diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-05-04 12:15:05 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-05-04 12:15:05 +0000 |
commit | 46651ce6fe013220ed397add242004d764fc0153 (patch) | |
tree | 6e5299f990f88e60174a1d3ae6e48eedd2688b2b /doc/src/sgml/html/parallel-safety.html | |
parent | Initial commit. (diff) | |
download | postgresql-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.html | 83 |
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 |