summaryrefslogtreecommitdiffstats
path: root/doc/src/sgml/html/parallel-safety.html
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2024-05-04 12:15:05 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2024-05-04 12:15:05 +0000
commit46651ce6fe013220ed397add242004d764fc0153 (patch)
tree6e5299f990f88e60174a1d3ae6e48eedd2688b2b /doc/src/sgml/html/parallel-safety.html
parentInitial commit. (diff)
downloadpostgresql-14-upstream.tar.xz
postgresql-14-upstream.zip
Adding upstream version 14.5.upstream/14.5upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to '')
-rw-r--r--doc/src/sgml/html/parallel-safety.html83
1 files changed, 83 insertions, 0 deletions
diff --git a/doc/src/sgml/html/parallel-safety.html b/doc/src/sgml/html/parallel-safety.html
new file mode 100644
index 0000000..d1efc76
--- /dev/null
+++ b/doc/src/sgml/html/parallel-safety.html
@@ -0,0 +1,83 @@
+<?xml version="1.0" encoding="UTF-8" standalone="no"?>
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>15.4. Parallel Safety</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="parallel-plans.html" title="15.3. Parallel Plans" /><link rel="next" href="admin.html" title="Part III. Server Administration" /></head><body id="docContent" class="container-fluid col-10"><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">15.4. Parallel Safety</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="parallel-plans.html" title="15.3. Parallel Plans">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="parallel-query.html" title="Chapter 15. Parallel Query">Up</a></td><th width="60%" align="center">Chapter 15. Parallel Query</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 14.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="admin.html" title="Part III. Server Administration">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="PARALLEL-SAFETY"><div class="titlepage"><div><div><h2 class="title" style="clear: both">15.4. Parallel Safety</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="parallel-safety.html#PARALLEL-LABELING">15.4.1. Parallel Labeling for Functions and Aggregates</a></span></dt></dl></div><p>
+ The planner classifies operations involved in a query as either
+ <em class="firstterm">parallel safe</em>, <em class="firstterm">parallel restricted</em>,
+ or <em class="firstterm">parallel unsafe</em>. A parallel safe operation is one that
+ does not conflict with the use of parallel query. A parallel restricted
+ operation is one that cannot be performed in a parallel worker, but that
+ can be performed in the leader while parallel query is in use. Therefore,
+ parallel restricted operations can never occur below a <code class="literal">Gather</code>
+ or <code class="literal">Gather Merge</code> node, but can occur elsewhere in a plan that
+ contains such a node. A parallel unsafe operation is one that cannot
+ be performed while parallel query is in use, not even in the leader.
+ When a query contains anything that is parallel unsafe, parallel query
+ is completely disabled for that query.
+ </p><p>
+ The following operations are always parallel restricted:
+ </p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
+ Scans of common table expressions (CTEs).
+ </p></li><li class="listitem"><p>
+ Scans of temporary tables.
+ </p></li><li class="listitem"><p>
+ Scans of foreign tables, unless the foreign data wrapper has
+ an <code class="literal">IsForeignScanParallelSafe</code> API that indicates otherwise.
+ </p></li><li class="listitem"><p>
+ Plan nodes to which an <code class="literal">InitPlan</code> is attached.
+ </p></li><li class="listitem"><p>
+ Plan nodes that reference a correlated <code class="literal">SubPlan</code>.
+ </p></li></ul></div><div class="sect2" id="PARALLEL-LABELING"><div class="titlepage"><div><div><h3 class="title">15.4.1. Parallel Labeling for Functions and Aggregates</h3></div></div></div><p>
+ The planner cannot automatically determine whether a user-defined
+ function or aggregate is parallel safe, parallel restricted, or parallel
+ unsafe, because this would require predicting every operation that the
+ function could possibly perform. In general, this is equivalent to the
+ Halting Problem and therefore impossible. Even for simple functions
+ where it could conceivably be done, we do not try, since this would be expensive
+ and error-prone. Instead, all user-defined functions are assumed to
+ be parallel unsafe unless otherwise marked. When using
+ <a class="xref" href="sql-createfunction.html" title="CREATE FUNCTION"><span class="refentrytitle">CREATE FUNCTION</span></a> or
+ <a class="xref" href="sql-alterfunction.html" title="ALTER FUNCTION"><span class="refentrytitle">ALTER FUNCTION</span></a>, markings can be set by specifying
+ <code class="literal">PARALLEL SAFE</code>, <code class="literal">PARALLEL RESTRICTED</code>, or
+ <code class="literal">PARALLEL UNSAFE</code> as appropriate. When using
+ <a class="xref" href="sql-createaggregate.html" title="CREATE AGGREGATE"><span class="refentrytitle">CREATE AGGREGATE</span></a>, the
+ <code class="literal">PARALLEL</code> option can be specified with <code class="literal">SAFE</code>,
+ <code class="literal">RESTRICTED</code>, or <code class="literal">UNSAFE</code> as the corresponding value.
+ </p><p>
+ Functions and aggregates must be marked <code class="literal">PARALLEL UNSAFE</code> if
+ they write to the database, access sequences, change the transaction state
+ even temporarily (e.g., a PL/pgSQL function that establishes an
+ <code class="literal">EXCEPTION</code> block to catch errors), or make persistent changes to
+ settings. Similarly, functions must be marked <code class="literal">PARALLEL
+ RESTRICTED</code> if they access temporary tables, client connection state,
+ cursors, prepared statements, or miscellaneous backend-local state that
+ the system cannot synchronize across workers. For example,
+ <code class="literal">setseed</code> and <code class="literal">random</code> are parallel restricted for
+ this last reason.
+ </p><p>
+ In general, if a function is labeled as being safe when it is restricted or
+ unsafe, or if it is labeled as being restricted when it is in fact unsafe,
+ it may throw errors or produce wrong answers when used in a parallel query.
+ C-language functions could in theory exhibit totally undefined behavior if
+ mislabeled, since there is no way for the system to protect itself against
+ arbitrary C code, but in most likely cases the result will be no worse than
+ for any other function. If in doubt, it is probably best to label functions
+ as <code class="literal">UNSAFE</code>.
+ </p><p>
+ If a function executed within a parallel worker acquires locks that are
+ not held by the leader, for example by querying a table not referenced in
+ the query, those locks will be released at worker exit, not end of
+ transaction. If you write a function that does this, and this behavior
+ difference is important to you, mark such functions as
+ <code class="literal">PARALLEL RESTRICTED</code>
+ to ensure that they execute only in the leader.
+ </p><p>
+ Note that the query planner does not consider deferring the evaluation of
+ parallel-restricted functions or aggregates involved in the query in
+ order to obtain a superior plan. So, for example, if a <code class="literal">WHERE</code>
+ clause applied to a particular table is parallel restricted, the query
+ planner will not consider performing a scan of that table in the parallel
+ portion of a plan. In some cases, it would be
+ possible (and perhaps even efficient) to include the scan of that table in
+ the parallel portion of the query and defer the evaluation of the
+ <code class="literal">WHERE</code> clause so that it happens above the <code class="literal">Gather</code>
+ node. However, the planner does not do this.
+ </p></div></div><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navfooter"><hr></hr><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="parallel-plans.html" title="15.3. Parallel Plans">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="parallel-query.html" title="Chapter 15. Parallel Query">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="admin.html" title="Part III. Server Administration">Next</a></td></tr><tr><td width="40%" align="left" valign="top">15.3. Parallel Plans </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 14.5 Documentation">Home</a></td><td width="40%" align="right" valign="top"> Part III. Server Administration</td></tr></table></div></body></html> \ No newline at end of file