a lot in advance ,It is not clear what is meant by term "packages" in this context - i.e. are these PL/SQL packages or some other bunch of pl/sql statements or whatever else, but simply blindly changing init parameters you won't get anything good in a predictable future.
You have to find WHAT is slow and tune that, it might very well be that it has nothing to do with pga_aggregate_target or another init parameters.
Optimizing Oracle Performance by Cary Millsap and Jeff Holt can be the gretest help in this case.
You can also look at his website www.hotsos.com, it has many whitepapers that can help you if you haven't the book.
I for example see " 3 - node RAC" in your post and that makes me suspicious because if you are running batch jobs reading/writing the same data on all 3 nodes simultaneously, then the problem might be data pinging among nodes. BUT I DON'T KNOW FOR SURE, as well as you don't know, so the only predictable scenario is to get to know what is the reason.
Gints Plivna
http://www.gplivna.eu -
Resizing db_cache_size parameter
Hi all,
i have database performance problem, the server is slow because the RAm is used in 100%, taking statspack report i got the folowing output in instance efficiency:
Instance Efficiency Percentages (Target 100%)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Buffer Nowait %: 100.00 Redo NoWait %: 100.00
Buffer Hit %: 14.15 In-memory Sort %: 99.98
Library Hit %: 98.60 Soft Parse %: 95.77
Execute to Parse %: 74.71 Latch Hit %: 99.99
Parse CPU to Parse Elapsd %: 72.98 % Non-Parse CPU: 99.90
Shared Pool Statistics Begin End
Memory Usage %: 84.61 89.19
% SQL with executions>1: 4.97 5.09
% Memory for SQL w/exec>1: 14.23 16.09
The Buffer Hit is 14 which seems very slow, from statspack analyzer they are advising to resize the DB_CACHE_SIZE parameter following the DB_CACHE_ADVICE view output below:
80 10010 1.631 4847378181
160 20020 1.4917 4433425952
240 30030 1.4153 4206245775
320 40040 1.3496 4011022458
400 50050 1.2544 3727968097
480 60060 1.1851 3522007360
560 70070 1.1305 3359950649
640 80080 1.0732 3189644899
720 90090 1.0296 3060020775
768 96096 1 2972013360
800 100100 .9785 2908132320
880 110110 .8999 2674444052
960 120120 .8683 2580516085
1,040 130130 .8366 2486253576
1,120 140140 .6408 1904428818
1,200 150150 .569 1691134670
1,280 160160 .556 1652315333
1,360 170170 .5489 1631483768
1,440 180180 .5448 1619113935
1,520 190190 .5421 1611045498
1,600 200200 .5388 1601383512
can anybody help me on what is the right size of my db_cache_size as i have 4 Gb of RAM and the SGA_MAX_SIZE is set to 2Gb and pga_aggregate_target is set to 0.
thanks
LucienHi Jonathan,
here is my entire statspack report, please help.
SQL ordered by Reads for DB: GABON Instance: gabon Snaps: 51 -52
-> End Disk Reads Threshold: 1000
CPU Elapsd
Physical Reads Executions Reads per Exec %Total Time (s) Time (s) Hash Value
77,007 2 38,503.5 0.9 5.17 5.25 3171232606
Module: sqlservr.exe
select trunc(DATELOADED) STARDATE,FILENAME, RECORDS_IN_FILE,
RECORDS_LOADED ,filenameid,ABS(RECORDS_IN_FILE-RECORDS_LOADED) D
ELTA from sysadm.infiles where to_char(DATELOADED,'YYYYMMDD
')='20100617' and (RECORDS_IN_FILE <> RECORDS_LOADED OR RECOR
DS_LOADED=0)and ABS(RECORDS_IN_FILE-RECORDS_LOADED)>1 ORDER BY
42,729 1 42,729.0 0.5 1.13 1.10 220927140
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1140-552795-0670-29062010114304-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.23 1.21 319703435
Module: SQL*Plus
update infiles set records_loaded = 952 where filename = '091016
-0234-Ccn1CDRFile-100629-1140-552797-0677-29062010114405-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.11 1.08 398750951
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1142-552799-0700-29062010114905-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 5.05 5.00 585333435
Module: sqlservr.exe
select * from sysadm.infiles where to_char(DATELOADED,'YYYYM
M')='201006'
42,729 1 42,729.0 0.5 1.23 1.20 670820997
Module: SQL*Plus
update infiles set records_loaded = 809 where filename = '091016
-0234-Ccn1CDRFile-100629-1142-552799-0700-29062010114905-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.25 1.22 712766486
Module: SQL*Plus
update infiles set records_loaded = 1011 where filename = '09101
6-0234-Ccn1CDRFile-100629-1138-552792-0668-29062010114246-MINSAT
.txt20100629'
42,729 1 42,729.0 0.5 1.25 1.22 1267800290
Module: SQL*Plus
update infiles set records_loaded = 928 where filename = '091016
-0234-Ccn1CDRFile-100629-1142-552798-0690-29062010114841-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.24 1.22 1311160266
Module: SQL*Plus
update infiles set records_loaded = 959 where filename = '091016
-0234-Ccn1CDRFile-100629-1139-552793-0675-29062010114350-MINSAT.
txt20100629'
SQL ordered by Reads for DB: GABON Instance: gabon Snaps: 51 -52
-> End Disk Reads Threshold: 1000
CPU Elapsd
Physical Reads Executions Reads per Exec %Total Time (s) Time (s) Hash Value
42,729 1 42,729.0 0.5 1.24 1.21 1402052156
Module: SQL*Plus
update infiles set records_loaded = 784 where filename = '091016
-0234-Ccn1CDRFile-100629-1140-552795-0670-29062010114304-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.11 1.09 1617445411
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1138-552792-0668-29062010114246-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.11 1.08 1686434310
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1142-552798-0690-29062010114841-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.12 1.10 1778534429
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1140-552797-0677-29062010114405-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.11 1.08 1800594875
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1142-552800-0692-29062010114859-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.11 1.08 1952783295
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1139-552794-0671-29062010114326-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.12 1.10 1985306312
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1139-552793-0675-29062010114350-MINSAT.tx
t20100629'
42,729 1 42,729.0 0.5 1.24 1.21 2073655895
Module: SQL*Plus
update infiles set records_loaded = 936 where filename = '091016
-0234-Ccn1CDRFile-100629-1138-552791-0673-29062010114335-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.11 1.08 2132012199
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1140-552796-0672-29062010114329-MINSAT.tx
t20100629'
SQL ordered by Reads for DB: GABON Instance: gabon Snaps: 51 -52
-> End Disk Reads Threshold: 1000
CPU Elapsd
Physical Reads Executions Reads per Exec %Total Time (s) Time (s) Hash Value
42,729 1 42,729.0 0.5 1.24 1.21 2179069916
Module: SQL*Plus
update infiles set records_loaded = 936 where filename = '091016
-0234-Ccn1CDRFile-100629-1139-552794-0671-29062010114326-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.24 1.21 3172937007
Module: SQL*Plus
update infiles set records_loaded = 909 where filename = '091016
-0234-Ccn1CDRFile-100629-1140-552796-0672-29062010114329-MINSAT.
txt20100629'
42,729 1 42,729.0 0.5 1.22 1.21 3567256981
Module: SQL*Plus
select distinct 'Exists' from infiles where filename = '091016-0
234-Ccn1CDRFile-100629-1142-552801-0696-29062010114902-MINSAT.tx
t20100629'
SQL ordered by Executions for DB: GABON Instance: gabon Snaps: 51 -52
-> End Executions Threshold: 100
CPU per Elap per
Executions Rows Processed Rows per Exec Exec (s) Exec (s) Hash Value
22,803 22,803 1.0 0.00 0.00 2453141568
Module: SQL*Plus
DELETE from AIR_CDR where ROWID = :b1
21,452 21,374 1.0 0.00 0.01 1609964435
Module: SQL*Plus
INSERT INTO ADJUSTMENT_CDR ( NOMINALAMOUNT, MSISDN, S
ERVICECLASSID, ORIGINNODEID, ORIGINOPERATORID, DESCRIPTION
, DESCRIPTION2, AMOUNTDA1, AMOUNTDA2, AMOUNTDA3, AMOUNT
DA4, AMOUNTDA5, startdate, sequencenumber, FILEID )
VALUES ( :b15 , :b14 , :b13 , :b12 , :b11 , :b10 ,
5,268 22,389 4.3 0.00 0.00 4143084494
select privilege#,level from sysauth$ connect by grantee#=prior
privilege# and privilege#>0 start with grantee#=:1 and privilege
#>0
1,434 1,434 1.0 0.00 0.01 1782899448
Module: SQL*Plus
INSERT INTO VOUCHER_CDR ( MSISDN, STARTDATE, SERIALNU
MBER, COST, RECHARGEAMOUNTMARKET, VOUCHERGROUP, sequence
number, FILEID ) VALUES ( :b8, to_date(substr(
:b7,1,14),'YYYYMMDDHH24MISS'), :b6, :b5, :b4, :b3, :b2
, :b1 )
1,321 0 0.0 0.00 0.00 3611653548
Module: perl@kgateway (TNS V1-V3)
ALTER SESSION SET TIME_ZONE='+01:00'
1,317 5,268 4.0 0.00 0.00 888033327
select role# from defrole$ d,user$ u where d.user#=:1 and u.user
#=d.user# and u.defrole=2 union select privilege# from sysauth$
s,user$ u where (grantee#=:1 or grantee#=1) and privilege#>0 and
not exists (select null from defrole$ where user#=:1 and role#=
s.privilege#) and u.user#=:1 and u.defrole=3
1,020 0 0.0 0.00 0.00 2009857449
Module: sqlplus@cdranalysis (TNS V1-V3)
SELECT CHAR_VALUE FROM SYSTEM.PRODUCT_PRIVS WHERE (UPPER('SQL*
Plus') LIKE UPPER(PRODUCT)) AND ((UPPER(USER) LIKE USERID) OR
(USERID = 'PUBLIC')) AND (UPPER(ATTRIBUTE) = 'ROLES')
1,020 0 0.0 0.00 0.00 3096433403
Module: sqlplus@cdranalysis (TNS V1-V3)
SELECT ATTRIBUTE,SCOPE,NUMERIC_VALUE,CHAR_VALUE,DATE_VALUE FROM
SYSTEM.PRODUCT_PRIVS WHERE (UPPER('SQL*Plus') LIKE UPPER(PRODUCT
)) AND (UPPER(USER) LIKE USERID)
1,020 1,020 1.0 0.00 0.00 4119976668
Module: sqlplus@cdranalysis (TNS V1-V3)
SELECT USER FROM DUAL
1,020 1,020 1.0 0.00 0.00 4282642546
Module: SQL*Plusregards
Lucienot. -
Hi,
I try to reduce the db_cache_size to zero (Oracle 10g on Suse 9). Therefore, I've set SGA_TARGET to 0. After issueing alter system set db_cache_size=0;, I got the error message
ORA-00383: DEFAULT cache for blocksize 8192 cannot be reduced to zero
The manual says that all tablespaces that use this blocksize should be offlined and then the operation shall be performed again. My problem is that the SYSTEM, TEMP, and UNDOTBS1 cannot be brought offline (at least not with the Enterprise Manager GUI).
Can anybody help me setting db_cache_size to 0? Thanks in advance!Hi,<br>
<br>
Very strange manear to tune database... in fact. Will it be only one user to connect to your db ?<br>
Did you know what is cache size ? I'm not sure, 4Gb seems enormous, and 16Mb seems too small...<br>
Try to not set this parameter, and Oracle will choose for you the default value, it sometimes a good idea (at least >= 4MB * number of CPUs * granule size).<br>
Read the following doc, to help you to set a valuable value : Configuring and Using the Buffer Cache<br>
But, I think that you need to read the buffer cache definition : Database Buffer Cache<br>
<br>
Nicolas.
Maybe you are looking for
-
Issue with wifi on my new itouch 4g after upgrade to 4.3.1
Hi, I have itouch 4g and based on the recent update message from itunes, i have clicked on upgrade to 4.3.1 after that i'm not able to select my wi-fi location as it is greayed out in the settings of my ipod touch. I'm upset. I tried reset option but
-
Hello, I just tried to install the new file CRforVS_13_0_9.exe on a Windows 7 Ultimate machine. I have Visual Studio 2013 and wanted Crystal Reports to be fully integrated with it. By the way, I had already uninstalled CRforVS_13_0_8.exe before I att
-
Franchise contracts posting doc. SAP ISU
Hi experts, We are integrating Gas and Electricity in the same machine. Gas use to be also in SAP but in another machine. Now, I finished the customizing but in preproduction the franchise BP does not have any open items. I do not know why when I exe
-
CS 6 InDesign crashes at opening
HI, Just upgraded to CS6 and my InDesign crashes when i try to open it. It seems other mac users are having similar problems. Please help
-
Photoshop Laver masks & grads to trasparent
I am working with a web developer who is using Fireworks and I am designing pages in Photoshop. When I give him the photoshop files the layer compositing is not working on layers that I use a vector and layer mask on or if I have a vector mask that i