Tuning PGA Memory: Area. Due To The Memory-Intensive Nature of These Operations, Tuning The
Tuning PGA Memory: Area. Due To The Memory-Intensive Nature of These Operations, Tuning The
PGA memory is the most critical memory parameter for the resource-intensive
queries found in a data warehouse. In a data warehouse, PGA memory is
used by SQL operations, such as sorts, hash joins, bitmap merges, and bulk
loads. The amount of PGA memory used by each operation is called its work
area. Due to the memory-intensive nature of these operations, tuning the
work areas is very crucial to ensure good query performance. If enough memory
is not available, intermediate data may need to be written to temporary
segments on disk, which can slow down performance significantly. Prior to
Oracle 9i, in order to tune query performance, the DBA had to tune various
initialization parameters, such as SORT_AREA_SIZE, HASH_AREA_
SIZE, CREATE_BITMAP_AREA_SIZE, and BITMAP_MERGE_AREA_
SIZE, to get good performance. However, this was a very difficult and timeconsuming
process because, the ideal values for these parameters may vary
from query to query and may depend on the load on the system. Oracle 9i
introduced the automatic PGA memory management feature to ease this burden.
Additionally, in Oracle Database 10g the PGA advisor can be used to
determine the ideal memory size setting for the system.
While a query execution is in progress, you can monitor the work area
usage in the V$SQL_WORKAREA_ACTIVE view.
Now that we understand how to monitor PGA memory usage and identify
any multipass executions, let us see how we go about tuning it.
Tuning PGA_AGGREGATE_TARGET
As we mentioned earlier, available physical memory on a system running
Oracle must be distributed among the Operating System, SGA, and PGA.
For a data warehouse, a good rule of thumb is to set the
PGA_AGGREGATE_TARGET initially to about 40 percent to 50 percent
of available physical memory and then tune it based on execution of a real
workload.
Oracle Database 10g will continuously monitor how much PGA memory
is being used by the entire instance by collecting statistics during executions
of queries. This information can be seen in the V$PGASTAT view.
Using these statistics, it estimates how the performance would vary if the
PGA_ AGGREGATE_TARGET were set to different values. This information
is published in the V$PGA_ TARGET_ADVICE view and can be
used to tune the PGA_AGGREGATE_TARGET parameter.
Hint: Set the initialization parameter STATISTICS_LEVEL to TYPICAL
(default) or ALL; otherwise, the V$PGA_TARGET_ADVICE view is not
available.
The following query shows the various statistics in V$PGASTAT:
The first step is to look at the top two lines of this output: the aggregate
PGA target parameter, which is the current setting of PGA_
AGGREGATE_TARGET, and the aggregate PGA auto target, which Oracle
has calculated as the total memory it can use for SQL work areas. The
difference is the estimated memory needed for other processing, such as PL/
SQL, and is not tuned by the automatic memory management feature. In
our example, the PGA_AGGREGATE_TARGET is 15.7MB, and the total
amount available for work areas is 4.2MB. Note that if you find the auto
target is much smaller than the PGA_AGGREGATE_TARGET, as is the
case in our example, this is one indication that there is not enough PGA
memory for work areas and you may need to increase it.
To confirm whether your PGA_AGGREGATE_SETTING is too small,
you should look at two quantities - the cache hit percentage and the overallocation
count, underlined in the preceding output. The cache hit percentage
indicates the percentage of work areas that operated with an optimal
allocation of memory. The overallocation count indicates how many
times Oracle had to step over the user-defined limit for PGA_
AGGREGATE_TARGET, because there was not enough memory available.
In a well-tuned system, the overallocation count should be zero,
meaning the available PGA memory was sufficient, and the cache hit percentage
should be over 80 percent, meaning most queries execute with the
optimal amount of memory. If you find that your cache hit percentage is
too low or the overallocation count is nonzero, you have insufficient PGA
memory. In our example, the cache hit ratio is 61 percent, which is low, and
the overallocation count is 426, which is not good.
In these cases you should look at the V$PGA_TARGET_ADVICE view
for advice. The view shows the projected values of the cache hit percentage
and the overallocation count for various memory sizes. For each memory
size, the FACTOR column shows which factor of the current memory setting
it is. For example, the row where the FACTOR column is 1 is the current
setting, in our example 15.7MB.
PURPOSE
-------
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.
This Note is continuation to Note 147806.1 and Note 148346.1 and should
provide more information about this New 9i Feature.
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 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.
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, 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.
Note:
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.
The first question we will have when we set this parameter is what is the best
value for it ?
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
- V$PGASTAT
This view provides instance-level statistics on the PGA memory usage and
the automatic PGA memory manager. For example:
NAME VALUE
--------------------------------------------------------
aggregate PGA target parameter 524288000 bytes
aggregate PGA auto target 463435776 bytes
global memory bound 25600 bytes
total PGA inuse 9353216 bytes
total PGA allocated 73516032 bytes
maximum PGA allocated 698371072 bytes
total PGA used for auto workareas 0 bytes
maximum PGA used for auto workareas 560744448 bytes
total PGA used for manual workareas 0 bytes
maximum PGA used for manual workareas 0 bytes
over allocation count 0 bytes
total bytes processed 4.0072E+10 bytes
total extra bytes read/written 3.1517E+10 bytes
cache hit percentage 55.97 percent
(a) aggregate PGA auto target : This gives the amount of PGA memory Oracle can
use for work areas running in automatic mode. This part of memory represent the
tunable part of PGA memory,i.e. memory allocated for intensive memory SQL operators
like sorts, hash-join, group-by, bitmap merge and bitmap index create. This memory
part can be shrinked/expanded in function of the system load. Other parts of
PGA memory are known as untunable, i.e. they require a size that can't be negociated
(e.g. context information for each session, for each open/active cursor,
PL/SQL or Java memory).
So, the aggregate PGA auto target should not be small compared to the value of
PGA_AGGREGATE_TARGET. You must ensure that enough PGA memory is left for work areas
running in automatic mode.
(b) total PGA used for auto workarea: This gives the actual tunable PGA memory used by
the system. The 'maximum PGA used for auto workareas' gives the maximum reached
by previous statistic since instance startup.
(c) total PGA in used: This gives the total PGA memory in use. The detail of this
value can be found in the PGA_USED_MEM column of the v$process view.
Oracle92 only:
(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 only:
(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:
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 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$SQL_WORKAREA_HISTOGRAM to find the percentage of times work
areas were executed in optimal, one-pass, or multi-pass mode since start-up.
Example :
- V$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 :
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$SQL_WORKAREA_ACTIVE view.
- Note 148346.1 have some other queries we use to monitor SQL execution memory
- V$PGA_TARGET_ADVICE
- V$PGA_TARGET_ADVICE_HISTOGRAM
By examining these two views, you will be able to determine how key PGA statistics
will be impacted if you change the value of PGA_AGGREGATE_TARGET.
To enable automatic generation of PGA advice performance views, make sure the
following parameters are set:
- PGA_AGGREGATE_TARGET
- STATISTICS_LEVEL. Set this to TYPICAL (the default) or ALL; setting this
parameter to BASIC turns off generation of PGA performance advice views.
The content of these PGA advice performance views is reset at instance start-up
or when PGA_AGGREGATE_TARGET is altered.
V$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.
After eliminating over-allocations, the goal is to maximize the PGA cache hit
percentage, based on your response-time requirement and memory constraints.
Common issues
---------------
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.
6- Some customer reported that SQL LOADER in Oracle 9i is slower than SQL Loader
in Oracle 8i, and example of this is Bug 2230115 which was closed as not a
bug. Using PGA_AGGREGATE_TARGET alleviated the problem.
Known Bugs
-----------
- Bug 2109788
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
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.
RELATED DOCUMENTS
-----------------
- Note 147806.1 Oracle9i New Feature Automated SQL Execution Memory Management
- Note 148346.1 Oracle9i Monitoring Automated SQL Execution Memory Management
- Note 175216.1 Oracle9i Release Notes Release 1 (9.0.1) for Alpha OpenVMS
- Note 43507.1 > ALERT HP-UX Patch Levels Advised
- Oracle 9i Database Performance tuning Guide and reference, Chapter 14