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: Who and What is using the shared pool.

RE: Who and What is using the shared pool.

From: Philip West <P.West_at_g-icap.com>
Date: Thu, 24 Aug 2000 16:38:56 +0100
Message-Id: <10599.115482@fatcity.com>


This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible.

------_=_NextPart_001_01C00DE3.CFD168DE
Content-Type: text/plain;

        charset="iso-8859-1"

Lisa,  

I am baffled. I do not have access to 8.x at the moment. Is there anyone out there who can help Lisa?  

Phil West - Impex IT ltd
Unix Sys Admin and Oracle Financials DBA Services

all opinions are worth [at most] what you paid for them

-----Original Message-----
From: Koivu, Lisa [mailto:lkoivu_at_qode.com] Sent: 24 August 2000 16:52
To: Multiple recipients of list ORACLE-L Subject: RE: Who and What is using the shared pool.

Hi Phil,  

8.1.5 on HP/UX 11.0. Check it out below. I figured I would have to look at executions to determine if something that is constantly being reloaded really needs to be pinned. Seems to me if it's being executed a lot, and reloaded a lot, it should be pinned - and if it's being reloaded a lot but not executed more than once per load, it's not worth the effort.  

Thanks in advance for any comments.
Lisa  

SQL> select startup_time from v$instance;  

STARTUP_T



02-AUG-00     1 select owner || '.' || name OBJECT
  2          , type
  3          , to_char(sharable_mem/1024,'9,999.9') "SPACE(K)"
  4          , loads
  5          , executions execs
  6          , kept

  7 from v$db_object_cache
  8 where type in ('FUNCTION','PACKAGE','PACKAGE BODY','PROCEDURE')   9 and owner not in ('SYS')
 10* order by loads desc
SQL> /  
OBJECT                                   TYPE            SPACE(K)     LOADS
EXECS KEP
---------------------------------------- --------------- -------- ---------
--------- ---
QODE_LOAD.F_VALIDATE_ISBN                FUNCTION             1.9        31
0 NO
QODE_BC.PKG_GN_SPECIAL_OFFERS            PACKAGE BODY        12.3        29
0 NO
QODE_APP.CLEANUP_PARTNERNAME_STG2        PROCEDURE            2.8        25
0 NO
QODE_APP.PKG_GENERATE_KEYWORD            PACKAGE BODY         3.1        21
0 NO
QODE_APP.PKG_UTIL                        PACKAGE             28.0        17
0 NO
QODE_APP.PKG_CLEANUP                     PACKAGE BODY         9.3        17
0 NO
QODE_BC.PKG_GN_SPECIAL_OFFERS            PACKAGE             11.3        16
0 NO
QODE_BC.CATHIERARCHY                     PACKAGE              1.7        15
0 NO
QODE_BC.RPT_PARTNER_EMAIL                PACKAGE              1.7        15
0 NO
QODE_BC.PARTNER1                         PACKAGE              1.7        15
0 NO
QODE_APP.P_GUESTHOME                     PROCEDURE           73.7        15
0 NO
QODE_USER.WEBDBTEST012345678912          PROCEDURE            1.7        15
0 NO
QODE_BC.PARTRELATIONSHIP                 PACKAGE              1.2        15
0 NO
QODE_BC.PARTNER_RELATIONSHIP1            PACKAGE              1.7        15
0 NO
QODE_BC.CATHIERARCHY2                    PACKAGE              1.7        15
0 NO
QODE_APP.P_BUYQODER                      PROCEDURE           38.8        15
0 NO
QODE_LOAD.PKG_LOAD                       PACKAGE BODY         1.4        15
0 NO
QODE_APP.P_LOGIN                         PROCEDURE           75.2        15
0 NO
QODE_BC.PARTNER9                         PACKAGE              1.6        15
0 NO
QODE_APP.PKG_SPECIALOFFERS_OBJ           PACKAGE             35.6        14
0 NO
QODE_APP.P_SIGN_UP_ERROR_VALIDATE        PROCEDURE           17.6        13
0 NO
QODE_BC.PARTNER9                         PACKAGE BODY         1.4        13
0 NO
QODE_LOAD.PKG_PRODUCT_CATEGORY           PACKAGE BODY         2.6        13
0 NO
QODE_APP.PKG_LIST_MAINT                  PACKAGE             21.8        13
0 NO
QODE_BC.CATEGORY                         PACKAGE BODY         1.7        13
0 NO

-----Original Message-----
From: Philip West [mailto:P.West_at_g-icap.com] Sent: Thursday, August 24, 2000 7:27 AM
To: Multiple recipients of list ORACLE-L Subject: RE: Who and What is using the shared pool.

If your database has been up and running for a while I am VERY suprised that executions are 0 across the board. This is the way that I would always check for pinning candidates. Maybe someone else on the list could shed light. WHat versions are you running.

Phil West - Impex IT ltd
Unix Sys Admin and Oracle Financials DBA Services

all opinions are worth [at most] what you paid for them

-----Original Message-----
From: Koivu, Lisa [mailto:lkoivu_at_qode.com] Sent: 23 August 2000 21:58
To: Multiple recipients of list ORACLE-L Subject: RE: Who and What is using the shared pool.

Thanks Philip for sending this to the list.

However now I see that executions in v$db_object_cache is 0 everywhere. Is there an easy way to join to v$sqlarea to v$db_object_cache? I guess I was hoping to see if any heavily-used objects were being reloaded frequently and are candidates for pinning.

Anyone?? Thanks
Lisa
Ft. Lauderdale, FL, USA
*O*fficially *C*rabby and *P*eeved

-----Original Message-----
From: Philip West [ mailto:P.West_at_g-icap.com <mailto:P.West_at_g-icap.com> ] Sent: Monday, August 21, 2000 1:33 PM
To: Multiple recipients of list ORACLE-L Subject: RE: Who and What is using the shared pool.

Always keep to the shallow end of the shared pool. There are things swimming around in the depths that you do not want to meet!

Or, the problem is most likely shared pool fragmentation rather than the sizing per se. Fragmentation is a 'feature'. This is a big subject, which I would recommend you read up on (good papers on metalink on this subject. You do not tell us your db version. Ways to improve the situation include reserving a portion of the pool for larger objects. pinning critical or frequently reloaded objects into the pool. and, yes, increasing the size of the pool if necessary.

Look at v$sgastat to see what is where in your sga.

Try these (from the $AD_TOP/sql directory if you run Applications)

select owner || '.' || name OBJECT

, type
, to_char(sharable_mem/1024,'9,999.9') "SPACE(K)"
, loads
, executions execs
, kept

from v$db_object_cache
where type in ('FUNCTION','PACKAGE','PACKAGE BODY','PROCEDURE')   and owner not in ('SYS')
order by owner, name
/
select to_char(sum(sharable_mem)/1024,'9,999,999.9') "TOTAL SPACE (K)" from v$db_object_cache
where type in ('FUNCTION','PACKAGE','PACKAGE BODY','PROCEDURE')   and owner not in ('SYS')
/

also look here:        

  Bookmark Fixed font Go to End

Doc ID: Note:61623.1
Type: BULLETIN
Status: PUBLISHED
 Content Type: TEXT/PLAIN
Creation Date: 12-SEP-1997
Last Revision Date: 23-MAY-2000
Language: USAENG   

PURPOSE
  This document discusses some of the common issues associated with the shared
  pool in Oracle7 and describes how to diagnose and respond to these issues.

RELATED DOCUMENTS
  [NOTE:1012049.6] TUNING LIBRARY CACHE LATCH CONTENTION With Release 7.2 and 7.3, changes have been made to reduce usage of shared memory as well as per-user (UGA) memory. Also, memory is not being allocated
in large contiguous chunks, resulting in better shared-pool utilization and reduction in fragmentation.

  1. MEMORY FRAGMENTATION
The primary problem that occurs is that free memory in the shared pool becomes
fragmented into small pieces over time. Any attempt to allocate a large piece
of memory in the shared pool will cause large amount of objects in the library
cache to be flushed out and may result in an ORA-04031 out of shared memory error.
  1. DIAGNOSIS OF FRAGMENTATION
  2. ORA-04031 ERROR
One way to diagnose that this is happening is to look for ORA-04031 errors being returned from applications. When an attempt is made to allocate a large
contiguous piece of shared memory, and not enough contiguous memory can be created in the shared pool, the database will signal this error.

Before this error is signalled, all objects in the shared pool that are not currently in use will be flushed from the shared pool, and their memory will

be
freed and merged. This error only occurs when there is still not a large enough contiguous piece of free memory after this happens. There may be very
large amounts of total free memory in the shared pool, but just not enough contiguous memory.

ii) INIT.ORA PARAMETER

An init.ora parameter can be set so that whenever an ORA-04031 error is signalled a dump will occur into a trace file. By looking for these trace files, the DBA can determine that these errors are occurring. This is useful
when applications do not always report errors signalled by oracle, or if users
do not report the errors to the DBAs. The parameter is the following:

    event = "4031 trace name errorstack"

If you are using 7.0.16 or higher you can use the following:

    event = "4031 trace name errorstack level 4"

This will cause a dump of the Oracle state objects to occur when this error is
signalled. By looking in the dump for 'load=X' and then looking up a few lines
for 'name=' you can often tell whether an object was being loaded into the shared pool when this error occurred. If an object was being loaded then it

is
likely that this load is the cause of the problem and the object should be 'kept' in the shared pool. The object being loaded is the object printed after
the 'name='. Do not use the 'level 4' option in versions before 7.0.16 because
a bug existed that often caused the system to crash with this option enabled

due to a latch level violation.

Prior to version 7.3, there were a handful of cases where the RDBMS or PL/SQL
would attempt to allocate large pieces of contiguous memory. Most of this has
been fixed for 7.3. This problem was especially acute when running MTS, when

the UGA would be located in the SGA. This should also be fixed in 7.3 and using
MTS for a high OLTP scenario is recommended. As a result of all these changes,
the ORA-04031 error should be virtually eliminated. If an ORA-04031 error is

signalled, quite likely the shared pool is over 90% utilized and the alternative is to increase the shared pool. The only known situation is PL/SQL
packages (like STANDARD) where the package contains a very large number (over
400) procedure/function definitions. This still needs to be in contiguous memory and may request memory chunks as large as 15K. Packages like this should
be the only ones that should be kept.

iii) X$KSMLRU

There is a fixed table called x$ksmlru that tracks allocations in the shared pool that cause other objects in the shared pool to be aged out. This
fixed table can be used to identify what is causing the large allocation.

The columns of this fixed table are the following:

KSMLRCOM - allocation comment that describes the type of allocation.

If this comment is something like 'MPCODE' or 'PLSQL%' then there is a large

PL/SQL object being loaded into the shared pool. This PL/SQL object will need
to be 'kept' in the shared pool.

If this comment is 'kgltbtab' then the allocation is for a dependency table in
the library cache. This is only a problem when several hundred users are logged
on using distinct user ids. The solution in this case is to use fully qualified
names for all table references.

If you are running MTS and the comment is something like 'Fixed UGA' then the
problem is that the init.ora parameter 'open_cursors' is set too high.

KSMLRSIZ - amount of contiguous memory being allocated. Values over around 5K
start to be a problem, values over 10K are a serious problem, and values over
20K are very serious problems. Anything less then 5K should not be a problem.

KSMLRNUM - number of objects that were flushed from the shared pool in order

allocate the memory.

In release 7.1.3 or later, the following columns also exist:

KSMLRHON - the name of the object being loaded into the shared pool if the object is a PL/SQL object or a cursor.

KSMLROHV - hash value of object being loaded

KSMLRSES - SADDR of the session that loaded the object.

The advantage of X$KSMLRU is that it allows you to identify problems with fragmentation that are effecting performance, but that are not bad enough to

be
causing ORA-04031 errors to be signalled. If a lot of objects are being periodically flushed from the shared pool then this will cause response time

problems and will likely cause library cache latch contention problems when the
objects are reloaded into the shared pool. With version 7.2, the library cache
latch contention should be significantly reduced with the breaking up of the

library cache pin latch into a configurable set of symmetric library cache latches.

One unusual thing about the x$ksmlru fixed table is that the contents of the

fixed table are erased whenever someone selects from the fixed table. This is
done since the fixed table stores only the largest allocations that have occurred. The values are reset after being selected so that subsequent large
allocations can be noted even if they were not quite as large as others that

occurred previously. Because of this resetting, the output of selecting from
this table should be carefully noted since it cannot be reselected if it is forgotten. Also you should take care that there are not multiple people on one
database that select from this table because only one of them will select the
real data.

To monitor this fixed table just run the following:

    select * from x$ksmlru where ksmlrsiz > 5000;

iv) MTS

Oracle users using SQL*Net V2 can connect to the database using dedicated servers, or multiple clients can use a pool of shared (or MTS) servers. The biggest memory implication of this mode is that the session memory (also known
as the UGA) for every session needs to be accessible to every MTS server. This
implies that the logical UGA comes out of the physical SGA (or the shared pool) instead of the PGA (process memory).

In versions prior to 7.3, there were a few components in the UGA that would request large contiguous chunks of memory, contributing to fragmentation of the shared pool if using MTS. If the system had been up for a while, users would have failures when attempting to connect or executing sql. Starting with
7.3, all these allocations have been segmented such that the average size of

memory chunks allocated to the UGA should be about 5K.

B) CORRECTION OF FRAGMENTATION i) KEEPING OBJECTS

The primary source of problems is large PL/SQL objects. The means of correcting
these errors is to 'keep' large PL/SQL object in the shared pool at startup time. This will load the objects into the shared pool and will make sure that
the objects are never aged out of the shared pool. If the objects are never aged out then there will not be a problem with trying to load them and not having enough memory.

Objects are 'kept' in the shared pool using the dbms_shared_pool package that
is defined in the dbmspool.sql file. For example:

    execute dbms_shared_pool.keep('SYS.STANDARD');

All large packages that are shipped should be 'kept' if the customer uses PL/SQL. This includes 'STANDARD', 'DBMS_STANDARD', and 'DIUTIL'. With 7.3, the
only package left in this list is 'STANDARD'.

All large customer packages should also be marked 'kept'.

To mark all packages in the system 'kept' execute the following:

declare
  own varchar2(100);
  nam varchar2(100);
  cursor pkgs is
   select owner, object_name

     from dba_objects 
     where  object_type = 'PACKAGE'; 

begin
  open pkgs;
  loop
    fetch pkgs into own, nam;
    exit when pkgs%notfound;
    dbms_shared_pool.keep(own || '.' || nam, 'P');   end loop;
end;

The dbms_shared_pool package was introduced in 7.0 and has evolved over the versions. Until 7.1.5, 'keep' could only be used for packages. Starting with

7.1.6, this was extended to standalone procedures, cursors as well as triggers.
For detailed usage instructions, see the dbmspool.sql file. So, prior to this
version, if you have large procedures or large anonymous blocks, then these will need to be put into packages and marked kept. With 7.3, most packages do
not need to be kept any longer since PL/SQL no longer requires large amounts

of
contiguous memory to load packages/procedures in memory.

You can determine what large stored objects are in the shared pool by selecting
from the v$db_object_cache fixed view. This will also tell you which objects
have been marked kept. This can be done with the following query:

    select * from v$db_object_cache where sharable_mem > 10000;

Note that this query will not catch PL/SQ: objects that are only rarely used

and therefore the PL/SQL object is not currently loaded in the shared pool.

To determine what large PL/SQL objects are currently loaded in the shared pool
and are not marked 'kept' and therefore may cause a problem, execute the following:

    select name, sharable_mem
     from v$db_object_cache
    where sharable_mem > 10000

     and (type = 'PACKAGE' or type = 'PACKAGE BODY' or type = 'FUNCTION' 
          or type = 'PROCEDURE') 
     and kept = 'NO'; 

Another approach to the above is to use the dbms_shared_pool.sizes procedure.
To use this in SQLDBA:

    set serveroutput on;
    execute dbms_shared_pool.sizes(10);

This should show you the names of all the objects in the shared pool that take
more that 10K of memory as well as if they are marked kept or not. For SQL statements, if there are multiple versions of a query (usually a bug if the count is more than 3), they will also be indicated in parenthesis. Use the following query to check for problems:

  select sql_text, loaded_versions, version_count, sharable_mem    from v$sqlarea where loaded_versions > 3    order by sharable_mem;

In Oracle7.3 onwards the best candidates for keeping can be seen by querying

the table X$KSMSP to see if there are any chunks in the shared-pool that have
the KSMCHSIZ larger than 5K and KSMCHCOM like '%PL/SQL%'. If so then one can

identify the object name and owner of this chunk using the following SQL:

  select distinct
  decode(kglobtyp,0,'CURSOR',7,'PROCEDURE',8,'FUNCTION',9,'PACKAGE',

              11,'PACKAGE BODY',12,'TRIGGER',13,'TYPE',14,'TYPE BODY','OTHER')
   ||' - '||kglnaown||'.'||kglnaobj "Eligible PL/SQL objects"   from x$kglob
  where kglobhd4 in
    (select ksmchpar from x$ksmsp
      where ksmchcom='PL/SQL MPCODE' and ksmchsiz>5120)

If you are 'keeping' PL/SQL objects today and migrate to 7.3 or higher there

is
no need to re-assess the list of objects that you are keeping.

ii) USE BIND VARIABLES

One of the best things that can be done to reduce the amount of fragmentation
is to reduce or eliminate the number of sql statements in the shared pool that
are duplicates of each other except for a constant that is embedded in the statement. The statements should be replaced with one statement that uses a bind variable instead of a constant.

For example:

  select * from emp where empno=1; 
  select * from emp where empno=2; 
  select * from emp where empno=3; 

Should all be replaced with:

  select * from emp where empno=:1;

You can identify statements that potentially fall into this class with a query
like the following:

  select substr(sql_text, 1, 30) sql, count(*) copies     from v$sqlarea
   group by substr(sql_text, 1, 30)
   having count(*) > 3;

iii) MAX BIND SIZE

It is possible for a SQL statement to not be shared because the max bind variable lengths of the bind variables in the statement do not match. This is
automatically taken care of for precompiler programs and forms programs, but

could be a problem for programs that directly use OCI. The bind call in OCI

takes two arguments, one is the max length of the value, and the other is a pointer to the actual length. If the current length is always passed in as the
max length instead of the max possible length for the variable, then this could
cause the SQL statement not to be shared.

To identify statements that might potentially have this problem execute the following statement:

    select sql_text, version_count from v$sqlarea where version_count > 5;

Starting with 7.1.6 this should no longer be an issue as the server can graduate bind buffers even when the user's max bind lengths are jumping up or
down and continue to share cursors that are built for larger buffer lengths and flush the smaller sql compilation from the shared pool.

iv) ELIMINATING LARGE ANONYMOUS PL/SQL

Large anonymous PL/SQL blocks should be turned into small anonymous PL/SQL blocks that call packaged functions. The packages should be 'kept' in memory.
For version earlier that 7.3, this includes anonymous PL/SQL blocks that are

used for trigger definitions. With 7.3, triggers are compiled and stored to disk like standalone procedures and should be treated as such. Large anonymous
blocks can be identified with the following query:

  select sql_text from v$sqlarea
   where command_type=47 -- command type for anonymous block     and length(sql_text) > 500;

Note that this query will not catch PL/SQL blocks that are only rarely used and
therefore the PL/SQL block is not currently loaded in the shared pool.

Another option that can be used when an anonymous block cannot be turned into
a package is to mark the anonymous block with some string so that it can be identified in v$sqlarea and marked 'kept'.

For example, instead of using:
  declare x number; begin x := 5; end;;

you can use:
  declare /* KEEP_ME */ x number; begin x := 5; end;

You can then use the following procedure to select these statements out of the
shared pool and mark them 'kept' using the dbms_shared_pool.keep package.

declare
  /* DONT_KEEP_ME */
  addr varchar2(10);
  hash number;
  cursor anon is
   select address, hash_value

     from v$sqlarea 
    where command_type = 47               -- command type for anonymous 
block 
     and  sql_text like '% KEEP_ME %' 
     and  sql_text not like '%DONT_KEEP_ME%'; 
begin
  open anon;
  loop
    fetch anon into addr, hash;
    exit when anon%notfound;
    dbms_shared_pool.keep(addr || ',' || to_char(hash), 'C');   end loop;
end;

v) REDUCING USAGE

Another way to reducing fragmentation is to reduce consumption. This is of special importance when using MTS, when every user's session memory is in the
shared pool and the impact is multiplied by the total concurrent users.

Insert, update, delete and anonymous blocks complete the execution in one round
trip. All the memory that is allocated on the server for the execute comes from
the PGA and is freed before the call returns to the user. But in the case of

selects, memory required to execute the statement - which could be large if a
sort was involved - is not freed until the end-of-fetch is reached or the query
is cancelled. In these situations using the OCI features to do an exact fetch
and cancel helps free memory back to the pool.

If the application logic has been embedded into server side PL/SQL, a large number of cursors may be getting cached on the server for every user. Though

this results in reduced latch contention and faster response, it does use more
memory in the UGA. Setting the close_cached_open_cursors init.ora to TRUE closes the PL/SQL cached cursors on the server, freeing the memory.


2) COMMON FALLACIES There are a number of common fallacies about the shared pool that are often stated as fact.

  1. FREE MEMORY
One fallacy is that the amount of 'free memory' reported in v$sgastat needs to
be kept high. This is incorrect. The free memory reported in this table is

not
like the free memory reported by operating system statistics. Since the shared
pool acts as a cache, nothing will ever be aged out of the shared pool until

all
the free memory has been used up. This is entirely normal.

Free memory is more properly thought of as 'wasted memory'. You would rather
see this value be low than very high. In fact, a high value of free memory is
sometimes a symptom that a lot of objects have been aged out of the shared pool
and therefore the system is experiencing fragmentation problems.

B) FLUSH SHARED POOL Some people think that frequently executing 'alter system flush shared_pool'

improves the performance of the system and decreases the amount of fragmentation. This is incorrect. Executing this statement causes a big spike
in performance and does nothing to improve fragmentation.

The only time when it might be useful to run this statement is between shifts of
users so that the objects that are relevant to the last shift of users can be
flushed out before the next shift of users starts to use the system. This is
almost never needed though.


3) SIZING OF SHARED POOL One very difficult judgement that needs to be make in Oracle7 is to determine
the proper size of the shared pool. The following provides some guidelines for
this. It should be emphasized that these are just guidelines, there are no hard and fast rules here and experimentation will be needed to determine a good
value.

The shared pool size is highly application dependent. To determine the shared
pool size that will be needed for a production system it is generally necessary
to first develop the application and run it on a test system and take some measurements. The test system should be run with a very large value for the

shared pool size to make the measurements meaningful.

  1. OBJECTS STORED IN THE DATABASE
The amount of shared pool that needs to be allocated for objects that are stored in the database like packages and views is easy to measure. You can just measure their size directly with the following statement:

  select sum(sharable_mem) from v$db_object_cache;

This is especially effective because all large pl/sql object should be 'kept'
in the shared pool at all times.

B) SQL The amount of memory needed to store SQL statements in the shared pool is more
difficult to measure because of the needs of dynamic SQL. If an application

has no dynamic SQL then the amount of memory can simply be measured after the
application has run for a while by just selecting it out of the shared pool as
follows:

    select sum(sharable_mem) from v$sqlarea;

If the application has a moderate or large amount of dynamic SQL like most applications do, then a certain amount of memory will be needed for the shared
SQL, plus more for the dynamic SQL, and more so that the dynamic SQL does not
age the shared SQL out of the shared pool.

The amount of memory for the shared SQL can be approximated by the following:

    select sum(sharable_mem) from v$sqlarea where executions > 5;

The remaining memory in v$sqlarea is for dynamic SQL Some shared pool will need to be budgeted for this also, but there are few rules here.

C) PER-USER PER-CURSOR MEMORY You will need to allow around 250 bytes of memory in the shared pool per concurrent user for each open cursor that the user has whether the cursor is

shared or not. During the peak usage time of the production system, you can

measure this as follows:

    select sum(250 * users_opening) from v$sqlarea;

In a test system you can measure it by selecting the number of open cursors for a test user and multiplying by the total number of users:

  select 250 * value bytes_per_user
    from v$sesstat s, v$statname n

   where s.statistic# = n.statistic# 
    and  n.name = 'opened cursors current' 
    and  s.sid  =  23;    -- replace 23 with session id of user being 
measured

The per-user per-cursor memory is one of the classes of memory that shows up

as
'library cache' in v$sgastat.

D) MTS If you are using multi-threaded server, then you will need to allow enough memory for all the shared server users to put their session memory in the shared pool. This can be measured for one user with the following query:

  select value sess_mem
    from v$sesstat s, v$statname n

   where s.statistic# = n.statistic# 
    and  n.name = 'session uga memory' 
    and  s.sid  =  23;    -- replace 23 with session id of user being 
measured

A more conservative value to use is the maximum session memory that was ever allocated by the user:

  select value sess_max_mem
    from v$sesstat s, v$statname n

   where s.statistic# = n.statistic# 
    and  n.name = 'session uga memory max' 
    and  s.sid  =  23;    -- replace 23 with session id of user being 
measured

To select this value for all the currently logged on users the following query
can be used:

  select sum(value) all_sess_mem
    from v$sesstat s, v$statname n
   where s.statistic# = n.statistic#
    and n.name = 'session uga memory max';

E) OVERHEAD You will need to add a minimum of 30% overhead to the values calculated above
to allow for unexpected and unmeasured usage of the shared pool.


4) FINAL COMMENTS The most important point that needs to be understood by everyone using Oracle7
and PL/SQL (prior to release 7.3) is that all large PL/SQL objects must be made
into packages and those packages must be kept in the shared pool. This point

cannot be over emphasized. Many customers, especially those running a lot of
users, have had terrible performance problems that were completely cleared up
by doing this.

                  APPENDIX I:  Reserved Shared Pool 
                  ================================= 

  1. RESERVED SPACE FROM THE SHARED POOL
    On busy systems, the RDBMS may have difficulty finding a contiguous piece of

memory to satisfy a large request for memory. Because the RDBMS will search

for and free currently unused memory, the search for this large piece of memory
may disrupt the behavior of the share pool, leading to more fragmentation and
poor performance.

RDBMS 7.1.5 allows DBAs to reserve memory within the shared pool to satisfy these large allocations during RDBMS operations such as PL/SQL compilation and
trigger compilation. Smaller objects will not fragment the reserved list, helping to ensure the reserved list will have large contiguous chunks of memory.
Once the memory allocated from the reserved list is freed, it returns to the

reserved list.

The size of the reserved list, as well as the minimum size of the objects that
can be allocated from the reserved list are controlled via init.ora parameters:
shared_pool_reserved_size and shared_pool_reserved_min_alloc.

1.1 shared_pool_reserved_size



The init.ora parameter shared_pool_reserved_size controls the amount of shared_pool_size reserved for large allocations. In order to create a reserved
list, shared_pool_reserved_size must be greater than shared_pool_reserved_min_alloc.
  units  :  bytes 
  default:  0  (no reserved list) 
  minimum:  > shared_pool_reserved_min_alloc 
  maximum:  1/2 shared_pool_size 

1.2 shared_pool_reserved_min_alloc



The init.ora parameter shared_pool_reserved_min_alloc controls allocation for
the reserved memory. Only allocations larger than shared_pool_reserved_min_alloc are allowed to allocate space from the reserved
list if a chunk of memory of sufficient size is not found on the shared pool's
free lists.
  units  :  bytes 
  default:  5000 
  minimum:  5000 
  maximum:  < shared_pool_reserved_size 

The default value for shared_pool_reserved_min_alloc should be adequate for almost all systems.

2. CONTROLLING SPACE RECLAMATION OF THE SHARED POOL



RDBMS 7.1.5 also provides a new procedure, aborted_request_threshold, in package
dbms_shared_pool, which allows users to set the limit on the size of allocations
allowed to flush the shared pool if the free lists cannot satisfy the request
size.

Before the RDBMS signals the ORA-04031 error, it incrementally flushes unused
objects from the shared pool until there is sufficient memory to satisfy the

allocation request. In most cases, incrementally flushing objects from the shared pool frees enough memory for the allocation to complete succesfully. If
the RDBMS signals an ORA-04031 error, it has flushed all objects currently not
in use on the system without finding a large enough piece of contiguous memory.

On a busy system, the larger the space allocation, the more likely the RDBMS

will signal the ORA-04031 error. Flushing all objects, however, will impact

other users on the system, possibly causing a degradation in performance.

The aborted_request_threshold procedure allows the DBA to localize the impact
the ORA-04031 error to the process that couldn't allocate memory. The procedure
takes a numeric value between 5000 and 2147483647, representing the size, in

bytes, of the threshold.

3. NEW FIXED VIEW V$SHARED_POOL_RESERVED



RDBMS 7.1.5 has a new fixed view to help tune the reserved pool and space within
the shared pool. The name of the new fixed view is V$SHARED_POOL_RESERVED and
has the following columns:
Name                            Null?    Type 
------------------------------- -------- -------------- 
FREE_SPACE                               NUMBER 
AVG_FREE_SIZE                            NUMBER 
FREE_COUNT                               NUMBER 
MAX_FREE_SIZE                            NUMBER 
USED_SPACE                               NUMBER 
AVG_USED_SIZE                            NUMBER 
USED_COUNT                               NUMBER 
MAX_USED_SIZE                            NUMBER 
REQUESTS                                 NUMBER 
REQUEST_MISSES                           NUMBER 
LAST_MISS_SIZE                           NUMBER 
MAX_MISS_SIZE                            NUMBER 
REQUEST_FAILURES                         NUMBER 
LAST_FAILURE_SIZE                        NUMBER 
ABORTED_REQUEST_THRESHOLD                NUMBER 
ABORTED_REQUESTS                         NUMBER 
LAST_ABORTED_SIZE                        NUMBER 

These columns of V$SHARED_POOL_RESERVED are only valid if the parameter shared_pool_reserved_size is set to a valid value.

  FREE_SPACE is the total amount of free space on the reserved list.   

  AVG_FREE_SIZE is the average size of the free memory on the reserved list.

  FREE_COUNT is the number of free pieces of memory on the reserved list.

  MAX_FREE_SIZE is the size of the largest free piece of memory on the reserved

                list. 

  USED_SPACE is the total amount of used memory on the reserved list.

  AVG_USED_SIZE is the average size of the of the used memory on the reserved

                list. 

  USED_COUNT is the number of used pieces of memory on the reserved list.

  MAX_USED_SIZE is the size of the largest used piece of memory on the reserved

                list. 

  REQUESTS is the number of times that the reserved list was searched for a 
           free piece of memory. 

  REQUEST_MISSES is the number of times the reserved list didn't have a free

                 piece of memory to satisfy the request, and proceeded to 
start 
                 flushing objects from the LRU list. 

  LAST_MISS_SIZE is the request size of the last REQUEST_MISS.

  MAX_MISS_SIZE is the request size of the largest REQUEST_MISS.

The next set of columns contain values which are valid even if shared_pool_reserved_size is not set.

  REQUEST_FAILURES is the number of times that no memory was found to satisfy a

                   request (example: number of times ORA-04031 occurred) 

  LAST_FAILURE_SIZE is the request size of the last failed request 
                    (example: the request size of last ORA-04031). 

  ABORTED_REQUEST_THRESHOLD is the minimum size of a request which will signal

                            an ORA-04031 error without flushing objects. See


                            the procedure aborted_request_threshold 
described 
                            above. 

  LAST_ABORTED_SIZE is the last size of the request which returned an ORA-04031

                    error without flushing objects from the LRU list. 


4. TUNING HINTS BASED ON V$SHARED_POOL_RESERVED



Information in V$SHARED_POOL_RESERVED can help to set values for shared_pool_reserved_size and even shared_pool_size. This section assumes the
DBA has performed all other shared pool tuning on his system.

4.1 Initial Value for shared_pool_reserved_size



The DBA should make shared_pool_reserved_size 10% of the shared_pool_size. For
most systems, this value should be sufficient, if the DBA has already spent time
tuning the shared pool.

4.2 Initial Value for shared_pool_reserved_min_alloc



In most cases, the default value for this parameter is adequate. If the DBA increases this value, then the RDBMS will allow fewer allocations from the reserved list and will request more memory from the shared pool list.

4.4 Tuning shared_pool_reserved_size



Ideally, shared_pool_reserved_size should be made large enough to satisfy any
request scanning for memory on the reserved list without flushing objects from
the shared pool. The amount of operating system memory, however, may constrain
the size of the SGA, and therefore the size of the shared pool such that this
is not a feasible goal.

If the DBA has a system with ample free memory to increase his SGA, the goal

is
to have:

   REQUEST_MISS = 0 If the DBA is constrained for OS memory, his goal is:

   REQUEST_FAILURES = 0 or not increasing    LAST_FAILURE_SIZE > shared_pool_reserved_min_alloc    AVG_FREE_SIZE > shared_pool_reserved_min_alloc

If neither of these goals are met, increase shared_pool_reserved_size; the DBA
also needs to increase shared_pool_size by the same amount, since the reserved
list is taken from the shared pool.

4.5 shared_pool_reserved_size too low



The reserved pool is too small when:

   REQUEST_FAILURES > 0 (and increasing)

and at least one of the following is true:

   LAST_FAILURE_SIZE > shared_pool_reserved_min_alloc 
   MAX_FREE_SIZE     < shared_pool_reserved_min_alloc 
   FREE_MEMORY       < shared_pool_reserved_min_alloc 

The DBA has two options, depending on his SGA size constraints:

   o Increase shared_pool_reserved_size and shared_pool_size, accordingly    o Increase shared_pool_reserved_min_alloc (but may need to increase

      shared_pool_size)

The first option will increase the amount of memory available on the reserved
list without impacting users not allocating memory from the reserved list. The
second options reduces the number of allocations allowed to use memory from the
reserved list; doing so, however, will increase normal shared pool perhaps impacting other users on the system.

4.6 shared_pool_reserved_size too high



It is possible that too much memory has been allocated to the reserved list.

If:

   REQUEST_MISS      = 0 or not increasing 
   FREE_MEMORY       = > 50% of shared_pool_reserved_size minimum 

The DBA has two options:

   o Decrease shared_pool_reserved_size    o Decrease shared_pool_reserved_min_alloc (if not the default

      value)

4.7 shared_pool_size too small



The new fixed table can also indicate when shared_pool_size is too small. If:

   REQUEST_FAILURES > 0 and increasing    LAST_FAILURE_SIZE < shared_pool_reserved_min_alloc

Then the DBA has two options if he has enabled the reserved list:

   o Decrease shared_pool_reserved_size    o Decrease shared_pool_reserved_min_alloc (if set larger than the default)

Otherwise, the DBA the could:

   o Increase shared_pool_size

                  APPENDIX 2:  Procedure free_unused_memory 
                  ========================================= 

This text is also in the specification for this procedure in dbmsutil.sql. It is part of package dbms_session.

Procedure free_unused_memory --

Procedure for users to reclaim unused memory after performing operations requiring large amounts of memory (where large is >100K). Note that this procedure should only be used in cases where memory is at a premium.

Examples operations using lots of memory are:

   o large sorts where entire sort_area_size is used and

      sort_area_size is hundreds of KB
   o compiling large PL/SQL packages, procedures, or functions    o storing hundreds of KB of data within PL/SQL indexed tables

One can monitor user memory by tracking the statistics "session uga memory" and
"session pga memory" in the v$sesstat/v$statname fixed views. Monitoring these
statistics will also show how much memory this procedure has freed. The behavior
of this procedure depends upon the configuration of the server operating on behalf of the client:

   o dedicated server - returns unused PGA memory to the OS    o MTS server - returns unused session memory to the shared_pool

In order to free memory using this procedure, the memory must not be in use.

Once an operation allocates memory, only the same type of operation can reuse
the allocated memory. For example, once memory is allocated for sort, even if
the sort is complete and the memory is no longer in use, only another sort can
reuse the sort-allocated memory. For both sort and compilation, after the operation is complete, the memory is no longer in use and the user can invoke
this procedure to free the unused memory.

An indexed table implicitly allocates memory to store values assigned to the

indexed table's elements. Thus, the more elements in an indexed table, the more
memory the RDBMS allocates to the indexed table. As long as there are elements
within the indexed table, the memory associated with an indexed table is in use.

The scope of indexed tables determines how long their memory is in use. Indexed
tables declared globally are indexed tables declared in packages or package bodies. They allocate memory from session memory. For an indexed table declared globally, the memory will remain in use for the lifetime of a user's
login (lifetime of a user's session), and is freed after the user disconnects
from ORACLE.

Indexed tables declared locally are indexed tables declared within functions,
procedures, or anonymous blocks. These indexed tables allocate memory from PGA
memory. For an indexed table declared locally, the memory will remain in use
for as long as the user is still executing the procedure, function, or anonymous
block in which the indexed table is declared. After the procedure, function, or
anonymous block is finished executing, the memory is then available for other
locally declared indexed tables to use (i.e., the memory is no longer in use).

Assigning an uninitialized, "empty," indexed table to an existing index table is
a method to explicitly re-initialize the indexed table and the memory associated
with the indexed table. After this operation, the memory associated with the
indexed table will no longer be in use, making it available to be freed by calling this procedure. This method is particularly useful on indexed tables
declared globally which can grow during the lifetime of a user's session, as

long as the user no longer needs the contents of the indexed table.

The memory rules associated with an indexed table's scope still apply; this method and this procedure, however, allow users to intervene and to explictly
free the memory associated with an indexed table.

The PL/SQL fragment below illustrates the method and the use of procedure free_unused_user_memory.

 create package foobar
   type number_idx_tbl is table of number indexed by binary_integer;

   store1_table  number_idx_tbl;     --  PL/SQL indexed table 
   store2_table  number_idx_tbl;     --  PL/SQL indexed table 
   store3_table  number_idx_tbl;     --  PL/SQL indexed table 
   ... 
 end;            --  end of foobar 

 declare
   ...
   empty_table number_idx_tbl; -- uninitialized ("empty") version  begin
   for i in 1..1000000 loop
     store1_table(i) := i; -- load data    end loop;
   ...
   store1_table := empty_table; -- "truncate" the indexed table    ...
   -
   dbms_session.free_unused_user_memory; -- give memory back to system
   store1_table(1) := 100;           --  index tables still declared; 
   store2_table(2) := 200;           --  but truncated. 
   ...
  end;

 Copyright (c) 1995,1999 Oracle Corporation. All Rights Reserved. Legal Notices and Terms of Use.

Phil West - Impex IT ltd
Unix Sys Admin and Oracle Financials DBA Services

all opinions are worth [at most] what you paid for them



***
The information in this Internet e-mail is confidential and may be legally privileged. It is intended solely for the addressee. Access to this Internet e-mail by anyone else is unauthorised and any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful.

 When addressed to our clients any opinions or advice contained in this Internet e-mail are subject to the terms and conditions expressed in any applicable documentation or market practices governing the relationship between Garban Intercapital plc and its clients.

Any views expressed in this message are those of the individual sender except where they are stated to be the views of Garban Intercapital plc.



***
-- 
Author: Philip West 
  INET: P.West_at_g-icap.com 

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051 
San Diego, California        -- Public Internet access / Mailing Lists 
-------------------------------------------------------------------- 
To REMOVE yourself from this mailing list, send an E-Mail message 
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in 
the message BODY, include a line containing: UNSUB ORACLE-L 
(or the name of mailing list you want to be removed from).  You may 
also send the HELP command for other information (like subscribing). 


------_=_NextPart_001_01C00DE3.CFD168DE
Content-Type: text/html;
	charset="iso-8859-1"

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<TITLE>RE: Who and What is using the shared pool.</TITLE>

<META content="MSHTML 5.00.2314.1000" name=GENERATOR></HEAD>
<BODY>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=343393315-24082000>Lisa,</SPAN></FONT></DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN 
class=343393315-24082000></SPAN></FONT>&nbsp;</DIV>
<DIV><FONT color=#0000ff face=Arial size=2><SPAN class=343393315-24082000>I am 
baffled.&nbsp; I do not have access to 8.x at the moment.&nbsp; Is there anyone 
out there who can help Lisa?&nbsp; </SPAN></FONT></DIV>
<DIV>&nbsp;</DIV>
<P><FONT face=Arial>Phil West - Impex IT ltd</FONT> <BR><FONT 
face="Arial Narrow" size=1>Unix Sys Admin and Oracle Financials DBA 
Services</FONT> </P>
<P><FONT face="Arial Narrow" size=1>all opinions are worth [at most] what you 
paid for them</FONT> </P>
<BLOCKQUOTE style="MARGIN-RIGHT: 0px">
  <DIV align=left class=OutlookMessageHeader dir=ltr><FONT face=Tahoma 
  size=2>-----Original Message-----<BR><B>From:</B> Koivu, Lisa 
  [mailto:lkoivu_at_qode.com]<BR><B>Sent:</B> 24 August 2000 16:52<BR><B>To:</B> 
  Multiple recipients of list ORACLE-L<BR><B>Subject:</B> RE: Who and What is 
  using the shared pool.<BR><BR></DIV></FONT>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN class=473444114-24082000>Hi 
  Phil, </SPAN></FONT></DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000>8.1.5 on HP/UX 11.0.&nbsp; Check it out below.&nbsp; 
  I figured I would have to look at executions to determine if something that is 
  constantly being reloaded really needs to be pinned.&nbsp; Seems to me if it's 
  being executed a lot, and reloaded a lot, it should be pinned - and if it's 
  being reloaded a lot but not executed more than once per load, it's not worth 
  the effort. </SPAN></FONT></DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000>Thanks in advance for any comments. 
  </SPAN></FONT></DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000>Lisa</SPAN></FONT></DIV>
  <DIV><FONT color=#0000ff face=Arial size=2><SPAN 
  class=473444114-24082000></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face="Courier New" size=2><SPAN 
  class=473444114-24082000>SQL&gt; select startup_time&nbsp; from 
  v$instance;</SPAN></FONT></DIV>
  <DIV>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face="Courier New" size=2><SPAN 
  class=473444114-24082000>STARTUP_T<BR>---------<BR>02-AUG-00</SPAN></FONT></DIV>
  <DIV><FONT color=#0000ff><SPAN 
  class=473444114-24082000></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face="Courier New" size=2><SPAN 
  class=473444114-24082000>&nbsp; 1&nbsp; select owner || '.' || name 
  OBJECT<BR>&nbsp; 2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , 
  type<BR>&nbsp; 3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , 
  to_char(sharable_mem/1024,'9,999.9') "SPACE(K)"<BR>&nbsp; 
  4&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , loads<BR>&nbsp; 
  5&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , executions 
  execs<BR>&nbsp; 6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , 
  kept<BR>&nbsp; 7&nbsp; from v$db_object_cache<BR>&nbsp; 8&nbsp; where type in 
  ('FUNCTION','PACKAGE','PACKAGE BODY','PROCEDURE')<BR>&nbsp; 
  9&nbsp;&nbsp;&nbsp; and owner not in ('SYS')<BR>&nbsp;10* order by loads 
  desc<BR>SQL&gt; /</SPAN></FONT></DIV>
  <DIV>&nbsp;</DIV>
  <DIV><FONT color=#0000ff face="Courier New" size=2><SPAN 
  class=473444114-24082000>OBJECT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  TYPE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  SPACE(K)&nbsp;&nbsp;&nbsp;&nbsp; LOADS&nbsp;&nbsp;&nbsp;&nbsp; EXECS 
  KEP<BR>---------------------------------------- --------------- -------- 
  --------- --------- 
  ---<BR>QODE_LOAD.F_VALIDATE_ISBN&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  FUNCTION&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.9&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  31&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PKG_GN_SPECIAL_OFFERS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  12.3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  29&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.CLEANUP_PARTNERNAME_STG2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  2.8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  25&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.PKG_GENERATE_KEYWORD&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  3.1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  21&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.PKG_UTIL&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  28.0&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  17&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.PKG_CLEANUP&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  9.3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  17&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PKG_GN_SPECIAL_OFFERS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  11.3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  16&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.CATHIERARCHY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.RPT_PARTNER_EMAIL&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PARTNER1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.P_GUESTHOME&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  73.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_USER.WEBDBTEST012345678912&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PARTRELATIONSHIP&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PARTNER_RELATIONSHIP1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.CATHIERARCHY2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.P_BUYQODER&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  38.8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_LOAD.PKG_LOAD&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.4&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.P_LOGIN&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  75.2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PARTNER9&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.PKG_SPECIALOFFERS_OBJ&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  35.6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  14&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.P_SIGN_UP_ERROR_VALIDATE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PROCEDURE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  17.6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.PARTNER9&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.4&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_LOAD.PKG_PRODUCT_CATEGORY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  2.6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_APP.PKG_LIST_MAINT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  21.8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 
  NO<BR>QODE_BC.CATEGORY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  PACKAGE BODY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  1.7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
  13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0 NO</SPAN></FONT></DIV>
  <BLOCKQUOTE style="MARGIN-RIGHT: 0px">
    <DIV align=left class=OutlookMessageHeader dir=ltr><FONT face=Tahoma 
    size=2>-----Original Message-----<BR><B>From:</B> Philip West 
    [mailto:P.West_at_g-icap.com]<BR><B>Sent:</B> Thursday, August 24, 2000 7:27 
    AM<BR><B>To:</B> Multiple recipients of list ORACLE-L<BR><B>Subject:</B> RE: 
    Who and What is using the shared pool.<BR><BR></DIV></FONT>
    <DIV><FONT color=#0000ff face=Arial size=2><SPAN class=781273408-24082000>If 
    your database has been up and running for a while I am VERY suprised that 
    executions are 0 across the board.&nbsp; This is the way that I would always 
    check for pinning candidates.&nbsp; Maybe someone else on the list could 
    shed light.&nbsp; WHat versions are you running.</SPAN></FONT></DIV>
    <P><FONT face=Arial>Phil West - Impex IT ltd</FONT> <BR><FONT 
    face="Arial Narrow" size=1>Unix Sys Admin and Oracle Financials DBA 
    Services</FONT> </P>
    <P><FONT face="Arial Narrow" size=1>all opinions are worth [at most] what 
    you paid for them</FONT> </P>
    <BLOCKQUOTE style="MARGIN-RIGHT: 0px">
      <DIV align=left class=OutlookMessageHeader dir=ltr><FONT face=Tahoma 
      size=2>-----Original Message-----<BR><B>From:</B> Koivu, Lisa 
      [mailto:lkoivu_at_qode.com]<BR><B>Sent:</B> 23 August 2000 
      21:58<BR><B>To:</B> Multiple recipients of list 
      ORACLE-L<BR><B>Subject:</B> RE: Who and What is using the shared 
      pool.<BR><BR></DIV></FONT>
      <P><FONT size=2>Thanks Philip for sending this to the list.&nbsp; 
      </FONT></P>
      <P><FONT size=2>However now I see that executions in v$db_object_cache is 
      0 everywhere.&nbsp; Is there an easy way to join to v$sqlarea to 
      v$db_object_cache?&nbsp; I guess I was hoping to see if any heavily-used 
      objects were being reloaded frequently and are candidates for 
      pinning.</FONT></P>
      <P><FONT size=2>Anyone??&nbsp; Thanks</FONT> <BR><FONT size=2>Lisa</FONT> 
      <BR><FONT size=2>Ft. Lauderdale, FL, USA</FONT> <BR><FONT 
      size=2>*O*fficially *C*rabby and *P*eeved</FONT> </P>
      <P><FONT size=2>-----Original Message-----</FONT> <BR><FONT size=2>From: 
      Philip West [<A 
      href="mailto:P.West_at_g-icap.com">mailto:P.West_at_g-icap.com</A>]</FONT> 
      <BR><FONT size=2>Sent: Monday, August 21, 2000 1:33 PM</FONT> <BR><FONT 
      size=2>To: Multiple recipients of list ORACLE-L</FONT> <BR><FONT 
      size=2>Subject: RE: Who and What is using the shared pool.</FONT> </P><BR>
      <P><FONT size=2>Always keep to the shallow end of the shared pool.&nbsp; 
      There are things</FONT> <BR><FONT size=2>swimming around in the depths 
      that you do not want to meet!</FONT> </P>
      <P><FONT size=2>Or,&nbsp; the problem is most likely shared pool 
      fragmentation rather than the</FONT> <BR><FONT size=2>sizing per se.&nbsp; 
      Fragmentation is a 'feature'.&nbsp; This is a big subject, which</FONT> 
      <BR><FONT size=2>I would recommend you read up on (good papers on metalink 
      on this subject.</FONT> <BR><FONT size=2>You do not tell us your db 
      version.&nbsp; Ways to improve the situation include</FONT> <BR><FONT 
      size=2>reserving a portion of the pool for larger objects.&nbsp; pinning 
      critical or</FONT> <BR><FONT size=2>frequently reloaded objects into the 
      pool. and, yes, increasing the size of</FONT> <BR><FONT size=2>the pool if 
      necessary.</FONT> </P>
      <P><FONT size=2>Look at v$sgastat to see what is where in your sga.</FONT> 
      </P>
      <P><FONT size=2>Try these (from the $AD_TOP/sql directory if you run 
      Applications)</FONT> </P>
      <P><FONT size=2>select owner || '.' || name OBJECT</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , type</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , 
      to_char(sharable_mem/1024,'9,999.9') "SPACE(K)"</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , loads</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , executions 
      execs</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; , 
      kept</FONT> <BR><FONT size=2>from v$db_object_cache</FONT> <BR><FONT 
      size=2>where type in ('FUNCTION','PACKAGE','PACKAGE 
      BODY','PROCEDURE')</FONT> <BR><FONT size=2>&nbsp; and owner not in 
      ('SYS')</FONT> <BR><FONT size=2>order by owner, name</FONT> <BR><FONT 
      size=2>/</FONT> <BR><FONT size=2>select 
      to_char(sum(sharable_mem)/1024,'9,999,999.9') "TOTAL SPACE (K)"</FONT> 
      <BR><FONT size=2>from v$db_object_cache</FONT> <BR><FONT size=2>where type 
      in ('FUNCTION','PACKAGE','PACKAGE BODY','PROCEDURE')</FONT> <BR><FONT 
      size=2>&nbsp; and owner not in ('SYS')</FONT> <BR><FONT size=2>/</FONT> 
      </P><BR>
      <P><FONT size=2>also look here:</FONT> <BR><FONT size=2>&nbsp; 
      </FONT><BR><FONT size=2>&nbsp;&nbsp;&nbsp; </FONT><BR><FONT size=2>&nbsp; 
      Bookmark Fixed font&nbsp; Go to End </FONT></P>
      <P><FONT size=2>Doc ID:&nbsp; Note:61623.1 </FONT><BR><FONT 
      size=2>Type:&nbsp; BULLETIN </FONT><BR><FONT size=2>Status:&nbsp; 
      PUBLISHED </FONT><BR><FONT size=2>&nbsp;Content Type:&nbsp; TEXT/PLAIN 
      </FONT><BR><FONT size=2>Creation Date:&nbsp; 12-SEP-1997 </FONT><BR><FONT 
      size=2>Last Revision Date:&nbsp; 23-MAY-2000 </FONT><BR><FONT 
      size=2>Language:&nbsp; USAENG </FONT><BR><FONT size=2>&nbsp;</FONT> </P>
      <P><FONT size=2>PURPOSE</FONT> <BR><FONT size=2>&nbsp; This document 
      discusses some of the common issues associated with the</FONT> <BR><FONT 
      size=2>shared </FONT><BR><FONT size=2>&nbsp; pool in Oracle7 and describes 
      how to diagnose and respond to these issues.</FONT> </P>
      <P><FONT size=2>RELATED DOCUMENTS</FONT> <BR><FONT size=2>&nbsp; 
      [NOTE:1012049.6]&nbsp;&nbsp; TUNING LIBRARY CACHE LATCH CONTENTION</FONT> 
      </P><BR>
      <P><FONT size=2>With Release 7.2 and 7.3, changes have been made to reduce 
      usage of shared </FONT><BR><FONT size=2>memory as well as per-user (UGA) 
      memory.&nbsp; Also, memory is not being</FONT> <BR><FONT size=2>allocated 
      </FONT><BR><FONT size=2>in large contiguous chunks, resulting in better 
      shared-pool utilization and </FONT><BR><FONT size=2>reduction in 
      fragmentation.</FONT> </P>
      <P><FONT size=2>1) MEMORY FRAGMENTATION</FONT> </P>
      <P><FONT size=2>The primary problem that occurs is that free memory in the 
      shared pool</FONT> <BR><FONT size=2>becomes </FONT><BR><FONT 
      size=2>fragmented into small pieces over time.&nbsp; Any attempt to 
      allocate a large</FONT> <BR><FONT size=2>piece </FONT><BR><FONT size=2>of 
      memory in the shared pool will cause large amount of objects in the</FONT> 
      <BR><FONT size=2>library</FONT> <BR><FONT size=2>cache to be flushed out 
      and may result in an ORA-04031 out of shared memory </FONT><BR><FONT 
      size=2>error.</FONT> </P>
      <P><FONT size=2>A) DIAGNOSIS OF FRAGMENTATION</FONT> </P>
      <P><FONT size=2>i) ORA-04031 ERROR</FONT> </P>
      <P><FONT size=2>One way to diagnose that this is happening is to look for 
      ORA-04031 errors</FONT> <BR><FONT size=2>being returned from 
      applications.&nbsp; When an attempt is made to allocate a</FONT> <BR><FONT 
      size=2>large </FONT><BR><FONT size=2>contiguous piece of shared memory, 
      and not enough contiguous memory can be </FONT><BR><FONT size=2>created in 
      the shared pool, the database will signal this error.</FONT> </P>
      <P><FONT size=2>Before this error is signalled, all objects in the shared 
      pool that are not</FONT> <BR><FONT size=2>currently in use will be flushed 
      from the shared pool, and their memory will</FONT> <BR><FONT 
      size=2>be</FONT> <BR><FONT size=2>freed and merged.&nbsp; This error only 
      occurs when there is still not a large </FONT><BR><FONT size=2>enough 
      contiguous piece of free memory after this happens.&nbsp; There may 
      be</FONT> <BR><FONT size=2>very </FONT><BR><FONT size=2>large amounts of 
      total free memory in the shared pool, but just not enough </FONT><BR><FONT 
      size=2>contiguous memory.</FONT> </P>
      <P><FONT size=2>ii) INIT.ORA PARAMETER</FONT> </P>
      <P><FONT size=2>An init.ora parameter can be set so that whenever an 
      ORA-04031 error is</FONT> <BR><FONT size=2>signalled a dump will occur 
      into a trace file.&nbsp; By looking for these trace </FONT><BR><FONT 
      size=2>files, the DBA can determine that these errors are occurring.&nbsp; 
      This is</FONT> <BR><FONT size=2>useful</FONT> <BR><FONT size=2>when 
      applications do not always report errors signalled by oracle, or if</FONT> 
      <BR><FONT size=2>users</FONT> <BR><FONT size=2>do not report the errors to 
      the DBAs.&nbsp; The parameter is the following:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; event = "4031 trace name 
      errorstack"</FONT> </P>
      <P><FONT size=2>If you are using 7.0.16 or higher you can use the 
      following:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; event = "4031 trace name errorstack 
      level 4"</FONT> </P>
      <P><FONT size=2>This will cause a dump of the Oracle state objects to 
      occur when this error</FONT> <BR><FONT size=2>is </FONT><BR><FONT 
      size=2>signalled.&nbsp; By looking in the dump for 'load=X' and then 
      looking up a few</FONT> <BR><FONT size=2>lines</FONT> <BR><FONT size=2>for 
      'name=' you can often tell whether an object was being loaded into the 
      </FONT><BR><FONT size=2>shared pool when this error occurred.&nbsp; If an 
      object was being loaded then it</FONT> <BR><FONT size=2>is</FONT> 
      <BR><FONT size=2>likely that this load is the cause of the problem and the 
      object should be </FONT><BR><FONT size=2>'kept' in the shared pool.&nbsp; 
      The object being loaded is the object printed</FONT> <BR><FONT 
      size=2>after</FONT> <BR><FONT size=2>the 'name='. Do not use the 'level 4' 
      option in versions before 7.0.16</FONT> <BR><FONT size=2>because</FONT> 
      <BR><FONT size=2>a bug existed that often caused the system to crash with 
      this option enabled</FONT> </P>
      <P><FONT size=2>due to a latch level violation.</FONT> </P>
      <P><FONT size=2>Prior to version 7.3, there were a handful of cases where 
      the RDBMS or</FONT> <BR><FONT size=2>PL/SQL</FONT> <BR><FONT size=2>would 
      attempt to allocate large pieces of contiguous memory. Most of this</FONT> 
      <BR><FONT size=2>has</FONT> <BR><FONT size=2>been fixed for 7.3. This 
      problem was especially acute when running MTS, when</FONT> </P>
      <P><FONT size=2>the UGA would be located in the SGA. This should also be 
      fixed in 7.3 and</FONT> <BR><FONT size=2>using</FONT> <BR><FONT size=2>MTS 
      for a high OLTP scenario is recommended. As a result of all these</FONT> 
      <BR><FONT size=2>changes, </FONT><BR><FONT size=2>the ORA-04031 error 
      should be virtually eliminated. If an ORA-04031 error is</FONT> <BR><FONT 
      size=2>signalled, quite likely the shared pool is over 90% utilized and 
      the </FONT><BR><FONT size=2>alternative is to increase the shared pool. 
      The only known situation is</FONT> <BR><FONT size=2>PL/SQL</FONT> 
      <BR><FONT size=2>packages (like STANDARD) where the package contains a 
      very large number</FONT> <BR><FONT size=2>(over</FONT> <BR><FONT 
      size=2>400) procedure/function definitions. This still needs to be in 
      contiguous</FONT> <BR><FONT size=2>memory and may request memory chunks as 
      large as 15K. Packages like this</FONT> <BR><FONT size=2>should 
      </FONT><BR><FONT size=2>be the only ones that should be kept.</FONT> </P>
      <P><FONT size=2>iii) X$KSMLRU</FONT> </P>
      <P><FONT size=2>There is a&nbsp; fixed table called&nbsp; x$ksmlru 
      that&nbsp; tracks&nbsp; allocations in the </FONT><BR><FONT size=2>shared 
      pool that cause other objects in&nbsp; the shared pool&nbsp; to be aged 
      out.</FONT> <BR><FONT size=2>This</FONT> <BR><FONT size=2>fixed table can 
      be&nbsp; used&nbsp; to identify what is causing the large 
      allocation.</FONT> </P>
      <P><FONT size=2>The columns of this fixed table are the following:</FONT> 
      </P>
      <P><FONT size=2>KSMLRCOM - allocation comment that describes the type of 
      allocation.</FONT> </P>
      <P><FONT size=2>If this comment is something like 'MPCODE' or 'PLSQL%' 
      then there is a large</FONT> </P>
      <P><FONT size=2>PL/SQL object being loaded into the shared pool.&nbsp; 
      This PL/SQL object will</FONT> <BR><FONT size=2>need </FONT><BR><FONT 
      size=2>to be 'kept' in the shared pool.</FONT> </P>
      <P><FONT size=2>If this comment is 'kgltbtab' then the allocation is for a 
      dependency table</FONT> <BR><FONT size=2>in </FONT><BR><FONT size=2>the 
      library cache.&nbsp; This is only a problem when several hundred users 
      are</FONT> <BR><FONT size=2>logged</FONT> <BR><FONT size=2>on using 
      distinct user ids.&nbsp; The solution in this case is to use fully</FONT> 
      <BR><FONT size=2>qualified</FONT> <BR><FONT size=2>names for all table 
      references.</FONT> </P>
      <P><FONT size=2>If you are running MTS and the comment is something like 
      'Fixed UGA' then</FONT> <BR><FONT size=2>the </FONT><BR><FONT 
      size=2>problem is that the init.ora parameter 'open_cursors' is set too 
      high.</FONT> </P>
      <P><FONT size=2>KSMLRSIZ - amount of contiguous memory being 
      allocated.&nbsp; Values over around</FONT> <BR><FONT size=2>5K 
      </FONT><BR><FONT size=2>start to be a problem, values over 10K are a 
      serious problem, and values</FONT> <BR><FONT size=2>over </FONT><BR><FONT 
      size=2>20K are very serious problems.&nbsp; Anything less then 5K should 
      not be a</FONT> <BR><FONT size=2>problem.</FONT> </P>
      <P><FONT size=2>KSMLRNUM - number of objects that were flushed from the 
      shared pool in order</FONT> <BR><FONT size=2>allocate the memory.</FONT> 
      </P>
      <P><FONT size=2>In release 7.1.3 or later, the following columns also 
      exist:</FONT> </P>
      <P><FONT size=2>KSMLRHON - the name of the object being loaded into the 
      shared pool if the</FONT> <BR><FONT size=2>object is a PL/SQL object or a 
      cursor.</FONT> </P>
      <P><FONT size=2>KSMLROHV - hash value of object being loaded</FONT> </P>
      <P><FONT size=2>KSMLRSES - SADDR of the session that loaded the 
      object.</FONT> </P>
      <P><FONT size=2>The advantage of X$KSMLRU is that it allows you to 
      identify problems with</FONT> <BR><FONT size=2>fragmentation that are 
      effecting performance, but that are not bad enough to</FONT> <BR><FONT 
      size=2>be</FONT> <BR><FONT size=2>causing ORA-04031 errors to be 
      signalled. If a lot of objects are being </FONT><BR><FONT 
      size=2>periodically flushed from the shared pool then this will cause 
      response time</FONT> </P>
      <P><FONT size=2>problems and will likely cause library cache latch 
      contention problems when</FONT> <BR><FONT size=2>the</FONT> <BR><FONT 
      size=2>objects are reloaded into the shared pool. With version 7.2, the 
      library</FONT> <BR><FONT size=2>cache </FONT><BR><FONT size=2>latch 
      contention should be significantly reduced with the breaking up of 
      the</FONT> </P>
      <P><FONT size=2>library cache pin latch into a configurable set of 
      symmetric library cache </FONT><BR><FONT size=2>latches.</FONT> </P>
      <P><FONT size=2>One unusual thing about the x$ksmlru fixed table is that 
      the contents of the</FONT> </P>
      <P><FONT size=2>fixed table are erased whenever someone selects from the 
      fixed table. This</FONT> <BR><FONT size=2>is </FONT><BR><FONT size=2>done 
      since the fixed table stores only the largest allocations that have 
      </FONT><BR><FONT size=2>occurred.&nbsp; The values are reset after being 
      selected so that subsequent</FONT> <BR><FONT size=2>large </FONT><BR><FONT 
      size=2>allocations can be noted even if they were not quite as large as 
      others that</FONT> </P>
      <P><FONT size=2>occurred previously.&nbsp; Because of this resetting, the 
      output of selecting</FONT> <BR><FONT size=2>from </FONT><BR><FONT 
      size=2>this table should be carefully noted since it cannot be reselected 
      if it is </FONT><BR><FONT size=2>forgotten.&nbsp; Also you should take 
      care that there are not multiple people on</FONT> <BR><FONT 
      size=2>one</FONT> <BR><FONT size=2>database that select from this table 
      because only one of them will select</FONT> <BR><FONT size=2>the 
      </FONT><BR><FONT size=2>real data.</FONT> </P>
      <P><FONT size=2>To monitor this fixed table just run the following:</FONT> 
      </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select * from x$ksmlru where ksmlrsiz 
      &gt; 5000;</FONT> </P>
      <P><FONT size=2>iv) MTS</FONT> </P>
      <P><FONT size=2>Oracle users using SQL*Net V2 can connect to the database 
      using dedicated</FONT> <BR><FONT size=2>servers, or multiple clients can 
      use a pool of shared (or MTS) servers. The</FONT> <BR><FONT size=2>biggest 
      memory implication of this mode is that the session memory (also</FONT> 
      <BR><FONT size=2>known</FONT> <BR><FONT size=2>as the UGA) for every 
      session needs to be accessible to every MTS server.</FONT> <BR><FONT 
      size=2>This</FONT> <BR><FONT size=2>implies that the logical UGA comes out 
      of the physical SGA (or the shared</FONT> <BR><FONT size=2>pool) instead 
      of the PGA (process memory).</FONT> </P>
      <P><FONT size=2>In versions prior to 7.3, there were a few components in 
      the UGA that would</FONT> <BR><FONT size=2>request large contiguous chunks 
      of memory, contributing to fragmentation of</FONT> <BR><FONT size=2>the 
      shared pool if using MTS. If the system had been up for a while, 
      users</FONT> <BR><FONT size=2>would have failures when attempting to 
      connect or executing sql. Starting</FONT> <BR><FONT size=2>with</FONT> 
      <BR><FONT size=2>7.3, all these allocations have been segmented such that 
      the average size of</FONT> <BR><FONT size=2>memory chunks allocated to the 
      UGA should be about 5K.</FONT> </P>
      <P><FONT size=2>B) CORRECTION OF FRAGMENTATION</FONT> </P>
      <P><FONT size=2>i) KEEPING OBJECTS</FONT> </P>
      <P><FONT size=2>The primary source of problems is large PL/SQL objects. 
      The means of</FONT> <BR><FONT size=2>correcting</FONT> <BR><FONT 
      size=2>these errors is to 'keep' large PL/SQL object in the shared pool at 
      startup </FONT><BR><FONT size=2>time. This will load the objects into the 
      shared pool and will make sure</FONT> <BR><FONT size=2>that 
      </FONT><BR><FONT size=2>the objects are never aged out of the shared pool. 
      If the objects are never </FONT><BR><FONT size=2>aged out then there will 
      not be a problem with trying to load them and not </FONT><BR><FONT 
      size=2>having enough memory.</FONT> </P>
      <P><FONT size=2>Objects are 'kept' in the shared pool using the 
      dbms_shared_pool package</FONT> <BR><FONT size=2>that </FONT><BR><FONT 
      size=2>is defined in the dbmspool.sql file.&nbsp; For example:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; execute 
      dbms_shared_pool.keep('SYS.STANDARD');</FONT> </P>
      <P><FONT size=2>All large packages that are shipped should be 'kept' if 
      the customer uses</FONT> <BR><FONT size=2>PL/SQL. This includes 
      'STANDARD', 'DBMS_STANDARD', and 'DIUTIL'. With 7.3,</FONT> <BR><FONT 
      size=2>the</FONT> <BR><FONT size=2>only package left in this list is 
      'STANDARD'.</FONT> </P>
      <P><FONT size=2>All large customer packages should also be marked 
      'kept'.</FONT> </P>
      <P><FONT size=2>To mark all packages in the system 'kept' execute the 
      following:</FONT> </P>
      <P><FONT size=2>declare</FONT> <BR><FONT size=2>&nbsp; own 
      varchar2(100);</FONT> <BR><FONT size=2>&nbsp; nam varchar2(100);</FONT> 
      <BR><FONT size=2>&nbsp; cursor pkgs is</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; select owner, object_name</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp; from dba_objects</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp; where&nbsp; object_type = 
      'PACKAGE';</FONT> <BR><FONT size=2>begin</FONT> <BR><FONT size=2>&nbsp; 
      open pkgs;</FONT> <BR><FONT size=2>&nbsp; loop</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; fetch pkgs into own, nam;</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; exit when pkgs%notfound;</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; dbms_shared_pool.keep(own || '.' || nam, 
      'P');</FONT> <BR><FONT size=2>&nbsp; end loop;</FONT> <BR><FONT 
      size=2>end;</FONT> </P>
      <P><FONT size=2>The dbms_shared_pool package was introduced in 7.0 and has 
      evolved over the </FONT><BR><FONT size=2>versions. Until 7.1.5, 'keep' 
      could only be used for packages. Starting with</FONT> </P>
      <P><FONT size=2>7.1.6, this was extended to standalone procedures, cursors 
      as well as</FONT> <BR><FONT size=2>triggers.</FONT> <BR><FONT size=2>For 
      detailed usage instructions, see the dbmspool.sql file. So, prior 
      to</FONT> <BR><FONT size=2>this </FONT><BR><FONT size=2>version, if you 
      have large procedures or large anonymous blocks, then these 
      </FONT><BR><FONT size=2>will need to be put into packages and marked kept. 
      With 7.3, most packages</FONT> <BR><FONT size=2>do </FONT><BR><FONT 
      size=2>not need to be kept any longer since PL/SQL no longer requires 
      large amounts</FONT> <BR><FONT size=2>of</FONT> <BR><FONT 
      size=2>contiguous memory to load packages/procedures in memory.</FONT> 
</P>
      <P><FONT size=2>You can determine what large stored objects are in the 
      shared pool by</FONT> <BR><FONT size=2>selecting</FONT> <BR><FONT 
      size=2>from the v$db_object_cache fixed view.&nbsp; This will also tell 
      you which</FONT> <BR><FONT size=2>objects </FONT><BR><FONT size=2>have 
      been marked kept.&nbsp; This can be done with the following query:</FONT> 
      </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select * from v$db_object_cache where 
      sharable_mem &gt; 10000;</FONT> </P>
      <P><FONT size=2>Note that this query will not catch PL/SQ: objects that 
      are only rarely used</FONT> </P>
      <P><FONT size=2>and therefore the PL/SQL object is not currently loaded in 
      the shared pool.</FONT> </P>
      <P><FONT size=2>To determine what large PL/SQL objects are currently 
      loaded in the shared</FONT> <BR><FONT size=2>pool </FONT><BR><FONT 
      size=2>and are not marked 'kept' and therefore may cause a problem, 
      execute the</FONT> <BR><FONT size=2>following:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select name, sharable_mem</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; from v$db_object_cache</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp; where sharable_mem &gt; 10000</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; and (type = 'PACKAGE' or type = 
      'PACKAGE BODY' or type = 'FUNCTION'</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; or type = 
      'PROCEDURE')</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; and kept = 
      'NO';</FONT> </P>
      <P><FONT size=2>Another approach to the above is to use the 
      dbms_shared_pool.sizes</FONT> <BR><FONT size=2>procedure.</FONT> <BR><FONT 
      size=2>To use this in SQLDBA:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; set serveroutput on;</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; execute dbms_shared_pool.sizes(10);</FONT> </P>
      <P><FONT size=2>This should show you the names of all the objects in the 
      shared pool that</FONT> <BR><FONT size=2>take</FONT> <BR><FONT size=2>more 
      that 10K of memory as well as if they are marked kept or not. For 
      SQL</FONT> <BR><FONT size=2>statements, if there are multiple versions of 
      a query (usually a bug if the</FONT> <BR><FONT size=2>count is more than 
      3), they will also be indicated in parenthesis. Use the</FONT> <BR><FONT 
      size=2>following query to check for problems:</FONT> </P>
      <P><FONT size=2>&nbsp; select sql_text, loaded_versions, version_count, 
      sharable_mem</FONT> <BR><FONT size=2>&nbsp;&nbsp; from v$sqlarea where 
      loaded_versions &gt; 3</FONT> <BR><FONT size=2>&nbsp;&nbsp; order by 
      sharable_mem;</FONT> </P>
      <P><FONT size=2>In Oracle7.3 onwards the best candidates for keeping can 
      be seen by querying</FONT> </P>
      <P><FONT size=2>the table X$KSMSP to see if there are any chunks in the 
      shared-pool that</FONT> <BR><FONT size=2>have </FONT><BR><FONT size=2>the 
      KSMCHSIZ larger than 5K and KSMCHCOM like '%PL/SQL%'. If so then one 
      can</FONT> </P>
      <P><FONT size=2>identify the object name and owner of this chunk using the 
      following SQL:</FONT> </P>
      <P><FONT size=2>&nbsp; select distinct</FONT> <BR><FONT size=2>&nbsp; 
      decode(kglobtyp,0,'CURSOR',7,'PROCEDURE',8,'FUNCTION',9,'PACKAGE',</FONT> 
      <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      11,'PACKAGE BODY',12,'TRIGGER',13,'TYPE',14,'TYPE</FONT> <BR><FONT 
      size=2>BODY','OTHER')</FONT> <BR><FONT size=2>&nbsp;&nbsp; ||' - 
      '||kglnaown||'.'||kglnaobj "Eligible PL/SQL objects"</FONT> <BR><FONT 
      size=2>&nbsp; from x$kglob</FONT> <BR><FONT size=2>&nbsp; where kglobhd4 
      in</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; (select&nbsp; ksmchpar from 
      x$ksmsp</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; where 
      ksmchcom='PL/SQL MPCODE' and ksmchsiz&gt;5120)</FONT> </P>
      <P><FONT size=2>If you are 'keeping' PL/SQL objects today and migrate to 
      7.3 or higher there</FONT> <BR><FONT size=2>is</FONT> <BR><FONT size=2>no 
      need to re-assess the list of objects that you are keeping.</FONT> </P>
      <P><FONT size=2>ii) USE BIND VARIABLES</FONT> </P>
      <P><FONT size=2>One of the best things that can be done to reduce the 
      amount of</FONT> <BR><FONT size=2>fragmentation</FONT> <BR><FONT size=2>is 
      to reduce or eliminate the number of sql statements in the shared 
      pool</FONT> <BR><FONT size=2>that</FONT> <BR><FONT size=2>are duplicates 
      of each other except for a constant that is embedded in the</FONT> 
      <BR><FONT size=2>statement.&nbsp; The statements should be replaced with 
      one statement that uses</FONT> <BR><FONT size=2>a bind variable instead of 
      a constant.</FONT> </P>
      <P><FONT size=2>For example:</FONT> </P>
      <P><FONT size=2>&nbsp; select * from emp where empno=1;</FONT> <BR><FONT 
      size=2>&nbsp; select * from emp where empno=2;</FONT> <BR><FONT 
      size=2>&nbsp; select * from emp where empno=3;</FONT> </P>
      <P><FONT size=2>Should all be replaced with:</FONT> </P>
      <P><FONT size=2>&nbsp; select * from emp where empno=:1;</FONT> </P>
      <P><FONT size=2>You can identify statements that potentially fall into 
      this class with a</FONT> <BR><FONT size=2>query </FONT><BR><FONT 
      size=2>like the following:</FONT> </P>
      <P><FONT size=2>&nbsp; select substr(sql_text, 1, 30) sql, count(*) 
      copies</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; from v$sqlarea</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp; group by substr(sql_text, 1, 30)</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp; having count(*) &gt; 3;</FONT> </P>
      <P><FONT size=2>iii) MAX BIND SIZE</FONT> </P>
      <P><FONT size=2>It is possible for a SQL statement to not be shared 
      because the max bind </FONT><BR><FONT size=2>variable lengths of the bind 
      variables in the statement do not match.&nbsp; This</FONT> <BR><FONT 
      size=2>is </FONT><BR><FONT size=2>automatically taken care of for 
      precompiler programs and forms programs, but</FONT> </P>
      <P><FONT size=2>could be a problem for programs that directly use 
      OCI.&nbsp; The bind call in OCI</FONT> </P>
      <P><FONT size=2>takes two arguments, one is the max length of the value, 
      and the other is a </FONT><BR><FONT size=2>pointer to the actual 
      length.&nbsp; If the current length is always passed in as</FONT> 
      <BR><FONT size=2>the</FONT> <BR><FONT size=2>max length instead of the max 
      possible length for the variable, then this</FONT> <BR><FONT size=2>could 
      </FONT><BR><FONT size=2>cause the SQL statement not to be shared.</FONT> 
      </P>
      <P><FONT size=2>To identify statements that might potentially have this 
      problem execute the </FONT><BR><FONT size=2>following statement:</FONT> 
      </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select sql_text, version_count from 
      v$sqlarea where version_count &gt; 5;</FONT> </P>
      <P><FONT size=2>Starting with 7.1.6 this should no longer be an issue as 
      the server can </FONT><BR><FONT size=2>graduate bind buffers even when the 
      user's max bind lengths are jumping up</FONT> <BR><FONT size=2>or</FONT> 
      <BR><FONT size=2>down and continue to share cursors that are built for 
      larger buffer lengths</FONT> <BR><FONT size=2>and flush the smaller sql 
      compilation from the shared pool.</FONT> </P>
      <P><FONT size=2>iv) ELIMINATING LARGE ANONYMOUS PL/SQL</FONT> </P>
      <P><FONT size=2>Large anonymous PL/SQL blocks should be turned into small 
      anonymous PL/SQL</FONT> <BR><FONT size=2>blocks that call packaged 
      functions.&nbsp; The packages should be 'kept' in</FONT> <BR><FONT 
      size=2>memory. </FONT><BR><FONT size=2>For version earlier that 7.3, this 
      includes anonymous PL/SQL blocks that are</FONT> </P>
      <P><FONT size=2>used for trigger definitions. With 7.3, triggers are 
      compiled and stored to </FONT><BR><FONT size=2>disk like standalone 
      procedures and should be treated as such. Large</FONT> <BR><FONT 
      size=2>anonymous </FONT><BR><FONT size=2>blocks can be identified with the 
      following query:</FONT> </P>
      <P><FONT size=2>&nbsp; select sql_text from v$sqlarea</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; where 
      command_type=47&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -- command type for 
      anonymous block</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; and 
      length(sql_text) &gt; 500;</FONT> </P>
      <P><FONT size=2>Note that this query will not catch PL/SQL blocks that are 
      only rarely used</FONT> <BR><FONT size=2>and</FONT> <BR><FONT 
      size=2>therefore the PL/SQL block is not currently loaded in the shared 
      pool.</FONT> </P>
      <P><FONT size=2>Another option that can be used when an anonymous block 
      cannot be turned</FONT> <BR><FONT size=2>into</FONT> <BR><FONT size=2>a 
      package is to mark the anonymous block with some string so that it can 
      be</FONT> <BR><FONT size=2>identified in v$sqlarea and marked 
      'kept'.</FONT> </P>
      <P><FONT size=2>For example, instead of using:</FONT> <BR><FONT 
      size=2>&nbsp; declare x number; begin x := 5; end;;</FONT> </P>
      <P><FONT size=2>you can use:</FONT> <BR><FONT size=2>&nbsp; declare /* 
      KEEP_ME */ x number; begin x := 5; end;</FONT> </P>
      <P><FONT size=2>You can then use the following procedure to select these 
      statements out of</FONT> <BR><FONT size=2>the </FONT><BR><FONT 
      size=2>shared pool and mark them 'kept' using the dbms_shared_pool.keep 
      package.</FONT> </P>
      <P><FONT size=2>declare</FONT> <BR><FONT size=2>&nbsp; /* DONT_KEEP_ME 
      */</FONT> <BR><FONT size=2>&nbsp; addr varchar2(10);</FONT> <BR><FONT 
      size=2>&nbsp; hash number;</FONT> <BR><FONT size=2>&nbsp; cursor anon 
      is</FONT> <BR><FONT size=2>&nbsp;&nbsp; select address, hash_value</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; from v$sqlarea</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; where command_type = 
      47&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      -- command type for anonymous</FONT> <BR><FONT size=2>block</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; and&nbsp; sql_text like '% 
      KEEP_ME %'</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp; and&nbsp; 
      sql_text not like '%DONT_KEEP_ME%';</FONT> <BR><FONT size=2>begin</FONT> 
      <BR><FONT size=2>&nbsp; open anon;</FONT> <BR><FONT size=2>&nbsp; 
      loop</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; fetch anon into addr, 
      hash;</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; exit when 
      anon%notfound;</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; 
      dbms_shared_pool.keep(addr || ',' || to_char(hash), 'C');</FONT> <BR><FONT 
      size=2>&nbsp; end loop;</FONT> <BR><FONT size=2>end;</FONT> </P>
      <P><FONT size=2>v) REDUCING USAGE</FONT> </P>
      <P><FONT size=2>Another way to reducing fragmentation is to reduce 
      consumption. This is of</FONT> <BR><FONT size=2>special importance when 
      using MTS, when every user's session memory is in</FONT> <BR><FONT 
      size=2>the</FONT> <BR><FONT size=2>shared pool and the impact is 
      multiplied by the total concurrent users.</FONT> </P>
      <P><FONT size=2>Insert, update, delete and anonymous blocks complete the 
      execution in one</FONT> <BR><FONT size=2>round </FONT><BR><FONT 
      size=2>trip. All the memory that is allocated on the server for the 
      execute comes</FONT> <BR><FONT size=2>from</FONT> <BR><FONT size=2>the PGA 
      and is freed before the call returns to the user. But in the case 
      of</FONT> </P>
      <P><FONT size=2>selects, memory required to execute the statement - which 
      could be large if</FONT> <BR><FONT size=2>a </FONT><BR><FONT size=2>sort 
      was involved - is not freed until the end-of-fetch is reached or 
      the</FONT> <BR><FONT size=2>query </FONT><BR><FONT size=2>is cancelled. In 
      these situations using the OCI features to do an exact</FONT> <BR><FONT 
      size=2>fetch </FONT><BR><FONT size=2>and cancel helps free memory back to 
      the pool.</FONT> </P>
      <P><FONT size=2>If the application logic has been embedded into server 
      side PL/SQL, a large</FONT> <BR><FONT size=2>number of cursors may be 
      getting cached on the server for every user. Though</FONT> <BR><FONT 
      size=2>this results in reduced latch contention and faster response, it 
      does use</FONT> <BR><FONT size=2>more</FONT> <BR><FONT size=2>memory in 
      the UGA. Setting the close_cached_open_cursors init.ora to TRUE</FONT> 
      <BR><FONT size=2>closes the PL/SQL cached cursors on the server, freeing 
      the memory.</FONT> </P>
      <P><FONT 
      size=2>*************************************************************************</FONT> 
      </P>
      <P><FONT size=2>2) COMMON FALLACIES</FONT> </P>
      <P><FONT size=2>There are a number of common fallacies about the shared 
      pool that are often</FONT> <BR><FONT size=2>stated as fact.</FONT> </P>
      <P><FONT size=2>A) FREE MEMORY</FONT> </P>
      <P><FONT size=2>One fallacy is that the amount of 'free memory' reported 
      in v$sgastat needs</FONT> <BR><FONT size=2>to </FONT><BR><FONT size=2>be 
      kept high.&nbsp; This is incorrect.&nbsp; The free memory reported in this 
      table is</FONT> <BR><FONT size=2>not</FONT> <BR><FONT size=2>like the free 
      memory reported by operating system statistics.&nbsp; Since the</FONT> 
      <BR><FONT size=2>shared</FONT> <BR><FONT size=2>pool acts as a cache, 
      nothing will ever be aged out of the shared pool until</FONT> <BR><FONT 
      size=2>all</FONT> <BR><FONT size=2>the free memory has been used up. This 
      is entirely normal.</FONT> </P>
      <P><FONT size=2>Free memory is more properly thought of as 'wasted 
      memory'.&nbsp; You would</FONT> <BR><FONT size=2>rather</FONT> <BR><FONT 
      size=2>see this value be low than very high.&nbsp; In fact, a high value 
      of free memory</FONT> <BR><FONT size=2>is </FONT><BR><FONT 
      size=2>sometimes a symptom that a lot of objects have been aged out of the 
      shared</FONT> <BR><FONT size=2>pool</FONT> <BR><FONT size=2>and therefore 
      the system is experiencing fragmentation problems.</FONT> </P>
      <P><FONT size=2>B) FLUSH SHARED POOL</FONT> </P>
      <P><FONT size=2>Some people think that frequently executing 'alter system 
      flush shared_pool'</FONT> <BR><FONT size=2>improves the performance of the 
      system and decreases the amount of </FONT><BR><FONT 
      size=2>fragmentation.&nbsp; This is incorrect.&nbsp; Executing this 
      statement causes a big</FONT> <BR><FONT size=2>spike</FONT> <BR><FONT 
      size=2>in performance and does nothing to improve fragmentation.</FONT> 
      </P>
      <P><FONT size=2>The only time when it might be useful to run this 
      statement is between</FONT> <BR><FONT size=2>shifts of</FONT> <BR><FONT 
      size=2>users so that the objects that are relevant to the last shift of 
      users can</FONT> <BR><FONT size=2>be </FONT><BR><FONT size=2>flushed out 
      before the next shift of users starts to use the system.&nbsp; This</FONT> 
      <BR><FONT size=2>is </FONT><BR><FONT size=2>almost never needed 
      though.</FONT> </P>
      <P><FONT 
      size=2>*************************************************************************</FONT> 
      </P>
      <P><FONT size=2>3) SIZING OF SHARED POOL</FONT> </P>
      <P><FONT size=2>One very difficult judgement that needs to be make in 
      Oracle7 is to</FONT> <BR><FONT size=2>determine</FONT> <BR><FONT 
      size=2>the proper size of the shared pool.&nbsp; The following provides 
      some guidelines</FONT> <BR><FONT size=2>for</FONT> <BR><FONT 
      size=2>this.&nbsp; It should be emphasized that these are just guidelines, 
      there are no </FONT><BR><FONT size=2>hard and fast rules here and 
      experimentation will be needed to determine a</FONT> <BR><FONT 
      size=2>good</FONT> <BR><FONT size=2>value.</FONT> </P>
      <P><FONT size=2>The shared pool size is highly application 
      dependent.&nbsp; To determine the</FONT> <BR><FONT size=2>shared 
      </FONT><BR><FONT size=2>pool size that will be needed for a production 
      system it is generally</FONT> <BR><FONT size=2>necessary</FONT> <BR><FONT 
      size=2>to first develop the application and run it on a test system and 
      take some </FONT><BR><FONT size=2>measurements.&nbsp; The test system 
      should be run with a very large value for the</FONT> </P>
      <P><FONT size=2>shared pool size to make the measurements 
      meaningful.</FONT> </P>
      <P><FONT size=2>A) OBJECTS STORED IN THE DATABASE</FONT> </P>
      <P><FONT size=2>The amount of shared pool that needs to be allocated for 
      objects that are </FONT><BR><FONT size=2>stored in the database like 
      packages and views is easy to measure.&nbsp; You can </FONT><BR><FONT 
      size=2>just measure their size directly with the following 
      statement:</FONT> </P>
      <P><FONT size=2>&nbsp; select sum(sharable_mem) from 
      v$db_object_cache;</FONT> </P>
      <P><FONT size=2>This is especially effective because all large pl/sql 
      object should be</FONT> <BR><FONT size=2>'kept'</FONT> <BR><FONT size=2>in 
      the shared pool at all times.</FONT> </P>
      <P><FONT size=2>B) SQL</FONT> </P>
      <P><FONT size=2>The amount of memory needed to store SQL statements in the 
      shared pool is</FONT> <BR><FONT size=2>more </FONT><BR><FONT 
      size=2>difficult to measure because of the needs of dynamic SQL.&nbsp; If 
      an application</FONT> </P>
      <P><FONT size=2>has no dynamic SQL then the amount of memory can simply be 
      measured after</FONT> <BR><FONT size=2>the </FONT><BR><FONT 
      size=2>application has run for a while by just selecting it out of the 
      shared pool</FONT> <BR><FONT size=2>as </FONT><BR><FONT 
      size=2>follows:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select sum(sharable_mem) from 
      v$sqlarea;</FONT> </P>
      <P><FONT size=2>If the application has a moderate or large amount of 
      dynamic SQL like most </FONT><BR><FONT size=2>applications do, then a 
      certain amount&nbsp; of memory will be needed for the</FONT> <BR><FONT 
      size=2>shared </FONT><BR><FONT size=2>SQL, plus more for the dynamic SQL, 
      and more so that the dynamic SQL does</FONT> <BR><FONT size=2>not 
      </FONT><BR><FONT size=2>age the shared SQL out of the shared pool.</FONT> 
      </P>
      <P><FONT size=2>The amount of memory for the shared SQL can be 
      approximated by the</FONT> <BR><FONT size=2>following:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select sum(sharable_mem) from v$sqlarea 
      where executions &gt; 5;</FONT> </P>
      <P><FONT size=2>The remaining memory in v$sqlarea is for dynamic SQL&nbsp; 
      Some shared pool will</FONT> <BR><FONT size=2>need to be budgeted for this 
      also, but there are few rules here.</FONT> </P>
      <P><FONT size=2>C) PER-USER PER-CURSOR MEMORY</FONT> </P>
      <P><FONT size=2>You will need to allow around 250 bytes of memory in the 
      shared pool per</FONT> <BR><FONT size=2>concurrent user for each open 
      cursor that the user has whether the cursor is</FONT> </P>
      <P><FONT size=2>shared or not.&nbsp; During the peak usage time of the 
      production system, you can</FONT> </P>
      <P><FONT size=2>measure this as follows:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp;&nbsp; select sum(250 * users_opening) from 
      v$sqlarea;</FONT> </P>
      <P><FONT size=2>In a test system you can measure it by selecting the 
      number of open cursors</FONT> <BR><FONT size=2>for a test user and 
      multiplying by the total number of users:</FONT> </P>
      <P><FONT size=2>&nbsp; select 250 * value bytes_per_user</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; from v$sesstat s, v$statname n</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; where s.statistic# = n.statistic#</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; and&nbsp; n.name = 'opened cursors 
      current'</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; and&nbsp; s.sid&nbsp; 
      =&nbsp; 23;&nbsp;&nbsp;&nbsp; -- replace 23 with session id of user 
      being</FONT> <BR><FONT size=2>measured</FONT> </P>
      <P><FONT size=2>The per-user per-cursor memory is one of the classes of 
      memory that shows up</FONT> <BR><FONT size=2>as</FONT> <BR><FONT 
      size=2>'library cache' in v$sgastat.</FONT> </P>
      <P><FONT size=2>D) MTS</FONT> </P>
      <P><FONT size=2>If you are using multi-threaded server, then you will need 
      to allow enough</FONT> <BR><FONT size=2>memory for all the shared server 
      users to put their session memory in the </FONT><BR><FONT size=2>shared 
      pool.&nbsp; This can be measured for one user with the following 
      query:</FONT> </P>
      <P><FONT size=2>&nbsp; select value sess_mem</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; from v$sesstat s, v$statname n</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; where s.statistic# = n.statistic#</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; and&nbsp; n.name = 'session uga memory'</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp; and&nbsp; s.sid&nbsp; =&nbsp; 
      23;&nbsp;&nbsp;&nbsp; -- replace 23 with session id of user being</FONT> 
      <BR><FONT size=2>measured</FONT> </P>
      <P><FONT size=2>A more conservative value to use is the maximum session 
      memory that was</FONT> <BR><FONT size=2>ever allocated by the user:</FONT> 
      </P>
      <P><FONT size=2>&nbsp; select value sess_max_mem</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; from v$sesstat s, v$statname n</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; where s.statistic# = n.statistic#</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; and&nbsp; n.name = 'session uga memory 
      max'</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp; and&nbsp; s.sid&nbsp; 
      =&nbsp; 23;&nbsp;&nbsp;&nbsp; -- replace 23 with session id of user 
      being</FONT> <BR><FONT size=2>measured</FONT> </P>
      <P><FONT size=2>To select this value for all the currently logged on users 
      the following</FONT> <BR><FONT size=2>query</FONT> <BR><FONT size=2>can be 
      used:</FONT> </P>
      <P><FONT size=2>&nbsp; select sum(value) all_sess_mem</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; from v$sesstat s, v$statname n</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; where s.statistic# = n.statistic#</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp; and&nbsp; n.name = 'session uga memory 
      max';</FONT> </P>
      <P><FONT size=2>E) OVERHEAD</FONT> </P>
      <P><FONT size=2>You will need to add a minimum of 30% overhead to the 
      values calculated</FONT> <BR><FONT size=2>above </FONT><BR><FONT size=2>to 
      allow for unexpected and unmeasured usage of the shared pool.</FONT> </P>
      <P><FONT 
      size=2>*************************************************************************</FONT> 
      </P>
      <P><FONT size=2>4) FINAL COMMENTS</FONT> </P>
      <P><FONT size=2>The most important point that needs to be understood by 
      everyone using</FONT> <BR><FONT size=2>Oracle7 </FONT><BR><FONT size=2>and 
      PL/SQL (prior to release 7.3) is that all large PL/SQL objects must 
      be</FONT> <BR><FONT size=2>made </FONT><BR><FONT size=2>into packages and 
      those packages must be kept in the shared pool. This point</FONT> </P>
      <P><FONT size=2>cannot be over emphasized.&nbsp; Many customers, 
      especially those running a lot</FONT> <BR><FONT size=2>of </FONT><BR><FONT 
      size=2>users, have had terrible performance problems that were completely 
      cleared</FONT> <BR><FONT size=2>up </FONT><BR><FONT size=2>by doing 
      this.</FONT> </P><BR>
      <P><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      APPENDIX I:&nbsp; Reserved Shared Pool</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      =================================</FONT> </P>
      <P><FONT size=2>1. RESERVED SPACE FROM THE SHARED POOL</FONT> <BR><FONT 
      size=2>======================================</FONT> <BR><FONT size=2>On 
      busy systems, the RDBMS may have difficulty finding a contiguous piece 
      of</FONT> </P>
      <P><FONT size=2>memory to satisfy a large request for memory.&nbsp; 
      Because the RDBMS will search</FONT> </P>
      <P><FONT size=2>for and free currently unused memory, the search for this 
      large piece of</FONT> <BR><FONT size=2>memory </FONT><BR><FONT size=2>may 
      disrupt the behavior of the share pool, leading to more 
      fragmentation</FONT> <BR><FONT size=2>and </FONT><BR><FONT size=2>poor 
      performance.</FONT> </P>
      <P><FONT size=2>RDBMS 7.1.5 allows DBAs to reserve memory within the 
      shared pool to satisfy </FONT><BR><FONT size=2>these large allocations 
      during RDBMS operations such as PL/SQL compilation</FONT> <BR><FONT 
      size=2>and </FONT><BR><FONT size=2>trigger compilation.&nbsp; Smaller 
      objects will not fragment the reserved list, </FONT><BR><FONT 
      size=2>helping to ensure the reserved list will have large contiguous 
      chunks of</FONT> <BR><FONT size=2>memory.</FONT> <BR><FONT size=2>Once the 
      memory allocated from the reserved list is freed, it returns to the</FONT> 
      </P>
      <P><FONT size=2>reserved list.</FONT> </P>
      <P><FONT size=2>The size of the reserved list, as well as the minimum size 
      of the objects</FONT> <BR><FONT size=2>that </FONT><BR><FONT size=2>can be 
      allocated from the reserved list are controlled via init.ora</FONT> 
      <BR><FONT size=2>parameters:</FONT> <BR><FONT 
      size=2>shared_pool_reserved_size and 
      shared_pool_reserved_min_alloc.</FONT> </P>
      <P><FONT size=2>1.1&nbsp; shared_pool_reserved_size</FONT> <BR><FONT 
      size=2>------------------------------</FONT> <BR><FONT size=2>The init.ora 
      parameter shared_pool_reserved_size controls the amount of</FONT> 
      <BR><FONT size=2>shared_pool_size reserved for large allocations.&nbsp; In 
      order to create a</FONT> <BR><FONT size=2>reserved</FONT> <BR><FONT 
      size=2>list, shared_pool_reserved_size must be greater than 
      </FONT><BR><FONT size=2>shared_pool_reserved_min_alloc.</FONT> </P>
      <P><FONT size=2>&nbsp; units&nbsp; :&nbsp; bytes</FONT> <BR><FONT 
      size=2>&nbsp; default:&nbsp; 0&nbsp; (no reserved list)</FONT> <BR><FONT 
      size=2>&nbsp; minimum:&nbsp; &gt; shared_pool_reserved_min_alloc</FONT> 
      <BR><FONT size=2>&nbsp; maximum:&nbsp; 1/2 shared_pool_size</FONT> </P>
      <P><FONT size=2>1.2&nbsp; shared_pool_reserved_min_alloc</FONT> <BR><FONT 
      size=2>-----------------------------------</FONT> <BR><FONT size=2>The 
      init.ora parameter shared_pool_reserved_min_alloc controls 
      allocation</FONT> <BR><FONT size=2>for </FONT><BR><FONT size=2>the 
      reserved memory.&nbsp; Only allocations larger than </FONT><BR><FONT 
      size=2>shared_pool_reserved_min_alloc are allowed to allocate space from 
      the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT size=2>list if a 
      chunk of memory of sufficient size is not found on the shared</FONT> 
      <BR><FONT size=2>pool's </FONT><BR><FONT size=2>free lists.</FONT> </P>
      <P><FONT size=2>&nbsp; units&nbsp; :&nbsp; bytes</FONT> <BR><FONT 
      size=2>&nbsp; default:&nbsp; 5000</FONT> <BR><FONT size=2>&nbsp; 
      minimum:&nbsp; 5000</FONT> <BR><FONT size=2>&nbsp; maximum:&nbsp; &lt; 
      shared_pool_reserved_size</FONT> </P>
      <P><FONT size=2>The default value for shared_pool_reserved_min_alloc 
      should be adequate for </FONT><BR><FONT size=2>almost all systems.</FONT> 
      </P>
      <P><FONT size=2>2.&nbsp; CONTROLLING SPACE RECLAMATION OF THE SHARED 
      POOL</FONT> <BR><FONT 
      size=2>====================================================</FONT> 
      <BR><FONT size=2>RDBMS 7.1.5 also provides a new procedure, 
      aborted_request_threshold, in</FONT> <BR><FONT size=2>package</FONT> 
      <BR><FONT size=2>dbms_shared_pool, which allows users to set the limit on 
      the size of</FONT> <BR><FONT size=2>allocations</FONT> <BR><FONT 
      size=2>allowed to flush the shared pool if the free lists cannot satisfy 
      the</FONT> <BR><FONT size=2>request </FONT><BR><FONT size=2>size.</FONT> 
      </P>
      <P><FONT size=2>Before the RDBMS signals the ORA-04031 error, it 
      incrementally flushes</FONT> <BR><FONT size=2>unused </FONT><BR><FONT 
      size=2>objects from the shared pool until there is sufficient memory to 
      satisfy the</FONT> </P>
      <P><FONT size=2>allocation request.&nbsp; In most cases, incrementally 
      flushing objects from the </FONT><BR><FONT size=2>shared pool frees enough 
      memory for the allocation to complete succesfully.</FONT> <BR><FONT 
      size=2>If </FONT><BR><FONT size=2>the RDBMS signals an ORA-04031 error, it 
      has flushed all objects currently</FONT> <BR><FONT size=2>not 
      </FONT><BR><FONT size=2>in use on the system without finding a large 
      enough piece of contiguous</FONT> <BR><FONT size=2>memory.</FONT> </P>
      <P><FONT size=2>On a busy system, the larger the space allocation, the 
      more likely the RDBMS</FONT> </P>
      <P><FONT size=2>will signal the ORA-04031 error.&nbsp; Flushing all 
      objects, however, will impact</FONT> </P>
      <P><FONT size=2>other users on the system, possibly causing a degradation 
      in performance.</FONT> </P>
      <P><FONT size=2>The aborted_request_threshold procedure allows the DBA to 
      localize the</FONT> <BR><FONT size=2>impact </FONT><BR><FONT size=2>the 
      ORA-04031 error to the process that couldn't allocate memory. The</FONT> 
      <BR><FONT size=2>procedure</FONT> <BR><FONT size=2>takes a numeric value 
      between 5000 and 2147483647, representing the size, in</FONT> </P>
      <P><FONT size=2>bytes, of the threshold.</FONT> </P>
      <P><FONT size=2>3.&nbsp; NEW FIXED VIEW V$SHARED_POOL_RESERVED</FONT> 
      <BR><FONT size=2>=========================================</FONT> 
      <BR><FONT size=2>RDBMS 7.1.5 has a new fixed view to help tune the 
      reserved pool and space</FONT> <BR><FONT size=2>within</FONT> <BR><FONT 
      size=2>the shared pool.&nbsp; The name of the new fixed view is 
      V$SHARED_POOL_RESERVED</FONT> <BR><FONT size=2>and </FONT><BR><FONT 
      size=2>has the following columns:</FONT> </P>
      <P><FONT 
      size=2>Name&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      Null?&nbsp;&nbsp;&nbsp; Type</FONT> <BR><FONT 
      size=2>------------------------------- -------- --------------</FONT> 
      <BR><FONT 
      size=2>FREE_SPACE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>AVG_FREE_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>FREE_COUNT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>MAX_FREE_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>USED_SPACE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>AVG_USED_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>USED_COUNT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>MAX_USED_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>REQUESTS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>REQUEST_MISSES&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>LAST_MISS_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>MAX_MISS_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>REQUEST_FAILURES&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>LAST_FAILURE_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>ABORTED_REQUEST_THRESHOLD&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>ABORTED_REQUESTS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> <BR><FONT 
      size=2>LAST_ABORTED_SIZE&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      NUMBER</FONT> </P>
      <P><FONT size=2>These columns of V$SHARED_POOL_RESERVED are only valid if 
      the parameter</FONT> <BR><FONT size=2>shared_pool_reserved_size is set to 
      a valid value.</FONT> </P>
      <P><FONT size=2>&nbsp; FREE_SPACE is the total amount of free space on the 
      reserved list.</FONT> <BR><FONT size=2>&nbsp; </FONT><BR><FONT 
      size=2>&nbsp; AVG_FREE_SIZE is the average size of the free memory on the 
      reserved list.</FONT> </P>
      <P><FONT size=2>&nbsp; FREE_COUNT is the number of free pieces of memory 
      on the reserved list.</FONT> </P>
      <P><FONT size=2>&nbsp; MAX_FREE_SIZE is the size of the largest free piece 
      of memory on the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      list.</FONT> </P>
      <P><FONT size=2>&nbsp; USED_SPACE is the total amount of used memory on 
      the reserved list.</FONT> </P>
      <P><FONT size=2>&nbsp; AVG_USED_SIZE is the average size of the of the 
      used memory on the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      list.</FONT> </P>
      <P><FONT size=2>&nbsp; USED_COUNT is the number of used pieces of memory 
      on the reserved list.</FONT> </P>
      <P><FONT size=2>&nbsp; MAX_USED_SIZE is the size of the largest used piece 
      of memory on the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      list.</FONT> </P>
      <P><FONT size=2>&nbsp; REQUESTS is the number of times that the reserved 
      list was searched for a </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; free 
      piece of memory.</FONT> </P>
      <P><FONT size=2>&nbsp; REQUEST_MISSES is the number of times the reserved 
      list didn't have a free</FONT> </P>
      <P><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      piece of memory to satisfy the request, and proceeded to</FONT> <BR><FONT 
      size=2>start</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      flushing objects from the LRU list.</FONT> </P>
      <P><FONT size=2>&nbsp; LAST_MISS_SIZE is the request size of the last 
      REQUEST_MISS.</FONT> </P>
      <P><FONT size=2>&nbsp; MAX_MISS_SIZE is the request size of the largest 
      REQUEST_MISS.</FONT> </P>
      <P><FONT size=2>The next set of columns contain values which are valid 
      even if</FONT> <BR><FONT size=2>shared_pool_reserved_size is not 
      set.</FONT> </P>
      <P><FONT size=2>&nbsp; REQUEST_FAILURES is the number of times that no 
      memory was found to</FONT> <BR><FONT size=2>satisfy a</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      request (example: number of times ORA-04031 occurred)</FONT> </P>
      <P><FONT size=2>&nbsp; LAST_FAILURE_SIZE is the request size of the last 
      failed request </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      (example: the request size of last ORA-04031).</FONT> </P>
      <P><FONT size=2>&nbsp; ABORTED_REQUEST_THRESHOLD is the minimum size of a 
      request which will</FONT> <BR><FONT size=2>signal </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      an ORA-04031 error without flushing objects. See</FONT> </P>
      <P><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      the procedure aborted_request_threshold</FONT> <BR><FONT size=2>described 
      </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      above.</FONT> </P>
      <P><FONT size=2>&nbsp; LAST_ABORTED_SIZE is the last size of the request 
      which returned an</FONT> <BR><FONT size=2>ORA-04031 </FONT><BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      error without flushing objects from the LRU list.</FONT> </P><BR>
      <P><FONT size=2>4.&nbsp; TUNING HINTS BASED ON 
      V$SHARED_POOL_RESERVED</FONT> <BR><FONT 
      size=2>================================================</FONT> <BR><FONT 
      size=2>Information in V$SHARED_POOL_RESERVED can help to set values 
      for</FONT> <BR><FONT size=2>shared_pool_reserved_size and even 
      shared_pool_size. This section assumes</FONT> <BR><FONT size=2>the 
      </FONT><BR><FONT size=2>DBA has performed all other shared pool tuning on 
      his system.</FONT> </P>
      <P><FONT size=2>4.1&nbsp; Initial Value for 
      shared_pool_reserved_size</FONT> <BR><FONT 
      size=2>------------------------------------------------</FONT> <BR><FONT 
      size=2>The DBA should make shared_pool_reserved_size 10% of the 
      shared_pool_size.</FONT> <BR><FONT size=2>For</FONT> <BR><FONT size=2>most 
      systems, this value should be sufficient, if the DBA has already 
      spent</FONT> <BR><FONT size=2>time</FONT> <BR><FONT size=2>tuning the 
      shared pool.</FONT> </P>
      <P><FONT size=2>4.2&nbsp; Initial Value for 
      shared_pool_reserved_min_alloc</FONT> <BR><FONT 
      size=2>-----------------------------------------------------</FONT> 
      <BR><FONT size=2>In most cases, the default value for this parameter is 
      adequate. If the DBA </FONT><BR><FONT size=2>increases this value, then 
      the RDBMS will allow fewer allocations from the </FONT><BR><FONT 
      size=2>reserved list and will request more memory from the shared pool 
      list.</FONT> </P>
      <P><FONT size=2>4.4&nbsp; Tuning shared_pool_reserved_size</FONT> 
      <BR><FONT size=2>-------------------------------------</FONT> <BR><FONT 
      size=2>Ideally, shared_pool_reserved_size should be made large enough to 
      satisfy</FONT> <BR><FONT size=2>any </FONT><BR><FONT size=2>request 
      scanning for memory on the reserved list without flushing objects</FONT> 
      <BR><FONT size=2>from </FONT><BR><FONT size=2>the shared pool.&nbsp; The 
      amount of operating system memory, however, may</FONT> <BR><FONT 
      size=2>constrain</FONT> <BR><FONT size=2>the size of the SGA, and 
      therefore the size of the shared pool such that</FONT> <BR><FONT 
      size=2>this </FONT><BR><FONT size=2>is not a feasible goal.</FONT> </P>
      <P><FONT size=2>If the DBA has a system with ample free memory to increase 
      his SGA, the goal</FONT> <BR><FONT size=2>is </FONT><BR><FONT size=2>to 
      have:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; REQUEST_MISS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = 
      0</FONT> </P>
      <P><FONT size=2>If the DBA is constrained for OS memory, his goal 
      is:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; REQUEST_FAILURES&nbsp; = 0 or not 
      increasing</FONT> <BR><FONT size=2>&nbsp;&nbsp; LAST_FAILURE_SIZE &gt; 
      shared_pool_reserved_min_alloc</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      AVG_FREE_SIZE&nbsp;&nbsp;&nbsp;&nbsp; &gt; 
      shared_pool_reserved_min_alloc</FONT> </P>
      <P><FONT size=2>If neither of these goals are met, increase 
      shared_pool_reserved_size; the</FONT> <BR><FONT size=2>DBA 
      </FONT><BR><FONT size=2>also needs to increase shared_pool_size by the 
      same amount, since the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT 
      size=2>list is taken from the shared pool.</FONT> </P>
      <P><FONT size=2>4.5&nbsp; shared_pool_reserved_size too low</FONT> 
      <BR><FONT size=2>--------------------------------------</FONT> <BR><FONT 
      size=2>The reserved pool is too small when:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; REQUEST_FAILURES&nbsp; &gt; 0 (and 
      increasing)</FONT> </P>
      <P><FONT size=2>and at least one of the following is true:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; LAST_FAILURE_SIZE &gt; 
      shared_pool_reserved_min_alloc</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      MAX_FREE_SIZE&nbsp;&nbsp;&nbsp;&nbsp; &lt; 
      shared_pool_reserved_min_alloc</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      FREE_MEMORY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &lt; 
      shared_pool_reserved_min_alloc</FONT> </P>
      <P><FONT size=2>The DBA has two options, depending on his SGA size 
      constraints:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; Increase shared_pool_reserved_size 
      and shared_pool_size, accordingly</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      o&nbsp; Increase shared_pool_reserved_min_alloc (but may need to 
      increase</FONT> <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      shared_pool_size)</FONT> </P>
      <P><FONT size=2>The first option will increase the amount of memory 
      available on the</FONT> <BR><FONT size=2>reserved </FONT><BR><FONT 
      size=2>list without impacting users not allocating memory from the 
      reserved list.</FONT> <BR><FONT size=2>The </FONT><BR><FONT size=2>second 
      options reduces the number of allocations allowed to use memory 
      from</FONT> <BR><FONT size=2>the</FONT> <BR><FONT size=2>reserved list; 
      doing so, however, will increase normal shared pool perhaps 
      </FONT><BR><FONT size=2>impacting other users on the system.</FONT> </P>
      <P><FONT size=2>4.6&nbsp; shared_pool_reserved_size too high</FONT> 
      <BR><FONT size=2>---------------------------------------</FONT> <BR><FONT 
      size=2>It is possible that too much memory has been allocated to the 
      reserved list.</FONT> </P>
      <P><FONT size=2>If:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; REQUEST_MISS&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = 
      0 or not increasing</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      FREE_MEMORY&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; = &gt; 50% of 
      shared_pool_reserved_size minimum</FONT> </P>
      <P><FONT size=2>The DBA has two options:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; Decrease 
      shared_pool_reserved_size</FONT> <BR><FONT size=2>&nbsp;&nbsp; o&nbsp; 
      Decrease shared_pool_reserved_min_alloc (if not the default</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; value)</FONT> </P>
      <P><FONT size=2>4.7&nbsp; shared_pool_size too small</FONT> <BR><FONT 
      size=2>-------------------------------</FONT> <BR><FONT size=2>The new 
      fixed table can also indicate when shared_pool_size is too small.&nbsp; 
      </FONT><BR><FONT size=2>If:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; REQUEST_FAILURES&nbsp;&nbsp; &gt; 0 and 
      increasing</FONT> <BR><FONT size=2>&nbsp;&nbsp; LAST_FAILURE_SIZE&nbsp; 
      &lt; shared_pool_reserved_min_alloc</FONT> </P>
      <P><FONT size=2>Then the DBA has two options if he has enabled the 
      reserved list:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; Decrease 
      shared_pool_reserved_size</FONT> <BR><FONT size=2>&nbsp;&nbsp; o&nbsp; 
      Decrease shared_pool_reserved_min_alloc (if set larger than the</FONT> 
      <BR><FONT size=2>default)</FONT> </P>
      <P><FONT size=2>Otherwise, the DBA the could:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; Increase shared_pool_size</FONT> 
      </P><BR>
      <P><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      APPENDIX 2:&nbsp; Procedure free_unused_memory</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      =========================================</FONT> </P>
      <P><FONT size=2>This text is also in the specification for this procedure 
      in dbmsutil.sql.</FONT> <BR><FONT size=2>It is part of package 
      dbms_session.</FONT> </P>
      <P><FONT size=2>Procedure free_unused_memory --</FONT> </P>
      <P><FONT size=2>Procedure for users to reclaim unused memory after 
      performing operations</FONT> <BR><FONT size=2>requiring large amounts of 
      memory (where large is &gt;100K).&nbsp; Note that this </FONT><BR><FONT 
      size=2>procedure should only be used in cases where memory is at a 
      premium.</FONT> </P>
      <P><FONT size=2>Examples operations using lots of memory are:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; large sorts where entire 
      sort_area_size is used and</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; sort_area_size is hundreds of 
      KB</FONT> <BR><FONT size=2>&nbsp;&nbsp; o&nbsp; compiling large PL/SQL 
      packages, procedures, or functions</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      o&nbsp; storing hundreds of KB of data within PL/SQL indexed tables</FONT> 
      </P>
      <P><FONT size=2>One can monitor user memory by tracking the statistics 
      "session uga memory"</FONT> <BR><FONT size=2>and</FONT> <BR><FONT 
      size=2>"session pga memory" in the v$sesstat/v$statname fixed views.&nbsp; 
      Monitoring</FONT> <BR><FONT size=2>these </FONT><BR><FONT 
      size=2>statistics will also show how much memory this procedure has freed. 
      The</FONT> <BR><FONT size=2>behavior</FONT> <BR><FONT size=2>of this 
      procedure depends upon the configuration of the server operating on 
      </FONT><BR><FONT size=2>behalf of the client:</FONT> </P>
      <P><FONT size=2>&nbsp;&nbsp; o&nbsp; dedicated server - returns unused PGA 
      memory to the OS</FONT> <BR><FONT size=2>&nbsp;&nbsp; o&nbsp; MTS 
      server&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; - returns unused session memory 
      to the shared_pool</FONT> </P>
      <P><FONT size=2>In order to free memory using this procedure, the memory 
      must not be in use.</FONT> </P>
      <P><FONT size=2>Once an operation allocates memory, only the same type of 
      operation can</FONT> <BR><FONT size=2>reuse </FONT><BR><FONT size=2>the 
      allocated memory.&nbsp; For example, once memory is allocated for sort, 
      even</FONT> <BR><FONT size=2>if </FONT><BR><FONT size=2>the sort is 
      complete and the memory is no longer in use, only another sort</FONT> 
      <BR><FONT size=2>can </FONT><BR><FONT size=2>reuse the sort-allocated 
      memory.&nbsp; For both sort and compilation, after the </FONT><BR><FONT 
      size=2>operation is complete, the memory is no longer in use and the user 
      can</FONT> <BR><FONT size=2>invoke </FONT><BR><FONT size=2>this procedure 
      to free the unused memory.</FONT> </P>
      <P><FONT size=2>An indexed table implicitly allocates memory to store 
      values assigned to the</FONT> </P>
      <P><FONT size=2>indexed table's elements.&nbsp; Thus, the more elements in 
      an indexed table, the</FONT> <BR><FONT size=2>more</FONT> <BR><FONT 
      size=2>memory the RDBMS allocates to the indexed table.&nbsp; As long as 
      there are</FONT> <BR><FONT size=2>elements </FONT><BR><FONT size=2>within 
      the indexed table, the memory associated with an indexed table is 
      in</FONT> <BR><FONT size=2>use.</FONT> </P>
      <P><FONT size=2>The scope of indexed tables determines how long their 
      memory is in use.</FONT> <BR><FONT size=2>Indexed </FONT><BR><FONT 
      size=2>tables declared globally are indexed tables declared in packages or 
      package </FONT><BR><FONT size=2>bodies.&nbsp; They allocate memory from 
      session memory.&nbsp; For an indexed table </FONT><BR><FONT 
      size=2>declared globally, the memory will remain in use for the lifetime 
      of a</FONT> <BR><FONT size=2>user's </FONT><BR><FONT size=2>login 
      (lifetime of a user's session), and is freed after the user</FONT> 
      <BR><FONT size=2>disconnects </FONT><BR><FONT size=2>from ORACLE.</FONT> 
      </P>
      <P><FONT size=2>Indexed tables declared locally are indexed tables 
      declared within</FONT> <BR><FONT size=2>functions, </FONT><BR><FONT 
      size=2>procedures, or anonymous blocks.&nbsp; These indexed tables 
      allocate memory from</FONT> <BR><FONT size=2>PGA</FONT> <BR><FONT 
      size=2>memory.&nbsp; For an indexed table declared locally, the memory 
      will remain in</FONT> <BR><FONT size=2>use </FONT><BR><FONT size=2>for as 
      long as the user is still executing the procedure, function, or</FONT> 
      <BR><FONT size=2>anonymous</FONT> <BR><FONT size=2>block in which the 
      indexed table is declared.&nbsp; After the procedure,</FONT> <BR><FONT 
      size=2>function, or</FONT> <BR><FONT size=2>anonymous block is finished 
      executing, the memory is then available for</FONT> <BR><FONT 
      size=2>other</FONT> <BR><FONT size=2>locally declared indexed tables to 
      use (i.e., the memory is no longer in</FONT> <BR><FONT size=2>use).</FONT> 
      </P>
      <P><FONT size=2>Assigning an uninitialized, "empty," indexed table to an 
      existing index</FONT> <BR><FONT size=2>table is</FONT> <BR><FONT size=2>a 
      method to explicitly re-initialize the indexed table and the memory</FONT> 
      <BR><FONT size=2>associated</FONT> <BR><FONT size=2>with the indexed 
      table.&nbsp; After this operation, the memory associated with</FONT> 
      <BR><FONT size=2>the </FONT><BR><FONT size=2>indexed table will no longer 
      be in use, making it available to be freed by </FONT><BR><FONT 
      size=2>calling this procedure.&nbsp; This method is particularly useful on 
      indexed</FONT> <BR><FONT size=2>tables </FONT><BR><FONT size=2>declared 
      globally which can grow during the lifetime of a user's session, as</FONT> 
      </P>
      <P><FONT size=2>long as the user no longer needs the contents of the 
      indexed table.</FONT> </P>
      <P><FONT size=2>The memory rules associated with an indexed table's scope 
      still apply; this </FONT><BR><FONT size=2>method and this procedure, 
      however, allow users to intervene and to</FONT> <BR><FONT size=2>explictly 
      </FONT><BR><FONT size=2>free the memory associated with an indexed 
      table.</FONT> </P>
      <P><FONT size=2>The PL/SQL fragment below illustrates the method and the 
      use of procedure </FONT><BR><FONT size=2>free_unused_user_memory.</FONT> 
      </P>
      <P><FONT size=2>&nbsp;create package foobar</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; type number_idx_tbl is table of number indexed by 
      binary_integer;</FONT> <BR><FONT size=2>&nbsp;&nbsp; store1_table&nbsp; 
      number_idx_tbl;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; PL/SQL indexed 
      table</FONT> <BR><FONT size=2>&nbsp;&nbsp; store2_table&nbsp; 
      number_idx_tbl;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; PL/SQL indexed 
      table</FONT> <BR><FONT size=2>&nbsp;&nbsp; store3_table&nbsp; 
      number_idx_tbl;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; PL/SQL indexed 
      table</FONT> <BR><FONT size=2>&nbsp;&nbsp; ...</FONT> <BR><FONT 
      size=2>&nbsp;end;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
      --&nbsp; end of foobar</FONT> <BR><FONT size=2>&nbsp;declare</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp; ...</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      empty_table&nbsp;&nbsp; number_idx_tbl;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; 
      uninitialized ("empty") version</FONT> <BR><FONT size=2>&nbsp;begin</FONT> 
      <BR><FONT size=2>&nbsp;&nbsp; for i in 1..1000000 loop</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp;&nbsp;&nbsp; store1_table(i) := 
      i;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; 
      load data</FONT> <BR><FONT size=2>&nbsp;&nbsp; end loop;</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; ...</FONT> <BR><FONT size=2>&nbsp;&nbsp; store1_table 
      := empty_table;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; "truncate" the 
      indexed table</FONT> <BR><FONT size=2>&nbsp;&nbsp; ...</FONT> <BR><FONT 
      size=2>&nbsp;&nbsp; -</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      dbms_session.free_unused_user_memory;&nbsp; -- give memory back to 
      system</FONT> <BR><FONT size=2>&nbsp;&nbsp; store1_table(1) := 
      100;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; 
      index tables still declared;</FONT> <BR><FONT size=2>&nbsp;&nbsp; 
      store2_table(2) := 
      200;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; --&nbsp; 
      but truncated.</FONT> <BR><FONT size=2>&nbsp;&nbsp; ...</FONT> <BR><FONT 
      size=2>&nbsp; end;</FONT> </P>
      <P><FONT size=2>&nbsp;Copyright (c) 1995,1999 Oracle Corporation. All 
Received on Thu Aug 24 2000 - 10:38:56 CDT

Original text of this message

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