<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Resource Consumption</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK
REV="MADE"
HREF="mailto:pgsql-docs@postgresql.org"><LINK
REL="HOME"
TITLE="PostgreSQL 9.6.17 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Server Configuration"
HREF="runtime-config.html"><LINK
REL="PREVIOUS"
TITLE="Connections and Authentication"
HREF="runtime-config-connection.html"><LINK
REL="NEXT"
TITLE="Write Ahead Log"
HREF="runtime-config-wal.html"><LINK
REL="STYLESHEET"
TYPE="text/css"
HREF="stylesheet.css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=ISO-8859-1"><META
NAME="creation"
CONTENT="2020-02-24T19:56:52"></HEAD
><BODY
CLASS="SECT1"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="4"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.6.17 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="Connections and Authentication"
HREF="runtime-config-connection.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="runtime-config.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Chapter 19. Server Configuration</TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="Write Ahead Log"
HREF="runtime-config-wal.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RUNTIME-CONFIG-RESOURCE"
>19.4. Resource Consumption</A
></H1
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-MEMORY"
>19.4.1. Memory</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-SHARED-BUFFERS"
></A
><TT
CLASS="VARNAME"
>shared_buffers</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the amount of memory the database server uses for shared
memory buffers. The default is typically 128 megabytes
(<TT
CLASS="LITERAL"
>128MB</TT
>), but might be less if your kernel settings will
not support it (as determined during <SPAN
CLASS="APPLICATION"
>initdb</SPAN
>).
This setting must be at least 128 kilobytes. (Non-default
values of <TT
CLASS="SYMBOL"
>BLCKSZ</TT
> change the minimum.) However,
settings significantly higher than the minimum are usually needed
for good performance. This parameter can only be set at server start.
</P
><P
> If you have a dedicated database server with 1GB or more of RAM, a
reasonable starting value for <TT
CLASS="VARNAME"
>shared_buffers</TT
> is 25%
of the memory in your system. There are some workloads where even
large settings for <TT
CLASS="VARNAME"
>shared_buffers</TT
> are effective, but
because <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> also relies on the
operating system cache, it is unlikely that an allocation of more than
40% of RAM to <TT
CLASS="VARNAME"
>shared_buffers</TT
> will work better than a
smaller amount. Larger settings for <TT
CLASS="VARNAME"
>shared_buffers</TT
>
usually require a corresponding increase in
<TT
CLASS="VARNAME"
>max_wal_size</TT
>, in order to spread out the
process of writing large quantities of new or changed data over a
longer period of time.
</P
><P
> On systems with less than 1GB of RAM, a smaller percentage of RAM is
appropriate, so as to leave adequate space for the operating system.
Also, on Windows, large values for <TT
CLASS="VARNAME"
>shared_buffers</TT
>
aren't as effective. You may find better results keeping the setting
relatively low and using the operating system cache more instead. The
useful range for <TT
CLASS="VARNAME"
>shared_buffers</TT
> on Windows systems
is generally from 64MB to 512MB.
</P
></DD
><DT
><A
NAME="GUC-HUGE-PAGES"
></A
><TT
CLASS="VARNAME"
>huge_pages</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> Enables/disables the use of huge memory pages. Valid values are
<TT
CLASS="LITERAL"
>try</TT
> (the default), <TT
CLASS="LITERAL"
>on</TT
>,
and <TT
CLASS="LITERAL"
>off</TT
>.
</P
><P
> At present, this feature is supported only on Linux. The setting is
ignored on other systems when set to <TT
CLASS="LITERAL"
>try</TT
>.
</P
><P
> The use of huge pages results in smaller page tables and less CPU time
spent on memory management, increasing performance. For more details,
see <A
HREF="kernel-resources.html#LINUX-HUGE-PAGES"
>Section 18.4.5</A
>.
</P
><P
> With <TT
CLASS="VARNAME"
>huge_pages</TT
> set to <TT
CLASS="LITERAL"
>try</TT
>,
the server will try to use huge pages, but fall back to using
normal allocation if that fails. With <TT
CLASS="LITERAL"
>on</TT
>, failure
to use huge pages will prevent the server from starting up. With
<TT
CLASS="LITERAL"
>off</TT
>, huge pages will not be used.
</P
></DD
><DT
><A
NAME="GUC-TEMP-BUFFERS"
></A
><TT
CLASS="VARNAME"
>temp_buffers</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the maximum number of temporary buffers used by each database
session. These are session-local buffers used only for access to
temporary tables. The default is eight megabytes
(<TT
CLASS="LITERAL"
>8MB</TT
>). The setting can be changed within individual
sessions, but only before the first use of temporary tables
within the session; subsequent attempts to change the value will
have no effect on that session.
</P
><P
> A session will allocate temporary buffers as needed up to the limit
given by <TT
CLASS="VARNAME"
>temp_buffers</TT
>. The cost of setting a large
value in sessions that do not actually need many temporary
buffers is only a buffer descriptor, or about 64 bytes, per
increment in <TT
CLASS="VARNAME"
>temp_buffers</TT
>. However if a buffer is
actually used an additional 8192 bytes will be consumed for it
(or in general, <TT
CLASS="SYMBOL"
>BLCKSZ</TT
> bytes).
</P
></DD
><DT
><A
NAME="GUC-MAX-PREPARED-TRANSACTIONS"
></A
><TT
CLASS="VARNAME"
>max_prepared_transactions</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the maximum number of transactions that can be in the
<SPAN
CLASS="QUOTE"
>"prepared"</SPAN
> state simultaneously (see <A
HREF="sql-prepare-transaction.html"
>PREPARE TRANSACTION</A
>).
Setting this parameter to zero (which is the default)
disables the prepared-transaction feature.
This parameter can only be set at server start.
</P
><P
> If you are not planning to use prepared transactions, this parameter
should be set to zero to prevent accidental creation of prepared
transactions. If you are using prepared transactions, you will
probably want <TT
CLASS="VARNAME"
>max_prepared_transactions</TT
> to be at
least as large as <A
HREF="runtime-config-connection.html#GUC-MAX-CONNECTIONS"
>max_connections</A
>, so that every
session can have a prepared transaction pending.
</P
><P
> When running a standby server, you must set this parameter to the
same or higher value than on the master server. Otherwise, queries
will not be allowed in the standby server.
</P
></DD
><DT
><A
NAME="GUC-WORK-MEM"
></A
><TT
CLASS="VARNAME"
>work_mem</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the amount of memory to be used by internal sort operations
and hash tables before writing to temporary disk files. The value
defaults to four megabytes (<TT
CLASS="LITERAL"
>4MB</TT
>).
Note that for a complex query, several sort or hash operations might be
running in parallel; each operation will be allowed to use as much memory
as this value specifies before it starts to write data into temporary
files. Also, several running sessions could be doing such operations
concurrently. Therefore, the total memory used could be many
times the value of <TT
CLASS="VARNAME"
>work_mem</TT
>; it is necessary to
keep this fact in mind when choosing the value. Sort operations are
used for <TT
CLASS="LITERAL"
>ORDER BY</TT
>, <TT
CLASS="LITERAL"
>DISTINCT</TT
>, and
merge joins.
Hash tables are used in hash joins, hash-based aggregation, and
hash-based processing of <TT
CLASS="LITERAL"
>IN</TT
> subqueries.
</P
></DD
><DT
><A
NAME="GUC-MAINTENANCE-WORK-MEM"
></A
><TT
CLASS="VARNAME"
>maintenance_work_mem</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the maximum amount of memory to be used by maintenance
operations, such as <TT
CLASS="COMMAND"
>VACUUM</TT
>, <TT
CLASS="COMMAND"
>CREATE
INDEX</TT
>, and <TT
CLASS="COMMAND"
>ALTER TABLE ADD FOREIGN KEY</TT
>. It defaults
to 64 megabytes (<TT
CLASS="LITERAL"
>64MB</TT
>). Since only one of these
operations can be executed at a time by a database session, and
an installation normally doesn't have many of them running
concurrently, it's safe to set this value significantly larger
than <TT
CLASS="VARNAME"
>work_mem</TT
>. Larger settings might improve
performance for vacuuming and for restoring database dumps.
</P
><P
> Note that when autovacuum runs, up to
<A
HREF="runtime-config-autovacuum.html#GUC-AUTOVACUUM-MAX-WORKERS"
>autovacuum_max_workers</A
> times this memory
may be allocated, so be careful not to set the default value
too high. It may be useful to control for this by separately
setting <A
HREF="runtime-config-resource.html#GUC-AUTOVACUUM-WORK-MEM"
>autovacuum_work_mem</A
>.
</P
></DD
><DT
><A
NAME="GUC-REPLACEMENT-SORT-TUPLES"
></A
><TT
CLASS="VARNAME"
>replacement_sort_tuples</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> When the number of tuples to be sorted is smaller than this number,
a sort will produce its first output run using replacement selection
rather than quicksort. This may be useful in memory-constrained
environments where tuples that are input into larger sort operations
have a strong physical-to-logical correlation. Note that this does
not include input tuples with an <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>inverse</I
></SPAN
>
correlation. It is possible for the replacement selection algorithm
to generate one long run that requires no merging, where use of the
default strategy would result in many runs that must be merged
to produce a final sorted output. This may allow sort
operations to complete sooner.
</P
><P
> The default is 150,000 tuples. Note that higher values are typically
not much more effective, and may be counter-productive, since the
priority queue is sensitive to the size of available CPU cache, whereas
the default strategy sorts runs using a <I
CLASS="FIRSTTERM"
>cache
oblivious</I
> algorithm. This property allows the default sort
strategy to automatically and transparently make effective use
of available CPU cache.
</P
><P
> Setting <TT
CLASS="VARNAME"
>maintenance_work_mem</TT
> to its default
value usually prevents utility command external sorts (e.g.,
sorts used by <TT
CLASS="COMMAND"
>CREATE INDEX</TT
> to build B-Tree
indexes) from ever using replacement selection sort, unless the
input tuples are quite wide.
</P
></DD
><DT
><A
NAME="GUC-AUTOVACUUM-WORK-MEM"
></A
><TT
CLASS="VARNAME"
>autovacuum_work_mem</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the maximum amount of memory to be used by each
autovacuum worker process. It defaults to -1, indicating that
the value of <A
HREF="runtime-config-resource.html#GUC-MAINTENANCE-WORK-MEM"
>maintenance_work_mem</A
> should
be used instead. The setting has no effect on the behavior of
<TT
CLASS="COMMAND"
>VACUUM</TT
> when run in other contexts.
</P
></DD
><DT
><A
NAME="GUC-MAX-STACK-DEPTH"
></A
><TT
CLASS="VARNAME"
>max_stack_depth</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the maximum safe depth of the server's execution stack.
The ideal setting for this parameter is the actual stack size limit
enforced by the kernel (as set by <TT
CLASS="LITERAL"
>ulimit -s</TT
> or local
equivalent), less a safety margin of a megabyte or so. The safety
margin is needed because the stack depth is not checked in every
routine in the server, but only in key potentially-recursive routines
such as expression evaluation. The default setting is two
megabytes (<TT
CLASS="LITERAL"
>2MB</TT
>), which is conservatively small and
unlikely to risk crashes. However, it might be too small to allow
execution of complex functions. Only superusers can change this
setting.
</P
><P
> Setting <TT
CLASS="VARNAME"
>max_stack_depth</TT
> higher than
the actual kernel limit will mean that a runaway recursive function
can crash an individual backend process. On platforms where
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> can determine the kernel limit,
the server will not allow this variable to be set to an unsafe
value. However, not all platforms provide the information,
so caution is recommended in selecting a value.
</P
></DD
><DT
><A
NAME="GUC-DYNAMIC-SHARED-MEMORY-TYPE"
></A
><TT
CLASS="VARNAME"
>dynamic_shared_memory_type</TT
> (<TT
CLASS="TYPE"
>enum</TT
>)
</DT
><DD
><P
> Specifies the dynamic shared memory implementation that the server
should use. Possible values are <TT
CLASS="LITERAL"
>posix</TT
> (for POSIX shared
memory allocated using <TT
CLASS="LITERAL"
>shm_open</TT
>), <TT
CLASS="LITERAL"
>sysv</TT
>
(for System V shared memory allocated via <TT
CLASS="LITERAL"
>shmget</TT
>),
<TT
CLASS="LITERAL"
>windows</TT
> (for Windows shared memory), <TT
CLASS="LITERAL"
>mmap</TT
>
(to simulate shared memory using memory-mapped files stored in the
data directory), and <TT
CLASS="LITERAL"
>none</TT
> (to disable this feature).
Not all values are supported on all platforms; the first supported
option is the default for that platform. The use of the
<TT
CLASS="LITERAL"
>mmap</TT
> option, which is not the default on any platform,
is generally discouraged because the operating system may write
modified pages back to disk repeatedly, increasing system I/O load;
however, it may be useful for debugging, when the
<TT
CLASS="LITERAL"
>pg_dynshmem</TT
> directory is stored on a RAM disk, or when
other shared memory facilities are not available.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-DISK"
>19.4.2. Disk</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-TEMP-FILE-LIMIT"
></A
><TT
CLASS="VARNAME"
>temp_file_limit</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the maximum amount of disk space that a process can use
for temporary files, such as sort and hash temporary files, or the
storage file for a held cursor. A transaction attempting to exceed
this limit will be canceled.
The value is specified in kilobytes, and <TT
CLASS="LITERAL"
>-1</TT
> (the
default) means no limit.
Only superusers can change this setting.
</P
><P
> This setting constrains the total space used at any instant by all
temporary files used by a given <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> process.
It should be noted that disk space used for explicit temporary
tables, as opposed to temporary files used behind-the-scenes in query
execution, does <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
> count against this limit.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-KERNEL"
>19.4.3. Kernel Resource Usage</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-MAX-FILES-PER-PROCESS"
></A
><TT
CLASS="VARNAME"
>max_files_per_process</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the maximum number of simultaneously open files allowed to each
server subprocess. The default is one thousand files. If the kernel is enforcing
a safe per-process limit, you don't need to worry about this setting.
But on some platforms (notably, most BSD systems), the kernel will
allow individual processes to open many more files than the system
can actually support if many processes all try to open
that many files. If you find yourself seeing <SPAN
CLASS="QUOTE"
>"Too many open
files"</SPAN
> failures, try reducing this setting.
This parameter can only be set at server start.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-VACUUM-COST"
>19.4.4. Cost-based Vacuum Delay</A
></H2
><P
> During the execution of <A
HREF="sql-vacuum.html"
>VACUUM</A
>
and <A
HREF="sql-analyze.html"
>ANALYZE</A
>
commands, the system maintains an
internal counter that keeps track of the estimated cost of the
various I/O operations that are performed. When the accumulated
cost reaches a limit (specified by
<TT
CLASS="VARNAME"
>vacuum_cost_limit</TT
>), the process performing
the operation will sleep for a short period of time, as specified by
<TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
>. Then it will reset the
counter and continue execution.
</P
><P
> The intent of this feature is to allow administrators to reduce
the I/O impact of these commands on concurrent database
activity. There are many situations where it is not
important that maintenance commands like
<TT
CLASS="COMMAND"
>VACUUM</TT
> and <TT
CLASS="COMMAND"
>ANALYZE</TT
> finish
quickly; however, it is usually very important that these
commands do not significantly interfere with the ability of the
system to perform other database operations. Cost-based vacuum
delay provides a way for administrators to achieve this.
</P
><P
> This feature is disabled by default for manually issued
<TT
CLASS="COMMAND"
>VACUUM</TT
> commands. To enable it, set the
<TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> variable to a nonzero
value.
</P
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-VACUUM-COST-DELAY"
></A
><TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The length of time, in milliseconds, that the process will sleep
when the cost limit has been exceeded.
The default value is zero, which disables the cost-based vacuum
delay feature. Positive values enable cost-based vacuuming.
Note that on many systems, the effective resolution
of sleep delays is 10 milliseconds; setting
<TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> to a value that is
not a multiple of 10 might have the same results as setting it
to the next higher multiple of 10.
</P
><P
> When using cost-based vacuuming, appropriate values for
<TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> are usually quite small, perhaps
10 or 20 milliseconds. Adjusting vacuum's resource consumption
is best done by changing the other vacuum cost parameters.
</P
></DD
><DT
><A
NAME="GUC-VACUUM-COST-PAGE-HIT"
></A
><TT
CLASS="VARNAME"
>vacuum_cost_page_hit</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The estimated cost for vacuuming a buffer found in the shared buffer
cache. It represents the cost to lock the buffer pool, lookup
the shared hash table and scan the content of the page. The
default value is one.
</P
></DD
><DT
><A
NAME="GUC-VACUUM-COST-PAGE-MISS"
></A
><TT
CLASS="VARNAME"
>vacuum_cost_page_miss</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The estimated cost for vacuuming a buffer that has to be read from
disk. This represents the effort to lock the buffer pool,
lookup the shared hash table, read the desired block in from
the disk and scan its content. The default value is 10.
</P
></DD
><DT
><A
NAME="GUC-VACUUM-COST-PAGE-DIRTY"
></A
><TT
CLASS="VARNAME"
>vacuum_cost_page_dirty</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The estimated cost charged when vacuum modifies a block that was
previously clean. It represents the extra I/O required to
flush the dirty block out to disk again. The default value is
20.
</P
></DD
><DT
><A
NAME="GUC-VACUUM-COST-LIMIT"
></A
><TT
CLASS="VARNAME"
>vacuum_cost_limit</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> The accumulated cost that will cause the vacuuming process to sleep.
The default value is 200.
</P
></DD
></DL
></DIV
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> There are certain operations that hold critical locks and should
therefore complete as quickly as possible. Cost-based vacuum
delays do not occur during such operations. Therefore it is
possible that the cost accumulates far higher than the specified
limit. To avoid uselessly long delays in such cases, the actual
delay is calculated as <TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> *
<TT
CLASS="VARNAME"
>accumulated_balance</TT
> /
<TT
CLASS="VARNAME"
>vacuum_cost_limit</TT
> with a maximum of
<TT
CLASS="VARNAME"
>vacuum_cost_delay</TT
> * 4.
</P
></BLOCKQUOTE
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-BACKGROUND-WRITER"
>19.4.5. Background Writer</A
></H2
><P
> There is a separate server
process called the <I
CLASS="FIRSTTERM"
>background writer</I
>, whose function
is to issue writes of <SPAN
CLASS="QUOTE"
>"dirty"</SPAN
> (new or modified) shared
buffers. It writes shared buffers so server processes handling
user queries seldom or never need to wait for a write to occur.
However, the background writer does cause a net overall
increase in I/O load, because while a repeatedly-dirtied page might
otherwise be written only once per checkpoint interval, the
background writer might write it several times as it is dirtied
in the same interval. The parameters discussed in this subsection
can be used to tune the behavior for local needs.
</P
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-BGWRITER-DELAY"
></A
><TT
CLASS="VARNAME"
>bgwriter_delay</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Specifies the delay between activity rounds for the
background writer. In each round the writer issues writes
for some number of dirty buffers (controllable by the
following parameters). It then sleeps for <TT
CLASS="VARNAME"
>bgwriter_delay</TT
>
milliseconds, and repeats. When there are no dirty buffers in the
buffer pool, though, it goes into a longer sleep regardless of
<TT
CLASS="VARNAME"
>bgwriter_delay</TT
>. The default value is 200
milliseconds (<TT
CLASS="LITERAL"
>200ms</TT
>). Note that on many systems, the
effective resolution of sleep delays is 10 milliseconds; setting
<TT
CLASS="VARNAME"
>bgwriter_delay</TT
> to a value that is not a multiple of 10
might have the same results as setting it to the next higher multiple
of 10. This parameter can only be set in the
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-BGWRITER-LRU-MAXPAGES"
></A
><TT
CLASS="VARNAME"
>bgwriter_lru_maxpages</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> In each round, no more than this many buffers will be written
by the background writer. Setting this to zero disables
background writing. (Note that checkpoints, which are managed by
a separate, dedicated auxiliary process, are unaffected.)
The default value is 100 buffers.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-BGWRITER-LRU-MULTIPLIER"
></A
><TT
CLASS="VARNAME"
>bgwriter_lru_multiplier</TT
> (<TT
CLASS="TYPE"
>floating point</TT
>)
</DT
><DD
><P
> The number of dirty buffers written in each round is based on the
number of new buffers that have been needed by server processes
during recent rounds. The average recent need is multiplied by
<TT
CLASS="VARNAME"
>bgwriter_lru_multiplier</TT
> to arrive at an estimate of the
number of buffers that will be needed during the next round. Dirty
buffers are written until there are that many clean, reusable buffers
available. (However, no more than <TT
CLASS="VARNAME"
>bgwriter_lru_maxpages</TT
>
buffers will be written per round.)
Thus, a setting of 1.0 represents a <SPAN
CLASS="QUOTE"
>"just in time"</SPAN
> policy
of writing exactly the number of buffers predicted to be needed.
Larger values provide some cushion against spikes in demand,
while smaller values intentionally leave writes to be done by
server processes.
The default is 2.0.
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
><DT
><A
NAME="GUC-BGWRITER-FLUSH-AFTER"
></A
><TT
CLASS="VARNAME"
>bgwriter_flush_after</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Whenever more than <TT
CLASS="VARNAME"
>bgwriter_flush_after</TT
> bytes have
been written by the bgwriter, attempt to force the OS to issue these
writes to the underlying storage. Doing so will limit the amount of
dirty data in the kernel's page cache, reducing the likelihood of
stalls when an fsync is issued at the end of a checkpoint, or when
the OS writes data back in larger batches in the background. Often
that will result in greatly reduced transaction latency, but there
also are some cases, especially with workloads that are bigger than
<A
HREF="runtime-config-resource.html#GUC-SHARED-BUFFERS"
>shared_buffers</A
>, but smaller than the OS's page
cache, where performance might degrade. This setting may have no
effect on some platforms. The valid range is between
<TT
CLASS="LITERAL"
>0</TT
>, which disables forced writeback, and
<TT
CLASS="LITERAL"
>2MB</TT
>. The default is <TT
CLASS="LITERAL"
>512kB</TT
> on Linux,
<TT
CLASS="LITERAL"
>0</TT
> elsewhere. (If <TT
CLASS="SYMBOL"
>BLCKSZ</TT
> is not 8kB,
the default and maximum values scale proportionally to it.)
This parameter can only be set in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
file or on the server command line.
</P
></DD
></DL
></DIV
><P
> Smaller values of <TT
CLASS="VARNAME"
>bgwriter_lru_maxpages</TT
> and
<TT
CLASS="VARNAME"
>bgwriter_lru_multiplier</TT
> reduce the extra I/O load
caused by the background writer, but make it more likely that server
processes will have to issue writes for themselves, delaying interactive
queries.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="RUNTIME-CONFIG-RESOURCE-ASYNC-BEHAVIOR"
>19.4.6. Asynchronous Behavior</A
></H2
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
><A
NAME="GUC-EFFECTIVE-IO-CONCURRENCY"
></A
><TT
CLASS="VARNAME"
>effective_io_concurrency</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the number of concurrent disk I/O operations that
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> expects can be executed
simultaneously. Raising this value will increase the number of I/O
operations that any individual <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> session
attempts to initiate in parallel. The allowed range is 1 to 1000,
or zero to disable issuance of asynchronous I/O requests. Currently,
this setting only affects bitmap heap scans.
</P
><P
> For magnetic drives, a good starting point for this setting is the
number of separate
drives comprising a RAID 0 stripe or RAID 1 mirror being used for the
database. (For RAID 5 the parity drive should not be counted.)
However, if the database is often busy with multiple queries issued in
concurrent sessions, lower values may be sufficient to keep the disk
array busy. A value higher than needed to keep the disks busy will
only result in extra CPU overhead.
SSDs and other memory-based storage can often process many
concurrent requests, so the best value might be in the hundreds.
</P
><P
> Asynchronous I/O depends on an effective <CODE
CLASS="FUNCTION"
>posix_fadvise</CODE
>
function, which some operating systems lack. If the function is not
present then setting this parameter to anything but zero will result
in an error. On some operating systems (e.g., Solaris), the function
is present but does not actually do anything.
</P
><P
> The default is 1 on supported systems, otherwise 0. This value can
be overridden for tables in a particular tablespace by setting the
tablespace parameter of the same name (see
<A
HREF="sql-altertablespace.html"
>ALTER TABLESPACE</A
>).
</P
></DD
><DT
><A
NAME="GUC-MAX-WORKER-PROCESSES"
></A
><TT
CLASS="VARNAME"
>max_worker_processes</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the maximum number of background processes that the system
can support. This parameter can only be set at server start. The
default is 8.
</P
><P
> When running a standby server, you must set this parameter to the
same or higher value than on the master server. Otherwise, queries
will not be allowed in the standby server.
</P
></DD
><DT
><A
NAME="GUC-MAX-PARALLEL-WORKERS-PER-GATHER"
></A
><TT
CLASS="VARNAME"
>max_parallel_workers_per_gather</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the maximum number of workers that can be started by a single
<TT
CLASS="LITERAL"
>Gather</TT
> node. Parallel workers are taken from the
pool of processes established by
<A
HREF="runtime-config-resource.html#GUC-MAX-WORKER-PROCESSES"
>max_worker_processes</A
>. Note that the requested
number of workers may not actually be available at run time. If this
occurs, the plan will run with fewer workers than expected, which may
be inefficient. Setting this value to 0, which is the default,
disables parallel query execution.
</P
><P
> Note that parallel queries may consume very substantially more
resources than non-parallel queries, because each worker process is
a completely separate process which has roughly the same impact on the
system as an additional user session. This should be taken into
account when choosing a value for this setting, as well as when
configuring other settings that control resource utilization, such
as <A
HREF="runtime-config-resource.html#GUC-WORK-MEM"
>work_mem</A
>. Resource limits such as
<TT
CLASS="VARNAME"
>work_mem</TT
> are applied individually to each worker,
which means the total utilization may be much higher across all
processes than it would normally be for any single process.
For example, a parallel query using 4 workers may use up to 5 times
as much CPU time, memory, I/O bandwidth, and so forth as a query which
uses no workers at all.
</P
><P
> For more information on parallel query, see
<A
HREF="parallel-query.html"
>Chapter 15</A
>.
</P
></DD
><DT
><A
NAME="GUC-BACKEND-FLUSH-AFTER"
></A
><TT
CLASS="VARNAME"
>backend_flush_after</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Whenever more than <TT
CLASS="VARNAME"
>backend_flush_after</TT
> bytes have
been written by a single backend, attempt to force the OS to issue
these writes to the underlying storage. Doing so will limit the
amount of dirty data in the kernel's page cache, reducing the
likelihood of stalls when an fsync is issued at the end of a
checkpoint, or when the OS writes data back in larger batches in the
background. Often that will result in greatly reduced transaction
latency, but there also are some cases, especially with workloads
that are bigger than <A
HREF="runtime-config-resource.html#GUC-SHARED-BUFFERS"
>shared_buffers</A
>, but smaller
than the OS's page cache, where performance might degrade. This
setting may have no effect on some platforms. The valid range is
between <TT
CLASS="LITERAL"
>0</TT
>, which disables forced writeback,
and <TT
CLASS="LITERAL"
>2MB</TT
>. The default is <TT
CLASS="LITERAL"
>0</TT
>, i.e., no
forced writeback. (If <TT
CLASS="SYMBOL"
>BLCKSZ</TT
> is not 8kB,
the maximum value scales proportionally to it.)
</P
></DD
><DT
><A
NAME="GUC-OLD-SNAPSHOT-THRESHOLD"
></A
><TT
CLASS="VARNAME"
>old_snapshot_threshold</TT
> (<TT
CLASS="TYPE"
>integer</TT
>)
</DT
><DD
><P
> Sets the minimum time that a snapshot can be used without risk of a
<TT
CLASS="LITERAL"
>snapshot too old</TT
> error occurring when using the snapshot.
This parameter can only be set at server start.
</P
><P
> Beyond the threshold, old data may be vacuumed away. This can help
prevent bloat in the face of snapshots which remain in use for a
long time. To prevent incorrect results due to cleanup of data which
would otherwise be visible to the snapshot, an error is generated
when the snapshot is older than this threshold and the snapshot is
used to read a page which has been modified since the snapshot was
built.
</P
><P
> A value of <TT
CLASS="LITERAL"
>-1</TT
> disables this feature, and is the default.
Useful values for production work probably range from a small number
of hours to a few days. The setting will be coerced to a granularity
of minutes, and small numbers (such as <TT
CLASS="LITERAL"
>0</TT
> or
<TT
CLASS="LITERAL"
>1min</TT
>) are only allowed because they may sometimes be
useful for testing. While a setting as high as <TT
CLASS="LITERAL"
>60d</TT
> is
allowed, please note that in many workloads extreme bloat or
transaction ID wraparound may occur in much shorter time frames.
</P
><P
> When this feature is enabled, freed space at the end of a relation
cannot be released to the operating system, since that could remove
information needed to detect the <TT
CLASS="LITERAL"
>snapshot too old</TT
>
condition. All space allocated to a relation remains associated with
that relation for reuse only within that relation unless explicitly
freed (for example, with <TT
CLASS="COMMAND"
>VACUUM FULL</TT
>).
</P
><P
> This setting does not attempt to guarantee that an error will be
generated under any particular circumstances. In fact, if the
correct results can be generated from (for example) a cursor which
has materialized a result set, no error will be generated even if the
underlying rows in the referenced table have been vacuumed away.
Some tables cannot safely be vacuumed early, and so will not be
affected by this setting. Examples include system catalogs and any
table which has a hash index. For such tables this setting will
neither reduce bloat nor create a possibility of a <TT
CLASS="LITERAL"
>snapshot
too old</TT
> error on scanning.
</P
></DD
></DL
></DIV
></DIV
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="runtime-config-connection.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="runtime-config-wal.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Connections and Authentication</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="runtime-config.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>Write Ahead Log</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
distrib
>
Mageia
>
7
>
armv7hl
>
media
>
core-updates
>
by-pkgid
>
8898dd367b2cdca21ea56f51161fb737
>
files
>
667
