Know More about Oracle PGA Memory

Question :   1. How can we know available PGA and temporary tablespace before we issue a huge operation? 2. Can we estimate PGA and temporary tablespace for a huge operation? 3. As we know, there is limitation for a user process. Can we set unlimited and how to do it?   We say:   PROPOSED SOLUTION(S) ==================== ANSWER ====== Note 223730.1 Title Automatic PGA Memory Management in 9i and 10g This Document briefly describes how Oracle 9i manage PGA work area and how to tune it and some of the common issues and some of the common misunderstood issues. Automatic PGA Memory Management Automatic PGA Memory Management ***Checked for relevance on 07-Jan-2011*** note:note: Automatic PGA Memory Management ------------------------------- Process Global Area,  often known as the Program Global Area (PGA) resides in the process private memory of the server process. It contains global variables and data structures and control information for a server process. example of such information is the runtime area of a cursor. Each time a cursor is executed, a new runtime area is created for that cursor in the PGA memory region of the server process executing that cursor. The performance of complex long running queries, typical in a DSS environment, depend to a large extent on the memory available in the Program Global Area (PGA). which is also called work area. The size of a work area can be controlled and tuned. Generally, bigger work areas can significantly improve the performance of a particular operator at the cost of higher memory consumption. Ideally, the size of a work area is big enough that it can accommodate the input data and auxiliary memory structures allocated by its associated SQL operator. This is known as the optimal size of a work area (e.g. a memory sort). When the size of the work area is smaller than optimal (e.g. a disk sort), the response time increases, because an extra pass is performed over part of the input data. This is known as the one-pass size of the work area. Under the one-pass threshold, when the size of a work area is far too small compared to the input data size, multiple passes over the input data are needed. This could dramatically increase the response time of the operator. This is known as the multi-pass size of the work area. In Oracle8i administrators sized the PGA by carefully adjusting a number of initialization parameters, such as, SORT_AREA_SIZE, HASH_AREA_SIZE, BITMAP_MERGE_AREA_SIZE, and CREATE_BITMAP_AREA_SIZE, etc. Starting with Oracle9i, an option is provided to completely automate the management of PGA memory.  Administrators merely need to specify the maximum amount of PGA memory available to an instance using a newly introduced initialization parameter PGA_AGGREGATE_TARGET. The database server automatically distributes this memory among various active queries in an intelligent manner so as to ensure maximum performance benefits and the most efficient utilization of memory. Furthermore, Oracle9i and newer releases can adapt itself to changing workload thus utilizing resources efficiently regardless of the load on the system. The amount of the PGA memory available to an instance can be changed dynamically by altering the value of the PGA_AGGREGATE_TARGET parameter making it possible to add to and remove PGA memory from an active instance online. Since the database engine itself is better equipped to determine SQL execution memory requirements, database administrators should use this feature and not try to tune the PGA manually. This should translate to better throughput for large number of users on the system as well as improved response time for queries. The automatic SQL execution memory management feature is enabled by setting the parameter WORKAREA_SIZE_POLICY to AUTO and by specifying a size of PGA_AGGREGATE_TARGET in the initialization file. These two parameters can also be set dynamically using the ALTER SYSTEM command. In the absence of either of these parameters, the database will revert to manual PGA management mode. In Oracle9i Release 2, an advisory for PGA_AGGREGATE_TARGET was introduced. Just like in Buffer Cache Advisory, the PGA Advisory will suggest the appropriate size for PGA memory and thus make PGA tuning an even simpler task. Version specific notes: Until 9iR2, PGA_AGGREGATE_TARGET parameter controls the sizing of workareas for all dedicated server connections, but it has no effect on shared servers (aka MTS) connections and the *_AREA_SIZE parameters will take precedence in this case. In 10g, PGA_AGGREGATE_TARGET controls workareas allocated by both dedicated and shared connections. As of 11g, Automatic Memory Management (AMM) expands to managing both SGA and PGA memory.   Under memory pressure for PGA memory, SGA memory will be re-allocated for use by a process to accommodate workarea needs.  On the flip-side, if PGA memory is under allocated, memory can be added to the auto-tuned components in the SGA beyond the original SGA configuration. NOTE:   With AMM, setting an explicit value for PGA_AGGREGATE_TARGET will act as a minimum setting that AMM will not shrink below.  See note:443746.1 for more information. How To Tune PGA_AGGREGATE_TARGET ---------------------------------- The first question we will have when we set this parameter is what is the best value for it? To determine the appropriate setting for PGA_AGGREGATE_TARGET  parameter we recommend to follow the following steps 1- Make a first estimate for PGA_AGGREGATE_TARGET  based on the following rule - For OLTP systems PGA_AGGREGATE_TARGET  = ( * 80%) * 20% - For DSS systems PGA_AGGREGATE_TARGET  = ( * 80%) * 50% So for example, if we have an Oracle instance configured on system with 16G of Physical memory, then the suggested PGA_AGGREGATE_TARGET parameter value we should start with incase we have OLTP system is (16 G * 80%)*20% ~= 2.5G and incase we have DSS system is (16 G * 80%)* 50% ~= 6.5 G. In the above equation, we assume that 20% of the memory will be used by the OS, and in OLTP system 20% of the remaining memory will be used for PGA_AGGREGATE_TARGET and the remaining memory is going for Oracle SGA memory and non-oracle processes memory. So make sure that you have enough memory for your SGA and also for non-oracle processes 2- A second step in tuning the PGA_AGGREGATE_TARGET is to monitor performance using available PGA statistics and see if PGA_AGGREGATE_TARGET is under sized or over sized. Several dynamic performance views are available for this purpose: - V$PGASTAT This view provides instance-level statistics on the PGA memory usage and the automatic PGA memory manager. For example: SELECT * FROM V$PGASTAT; NAME                                               VALUE -------------------------------------------------------- aggregate PGA target parameter                     524288000 bytes aggregate PGA auto target                          463435776 bytes global memory bound                                25600 bytes total PGA inuse                                    9353216 bytes total PGA allocated                                73516032 bytes maximum PGA allocated                              698371072 bytes total PGA used for auto workareas                  0 bytes maximum PGA used for auto workareas                560744448 bytes total PGA used for manual workareas                0 bytes maximum PGA used for manual workareas              0 bytes over allocation count                              0 bytes total bytes processed                              4.0072E+10 bytes total extra bytes read/written                     3.1517E+10 bytes cache hit percentage                               55.97 percent Main statistics to look at (a) aggregate PGA auto target : This gives the amount of PGA memory Oracle can use for work areas running in automatic mode. This part of memory represent the tunable part of PGA memory,i.e. memory allocated for intensive memory SQL operators like sorts, hash-join, group-by, bitmap merge and bitmap index create. This memory part can be shrinked/expanded in function of the system load. Other parts of PGA memory are known as untunable, i.e. they require a size that can't be negociated (e.g. context information for each session, for each open/active cursor, PL/SQL or Java memory). So, the aggregate PGA auto target should not be small compared to the value of PGA_AGGREGATE_TARGET. You must ensure that enough PGA memory is left for work areas running in automatic mode. (b) total PGA used for auto workarea: This gives the actual tunable PGA memory used by the system. The 'maximum PGA used for auto workareas' gives the maximum reached by previous statistic since instance startup. (c) total PGA in used: This gives the total PGA memory in use. The detail of this value can be found in the PGA_USED_MEM column of the v$process view. Oracle92, 10g, 11g: (d) over allocation count: Over-allocating PGA memory can happen if the value of PGA_AGGREGATE_TARGET is too small to accommodate the untunable PGA memory part plus the minimum memory required to execute the work area workload. When this happens, Oracle cannot honor the initialization parameter PGA_AGGREGATE_TARGET, and extra PGA memory needs to be allocated. over allocation count is the number of time the system was detected in this state since database startup. This count should ideally be equal to zero. Oracle92,10g, 11g: (e) cache hit percentage: This metric is computed by Oracle to reflect the performance of the PGA memory component. It is cumulative from instance start-up. A value of 100% means that all work areas executed by the system since instance start-up have used an optimal amount of PGA memory. This is, of course, ideal but rarely happens except maybe for pure OLTP systems. In reality, some work areas run one-pass or even multi-pass, depending on the overall size of the PGA memory. When a work area cannot run optimally, one or more extra passes is performed over the input data. This reduces the cache hit percentage in proportion to the size of the input data and the number of extra passes performed. this value if computed from the "total bytes processed" and "total extra bytes read/written" statistics available in the same view using the following formula: total bytes processed * 100 PGA Cache Hit Ratio =  ------------------------------------------------------ (total bytes processed + total extra bytes read/written) - V$SQL_WORKAREA_HISTOGRAM (Oracle92, 10g, 11g) This view shows the number of work areas executed with optimal memory size, one- pass memory size, and multi-pass memory size since instance start-up. Statistics in this view are subdivided into buckets that are defined by the optimal memory requirement of the work area. Each bucket is identified by a range of optimal memory requirements specified by the values of the columns LOW_OPTIMAL_SIZE and HIGH_OPTIMAL_SIZE. Example : The following query shows statistics for all nonempty buckets. SELECT LOW_OPTIMAL_SIZE/1024 low_kb,(HIGH_OPTIMAL_SIZE+1)/1024 high_kb, optimal_executions, onepass_executions, multipasses_executions FROM   v$sql_workarea_histogram WHERE  total_executions != 0; The result of the query might look like the following: LOW_KB HIGH_KB OPTIMAL_EXECUTIONS ONEPASS_EXECUTIONS MULTIPASSES_EXECUTIONS ------ ------- ------------------ ------------------ ---------------------- 8      16      156255             0                   0 16     32      150                0                   0 32     64      89                 0                   0 64     128     13                 0                   0 128    256     60                 0                   0 256    512     8                  0                   0 512    1024    657                0                   0 1024   2048    551                16                  0 2048   4096    538                26                  0 4096   8192    243                28                  0 8192   16384   137                35                  0 16384  32768   45                 107                 0 32768  65536   0                  153                 0 65536  131072  0                  73                  0 131072 262144  0                  44                  0 262144 524288  0                  22                  0 The query result shows that, in the 1024 KB to 2048 KB bucket, 551 work areas used an optimal amount of memory, while 16 ran in one-pass mode and none ran in multi-pass mode. It also shows that all work areas under 1 MB were able to run in optimal mode. You can also use V$SQL_WORKAREA_HISTOGRAM to find the percentage of times work areas were executed in optimal, one-pass, or multi-pass mode since start-up. Example : SELECT optimal_count, round(optimal_count*100/total, 2) optimal_perc, onepass_count, round(onepass_count*100/total, 2) onepass_perc, multipass_count, round(multipass_count*100/total, 2) multipass_perc FROM (SELECT decode(sum(total_executions), 0, 1, sum(total_executions)) total, sum(OPTIMAL_EXECUTIONS) optimal_count, sum(ONEPASS_EXECUTIONS) onepass_count, sum(MULTIPASSES_EXECUTIONS) multipass_count FROM   v$sql_workarea_histogram WHERE  low_optimal_size > 64*1024);   ---- for 64 K optimal size - V$SQL_WORKAREA_ACTIVE This view can be used to display the work areas that are active (or executing) in the instance. Small active sorts (under 64 KB) are excluded from the view. Use this view to precisely monitor the size of all active work areas and to determine if these active work areas spill to a temporary segment. Example : SELECT to_number(decode(SID, 65535, NULL, SID)) sid, operation_type OPERATION,trunc(EXPECTED_SIZE/1024) ESIZE, trunc(ACTUAL_MEM_USED/1024) MEM, trunc(MAX_MEM_USED/1024) "MAX MEM", NUMBER_PASSES PASS, trunc(TEMPSEG_SIZE/1024) TSIZE FROM V$SQL_WORKAREA_ACTIVE ORDER BY 1,2; SID OPERATION         ESIZE     MEM       MAX MEM    PASS TSIZE --- ----------------- --------- --------- --------- ----- ------- 8   GROUP BY (SORT)   315       280       904         0 8   HASH-JOIN         2995      2377      2430        1   20000 9   GROUP BY (SORT)   34300     22688     22688       0 11  HASH-JOIN         18044     54482     54482       0 12  HASH-JOIN         18044     11406     21406       1   120000 This output shows that session 12 (column SID) is running a hash-join having its work area running in one-pass mode (PASS column). This work area is currently using 11406 KB of memory (MEM column) and has used, in the past, up to 21406 KB of PGA memory (MAX MEM column). It has also spilled to a temporary segment of size 120000 KB. Finally, the column ESIZE indicates the maximum amount of memory that the PGA memory manager expects this hash-join to use. This maximum is dynamically computed by the PGA memory manager according to workload. When a work area is deallocated—that is, when the execution of its associated SQL operator is complete—the work area is automatically removed from the V$SQL_WORKAREA_ACTIVE view. - note: have some other queries we use to monitor SQL execution memory 3- The Third and last step is tuning the PGA_AGGREGATE_TARGET. In Oracle 9i Release 2 we have 2 new views that help us in this task - V$PGA_TARGET_ADVICE - V$PGA_TARGET_ADVICE_HISTOGRAM By examining these two views, you will be able to determine how key PGA statistics will be impacted if you change the value of PGA_AGGREGATE_TARGET. To enable automatic generation of PGA advice performance views, make sure the following parameters are set: - PGA_AGGREGATE_TARGET - STATISTICS_LEVEL. Set this to TYPICAL (the default) or ALL; setting this parameter to BASIC turns off generation of PGA performance advice views. The content of these PGA advice performance views is reset at instance start-up or when PGA_AGGREGATE_TARGET is altered.   NOTE:  PGA_AGGREGATE can change automatically over time starting with 11g as part of the Automatic Memory Management enhancements available at 11g.   See note:443746.1 for more details. V$PGA_TARGET_ADVICE view predicts how the statistics cache hit percentage and over allocation count in V$PGASTAT will be impacted if you change the value of the initialization parameter PGA_AGGREGATE_TARGET. The following select statement can be used to find this information SELECT round(PGA_TARGET_FOR_ESTIMATE/1024/1024) target_mb, ESTD_PGA_CACHE_HIT_PERCENTAGE cache_hit_perc, ESTD_OVERALLOC_COUNT FROM   v$pga_target_advice; The output of this query might look like the following: TARGET_MB  CACHE_HIT_PERC ESTD_OVERALLOC_COUNT ---------- -------------- -------------------- 63         23             367 125        24             30 250        30             3 375        39             0 500        58             0 600        59             0 700        59             0 800        60             0 900        60             0 1000       61             0 1500       67             0 2000       76             0 3000       83             0 4000       85             0 From the above results we should set the PGA_AGGREGATE_TARGET parameter to a value where we avoid any over allocation, so lowest PGA_AGGREGATE_TARGET value we can set is 375 ( where ESTD_OVERALLOC_COUNT is 0) After eliminating over-allocations, the goal is to maximize the PGA cache hit percentage, based on your response-time requirement and memory constraints. V$PGA_TARGET_ADVICE_HISTOGRAM view predicts how the statistics displayed by the performance view V$SQL_WORKAREA_HISTOGRAM will be impacted if you change the value of the initialization parameter PGA_AGGREGATE_TARGET. You can use the dynamic view V$PGA_TARGET_ADVICE_HISTOGRAM to view detailed information on the predicted number of optimal, one-pass and multi-pass work area executions for the set of PGA_AGGREGATE_TARGET values you use for the prediction. Common issues --------------- 1- When we set the PGA_AGGREGATE_TARGET  and WORKAREA_SIZE_POLICY to auto then the *_area_size parameter are automatically ignored and oracle will automatically use the computed value for these parameters. 2- In Oracle 8i and earlier, the PGA memory was static, once the process started and started to allocate memory for it's PGA area then it will not release it back to the OS unless the process exits or dies. But the OS and under heavy memory pressure will decide to page out unused memory pages belongs to a process PGA to the swap space. In Oracle 9i and under the automatic PGA memory management, Oracle will be able to unallocate memory from a process PGA which is not using it any more so another process can use it, also it can adjust the different work areas size to accommodate the current workload and the amount of memory can be used. 3- Using automatic PGA memory management feature will help limiting resources used by oracle process, and will use it more efficiently. 4- Using automatic PGA memory management will help also reducing the possibility of getting ora-4030 errors unless we hit a OS limit, because work area sizes will be controlled and adjusted automatically based on the PGA_AGGGREGATE_TARGET parameter first and then the current work load. 5- If column ESTD_OVERALLOCATION_COUNT in the V$PGA_TARGET_ADVICE VIEW is nonzero, It indicates that PGA_AGGREGATE_TARGET is too small to even meet the minimum PGA memory needs. If PGA_AGGREGATE_TARGET is set within the over allocation zone, the memory manager will over-allocate memory and actual PGA memory consumed will be more than the limit you set. It is therefore meaningless to set a value of PGA_AGGREGATE_TARGET in that zone. 6- Some customer reported that SQL LOADER in Oracle 9i is slower than SQL Loader in Oracle 8i, and example of this is bug: which was closed as not a bug. Using PGA_AGGREGATE_TARGET alleviated the problem. 7- PGA_AGGREGATE_TARGET  is not supported on VMS, for more information please refer to note: "Oracle9i Release Notes Release 1 (9.0.1) for Alpha OpenVMS". ORA-3113 is returned on instance startup when init.ora PGA_AGGREGATE_TARGET is set. 8- Setting PGA_AGGREGATE_TARGET in 9.0.1 on HP-UX 11.0 may panic the OS. for more information please refer to note: "ALERT HP-UX  Patch Levels Advised" and Bug:2122307. Known Bugs ----------- - bug: Details: Attempting to set pga_aggregate_target over 4000Gb should error with ORA-4032 but no error is signalled. Fixed-Releases: 9.2.0.1.0 - Bug:2122307 HP System crash when setting PGA_AGGREGATE_TARGET to 10M or more in Oracle 9.0.1. This is basically an OS Problem that cause the crash. The system call pattern automatic PGA management is using causing HP/UX to try to extend fixed region and leads to memory allocation failures. To resolve the bug both this patch and PHKL_25188 (or later) must be installed.   As the DBA you need to get confirmation from your operating system administrator that the amount of memory reported as being in use by a process includes or does not include shared memory. If shared memory is included in the value displayed by the operating system utility, then the shared pool size must be deducted from that value to know how much private memory the process is actually using. See note 174555.1 "UNIX  Determining the Size of an Oracle Process". If an RDBMS user process is using more private memory than expected, then the DBA has three options: - Do nothing - Monitor the RDBMS iuser session to find out what SQL statements are being performed or were being performed by that RDBMS session. Using the SQL*Trace functionality of the database would normally be done if information from the end user cannot be obtained directly as to what they were doing since the memory usage was higher than expected or what they are doing right now. - Kill that RDBMS user session. PGA_AGGREGATE_TARGET does not set a hard limit on pga usage. It is only a target value used to dynamically size the process work areas. It also does not affect other areas of the pga that are allowed to grow beyond this limit. There are certain areas of pga that cannot be controlled by initialization parameters. Such areas include pl/sql memory collections such as pl/sql tables and varrays. Depending on the programming code and amount of data being handled these areas can grow very large (up to 20G internal limit on 10) and can consume large amounts of memory. This memory growth can be controlled by good programming practices. As an example, use LIMIT clause with BULK COLLECT. Additionally, programming mistakes can also lead to excessive memory usage. You can take steps to control the size of a process. However, from within the database framework you cannot place a hard limit on the size of a process by setting any initialization parameters or database configuration. You can limit the size of a process from the OS side by setting kernel limits or user shell limits but this leads to the ORA-4030 and will cause transaction rollback. As noted in bug 7279150, "... this is not a hard limit and that we will exceed it when it is undersized and the workload increases rapidly, such as when they start the workload for their testing or when they spawn a new set of sessions from their application servers."   During the execution of SQL statements, server processes consume memory for various operations. Part of this memory is assigned to work areas for operations such as sorts and hash joins. Ideally, a work area should be large enough to support the SQL operation being performed. This size is known as the optimal size of a work area. When a work area is smaller than its optimal size, an extra pass is performed over part of the input data. This is known as the one-pass size of the work area. When the size of a work area is far too small compared to the input data size, multiple passes over the input data are needed. This is known as the multi-pass size of the work area. Operations performed in one-pass and multi-pass work area sizes increase response times, sometimes dramatically in the case of the latter. You can set the size of the various work areas with individual initialization parameters but the same amount of memory is allocated to each process. So if your parameters are set to avoid any multi-pass operations, processes will be assigned this much memory even for operations that could run in a much smaller optimal size. With Oracle9i, you can use the PGA_AGGREGATE_TARGET initialization parameter to assign memory that is shared by the server processes active in the instance and is automatically reallocated to the processes that currently need it. In the initial release of Oracle9i, additional information was added to various dynamic performance tables to help you monitor whether you had set a sufficiently high value for the PGA target. However, these statistics do not provide a lot of detail nor give you any guidance for setting a more appropriate value when it is under- or over-sized. Additional statistics are available in Oracle9i Database Release 2 to help you monitor and tune the PGA_AGGREGATE_TARGET parameter. Some of these statistics are provided through new views and others through modified views. In this section of the lesson, you can find more details about managing your PGA memory with the views listed above.   The new statistics to help you monitor the performance of the PGA memory component for a particular value of PGA_AGGREGATE_TARGET are based on two concepts: work areas groups and a computed PGA cache hit percentage value. By grouping work areas based on their optimal memory requirement, statistics can be collected for each group based on the number of optimal, one-pass, and multi-pass executions for each of these work area groups. With this finer granularity of work area statistics than previously available, you can more accurately predict how much memory is required across the instance to maximize the number of optimal executions. The PGA cache hit percentage summarizes statistics on work area executions into a single measure of PGA performance for a given value of PGA_AGGREGATE_TARGET. The PGA cache hit percentage is derived from the number of work areas that run optimal, the number of passes for the non-optimal work areas, and the sizes of the work areas. The new view, v$sql_workarea_histogram, enables you to study the nature of the work area workload over a chosen time period. The work areas are split into 33 groups based on their optimal memory requirements. Each group is defined by the lower bound on its optimal memory requirement, displayed in the low_optimal_size column, and its upper bound, displayed in the high_optimal_size column. For each group, the view accumulates the number of executions using optimal, one-pass, multi-pass memory since instance start up. These are displayed in the optimal_executions, onepass-executions, and multipasses_executions columns respectively. The total_executions column contains the sum of the other three execution counts. To determine if you have set a good value for the PGA target size, query the v$sql_workarea_histogram view using a statement such as the following, which converts the low_optimal_size and high_optimal_size column values to kilobytes: SQL> SELECT low_optimal_size/1024                          AS low_kb, 2     (high_optimal_size+1)/1024                         AS high_kb, 3     ROUND(100*optimal_executions/total_executions)     AS optimal, 4     ROUND(100*onepass_executions/total_executions)     AS onepass, 5     ROUND(100*multipasses_executions/total_executions) AS multipass 6  FROM v$sql_workarea_histogram 7  WHERE total_executions != 0 8  ORDER by low_kb; Of course, as with any monitoring effort using dynamic views, you should issue the query at the beginning and at the end of a typical work period and use the differences between the two results to determine what activity occurred during that period. Although it would be ideal for all work areas to execute in the optimal size, this goal is typically achieved by over-allocating memory to the PGA. If you graph the results of your query as a histogram, you can quickly identify the location in the graph where work groups begin to use one-pass, or even multi-pass, sizes. In the example shown the former occurs with a work area size of about 200KB and the latter about 100MB. Based on your knowledge of the type of work being done and the current level of performance, you can decide if this is acceptable or if the PGA_AGGREGATE_TARGET parameter value should be changed.   If a work area runs in one-pass or multi-pass mode, extra bytes will be processed since one or more extra pass over the input data will be performed. A new statistic, PGA cache hit percentage, condenses in one global numerical metric the relative performance of the PGA memory component. The PGA cache hit percentage is the percentage formed from the ratio of the number of bytes that need to be processed when all work areas run in optimal mode to the total bytes really processed. The higher the PGA cache hit percentage, the better the performance of PGA memory and hence of the system. The PGA cache hit percentage statistic is stored in the v$pgastat view and can be retrieved from the row with the value cache hit percentage in the name column. A sample query is shown above. The v$sql_workarea and v$sql_workarea_active views have been modified in Oracle9i Database Release 2 to reflect information on the temporary segment an operator (work area) uses. The v$sort_usage view is renamed to v$tempseg_usage to reflect that information about all temporary segments, not only sort segments, is recorded in this view. In the current release, the old view name is being maintained for backward compatibility.   If you monitor the PGA space consumption with the various views provided for that purpose (v$sql_workarea_histogram, v$pgastat, v$sql_workarea, and v$sql_workarea_active), you may decide that you need to reset your PGA_AGGREGATE_TARGET initialization parameter value to make better use of your available memory. In some cases, you may want to reduce the memory allocated, in other cases; you may realize that you should increase the value. To help you determine by what factor you should change the parameter value, you can use two new views, provided in Oracle9i Database Release 2, that provide PGA sizing advice. These views, v$pga_target_advice and v$pga_target_advice_histogram, accumulate advice statistics to help you tune the PGA_AGGREGATE_TARGET value. The views are only populated if PGA_AGGREGATE_TARGET is set to a non-zero value that enables auto management of the PGA. Also the initialization parameter STATISTICS_LEVEL value must be set to Typical or All. Further, the view contents are refreshed when you shut down the instance or when you change the value of PGA_AGGREGATE_TARGET dynamically. Rows in these two views correspond to setting the PGA_AGGREGATE_TARGET value to 0.125, 0.25, 0.5, 0.75, 1, 1.2, 1.4, 1.6, 1.8, 2, 3, 4, 6, and 8 times its current value. However, if these values are either less than 10MB or more than 256GB, they will not be included in the output. To begin monitoring and tuning the PGA target size, you should issue query against the v$pga_target_advice view similar to: SQL> SELECT 2    ROUND(pga_target_for_estimate/1024/1024) AS target_mb, 3    estd_pga_cache_hit_percentage    AS cache_hit_percent, 4    estd_overalloc_count 5  FROM v$pga_target_advice 6  ORDER BY target_mb; For output that is easier to read and use, this query converts the value of the pga_target_for_estimate column from bytes to megabytes. As mentioned earlier, you should issue this query at the beginning and at the end of a typical work period and use the differences between the two results to obtain advise related to the work performed during that period. The output from the query on v$pga_target_advice contains up to 14 rows that contain estimated statistics based on the multiples of the current PGA_AGGREGATE_TARGET value. Assuming that your query produced the output shown above with the PGA_AGGREGATE_TARGET value set to 500MB, follow these steps to evaluate the results: Step 1: Identify the first row with a value of zero in the estd_overallocation_count column. The rows above this one are for PGA_AGGREGATE_TARGET values (shown in the target_mb column) that are too small for the minimum PGA memory needs. In this case, this is the fifth row, which contains data for the current PGA_AGGREGATE_TARGET value, 500MB. Had the target_mb column value in the first row with a zero estd_overallocation_count been larger than the current setting, you should increase the PGA_AGGREGATE_TARGET parameter to at least this size. Step 2: Examine the rows with PGA values larger than the minimum required to find the first pair of adjacent rows with values in the cache_hit_percent column that are within a few percentage points of each other. These rows indicate where, were you to graph the values, you would see an inflection point (sometimes referred to as a “knee”) in the curve. The optimal value for the PGA_AGGREGATE_TARGET parameter is at this inflection point, represented by the target_mb column value in the first of these two rows. Based on the above results, you should set the PGA_AGGREGATE_TARGET to 3000MB if you have sufficient memory. If you have even more memory available, you should assign it to some other use, such as one of the SGA components, rather than increasing the PGA target size.