summaryrefslogtreecommitdiffstats
path: root/doc/src/sgml/html/bgworker.html
blob: e1970a0de8d6dab2e0fb3321280120f685359c48 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
<?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>Chapter 48. Background Worker Processes</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="spi-spi-start-transaction.html" title="SPI_start_transaction" /><link rel="next" href="logicaldecoding.html" title="Chapter 49. Logical Decoding" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">Chapter 48. Background Worker Processes</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="spi-spi-start-transaction.html" title="SPI_start_transaction">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="server-programming.html" title="Part V. Server Programming">Up</a></td><th width="60%" align="center">Part V. Server Programming</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="logicaldecoding.html" title="Chapter 49. Logical Decoding">Next</a></td></tr></table><hr /></div><div class="chapter" id="BGWORKER"><div class="titlepage"><div><div><h2 class="title">Chapter 48. Background Worker Processes</h2></div></div></div><a id="id-1.8.13.2" class="indexterm"></a><p>
  PostgreSQL can be extended to run user-supplied code in separate processes.
  Such processes are started, stopped and monitored by <code class="command">postgres</code>,
  which permits them to have a lifetime closely linked to the server's status.
  These processes are attached to <span class="productname">PostgreSQL</span>'s
  shared memory area and have the option to connect to databases internally; they can also run
  multiple transactions serially, just like a regular client-connected server
  process.  Also, by linking to <span class="application">libpq</span> they can connect to the
  server and behave like a regular client application.
 </p><div class="warning"><h3 class="title">Warning</h3><p>
   There are considerable robustness and security risks in using background
   worker processes because, being written in the <code class="literal">C</code> language,
   they have unrestricted access to data.  Administrators wishing to enable
   modules that include background worker processes should exercise extreme
   caution.  Only carefully audited modules should be permitted to run
   background worker processes.
  </p></div><p>
  Background workers can be initialized at the time that
  <span class="productname">PostgreSQL</span> is started by including the module name in
  <code class="varname">shared_preload_libraries</code>.  A module wishing to run a background
  worker can register it by calling
  <code class="function">RegisterBackgroundWorker(<code class="type">BackgroundWorker</code>
  *<em class="parameter"><code>worker</code></em>)</code>
  from its <code class="function">_PG_init()</code> function.
  Background workers can also be started
  after the system is up and running by calling
  <code class="function">RegisterDynamicBackgroundWorker(<code class="type">BackgroundWorker</code>
  *<em class="parameter"><code>worker</code></em>, <code class="type">BackgroundWorkerHandle</code>
  **<em class="parameter"><code>handle</code></em>)</code>.  Unlike
  <code class="function">RegisterBackgroundWorker</code>, which can only be called from
  within the postmaster process,
  <code class="function">RegisterDynamicBackgroundWorker</code> must be called
  from a regular backend or another background worker.
 </p><p>
  The structure <code class="structname">BackgroundWorker</code> is defined thus:
</p><pre class="programlisting">
typedef void (*bgworker_main_type)(Datum main_arg);
typedef struct BackgroundWorker
{
    char        bgw_name[BGW_MAXLEN];
    char        bgw_type[BGW_MAXLEN];
    int         bgw_flags;
    BgWorkerStartTime bgw_start_time;
    int         bgw_restart_time;       /* in seconds, or BGW_NEVER_RESTART */
    char        bgw_library_name[BGW_MAXLEN];
    char        bgw_function_name[BGW_MAXLEN];
    Datum       bgw_main_arg;
    char        bgw_extra[BGW_EXTRALEN];
    int         bgw_notify_pid;
} BackgroundWorker;
</pre><p>
  </p><p>
   <code class="structfield">bgw_name</code> and <code class="structfield">bgw_type</code> are
   strings to be used in log messages, process listings and similar contexts.
   <code class="structfield">bgw_type</code> should be the same for all background
   workers of the same type, so that it is possible to group such workers in a
   process listing, for example.  <code class="structfield">bgw_name</code> on the
   other hand can contain additional information about the specific process.
   (Typically, the string for <code class="structfield">bgw_name</code> will contain
   the type somehow, but that is not strictly required.)
  </p><p>
   <code class="structfield">bgw_flags</code> is a bitwise-or'd bit mask indicating the
   capabilities that the module wants.  Possible values are:
   </p><div class="variablelist"><dl class="variablelist"><dt><span class="term"><code class="literal">BGWORKER_SHMEM_ACCESS</code></span></dt><dd><p>
       <a id="id-1.8.13.8.2.1.2.1.1" class="indexterm"></a>
       Requests shared memory access.  This flag is required.
      </p></dd><dt><span class="term"><code class="literal">BGWORKER_BACKEND_DATABASE_CONNECTION</code></span></dt><dd><p>
       <a id="id-1.8.13.8.2.2.2.1.1" class="indexterm"></a>
       Requests the ability to establish a database connection through which it
       can later run transactions and queries.  A background worker using
       <code class="literal">BGWORKER_BACKEND_DATABASE_CONNECTION</code> to connect to a
       database must also attach shared memory using
       <code class="literal">BGWORKER_SHMEM_ACCESS</code>, or worker start-up will fail.
      </p></dd></dl></div><p>

  </p><p>
   <code class="structfield">bgw_start_time</code> is the server state during which
   <code class="command">postgres</code> should start the process; it can be one of
   <code class="literal">BgWorkerStart_PostmasterStart</code> (start as soon as
   <code class="command">postgres</code> itself has finished its own initialization; processes
   requesting this are not eligible for database connections),
   <code class="literal">BgWorkerStart_ConsistentState</code> (start as soon as a consistent state
   has been reached in a hot standby, allowing processes to connect to
   databases and run read-only queries), and
   <code class="literal">BgWorkerStart_RecoveryFinished</code> (start as soon as the system has
   entered normal read-write state).  Note the last two values are equivalent
   in a server that's not a hot standby.  Note that this setting only indicates
   when the processes are to be started; they do not stop when a different state
   is reached.
  </p><p>
   <code class="structfield">bgw_restart_time</code> is the interval, in seconds, that
   <code class="command">postgres</code> should wait before restarting the process in
   the event that it crashes.  It can be any positive value,
   or <code class="literal">BGW_NEVER_RESTART</code>, indicating not to restart the
   process in case of a crash.
  </p><p>
   <code class="structfield">bgw_library_name</code> is the name of a library in
   which the initial entry point for the background worker should be sought.
   The named library will be dynamically loaded by the worker process and
   <code class="structfield">bgw_function_name</code> will be used to identify the
   function to be called.  If loading a function from the core code, this must
   be set to "postgres".
  </p><p>
   <code class="structfield">bgw_function_name</code> is the name of a function in
   a dynamically loaded library which should be used as the initial entry point
   for a new background worker.
  </p><p>
   <code class="structfield">bgw_main_arg</code> is the <code class="type">Datum</code> argument
   to the background worker main function.  This main function should take a
   single argument of type <code class="type">Datum</code> and return <code class="type">void</code>.
   <code class="structfield">bgw_main_arg</code> will be passed as the argument.
   In addition, the global variable <code class="literal">MyBgworkerEntry</code>
   points to a copy of the <code class="structname">BackgroundWorker</code> structure
   passed at registration time; the worker may find it helpful to examine
   this structure.
  </p><p>
   On Windows (and anywhere else where <code class="literal">EXEC_BACKEND</code> is
   defined) or in dynamic background workers it is not safe to pass a
   <code class="type">Datum</code> by reference, only by value. If an argument is required, it
   is safest to pass an int32 or other small value and use that as an index
   into an array allocated in shared memory. If a value like a <code class="type">cstring</code>
   or <code class="type">text</code> is passed then the pointer won't be valid from the
   new background worker process.
  </p><p>
   <code class="structfield">bgw_extra</code> can contain extra data to be passed
   to the background worker.  Unlike <code class="structfield">bgw_main_arg</code>, this data
   is not passed as an argument to the worker's main function, but it can be
   accessed via <code class="literal">MyBgworkerEntry</code>, as discussed above.
  </p><p>
   <code class="structfield">bgw_notify_pid</code> is the PID of a PostgreSQL
   backend process to which the postmaster should send <code class="literal">SIGUSR1</code>
   when the process is started or exits.  It should be 0 for workers registered
   at postmaster startup time, or when the backend registering the worker does
   not wish to wait for the worker to start up.  Otherwise, it should be
   initialized to <code class="literal">MyProcPid</code>.
  </p><p>Once running, the process can connect to a database by calling
   <code class="function">BackgroundWorkerInitializeConnection(<em class="parameter"><code>char *dbname</code></em>, <em class="parameter"><code>char *username</code></em>, <em class="parameter"><code>uint32 flags</code></em>)</code> or
   <code class="function">BackgroundWorkerInitializeConnectionByOid(<em class="parameter"><code>Oid dboid</code></em>, <em class="parameter"><code>Oid useroid</code></em>, <em class="parameter"><code>uint32 flags</code></em>)</code>.
   This allows the process to run transactions and queries using the
   <code class="literal">SPI</code> interface.  If <code class="varname">dbname</code> is NULL or
   <code class="varname">dboid</code> is <code class="literal">InvalidOid</code>, the session is not connected
   to any particular database, but shared catalogs can be accessed.
   If <code class="varname">username</code> is NULL or <code class="varname">useroid</code> is
   <code class="literal">InvalidOid</code>, the process will run as the superuser created
   during <code class="command">initdb</code>. If <code class="literal">BGWORKER_BYPASS_ALLOWCONN</code>
   is specified as <code class="varname">flags</code> it is possible to bypass the restriction
   to connect to databases not allowing user connections.
   A background worker can only call one of these two functions, and only
   once.  It is not possible to switch databases.
  </p><p>
   Signals are initially blocked when control reaches the
   background worker's main function, and must be unblocked by it; this is to
   allow the process to customize its signal handlers, if necessary.
   Signals can be unblocked in the new process by calling
   <code class="function">BackgroundWorkerUnblockSignals</code> and blocked by calling
   <code class="function">BackgroundWorkerBlockSignals</code>.
  </p><p>
   If <code class="structfield">bgw_restart_time</code> for a background worker is
   configured as <code class="literal">BGW_NEVER_RESTART</code>, or if it exits with an exit
   code of 0 or is terminated by <code class="function">TerminateBackgroundWorker</code>,
   it will be automatically unregistered by the postmaster on exit.
   Otherwise, it will be restarted after the time period configured via
   <code class="structfield">bgw_restart_time</code>, or immediately if the postmaster
   reinitializes the cluster due to a backend failure.  Backends which need
   to suspend execution only temporarily should use an interruptible sleep
   rather than exiting; this can be achieved by calling
   <code class="function">WaitLatch()</code>. Make sure the
   <code class="literal">WL_POSTMASTER_DEATH</code> flag is set when calling that function, and
   verify the return code for a prompt exit in the emergency case that
   <code class="command">postgres</code> itself has terminated.
  </p><p>
   When a background worker is registered using the
   <code class="function">RegisterDynamicBackgroundWorker</code> function, it is
   possible for the backend performing the registration to obtain information
   regarding the status of the worker.  Backends wishing to do this should
   pass the address of a <code class="type">BackgroundWorkerHandle *</code> as the second
   argument to <code class="function">RegisterDynamicBackgroundWorker</code>.  If the
   worker is successfully registered, this pointer will be initialized with an
   opaque handle that can subsequently be passed to
   <code class="function">GetBackgroundWorkerPid(<em class="parameter"><code>BackgroundWorkerHandle *</code></em>, <em class="parameter"><code>pid_t *</code></em>)</code> or
   <code class="function">TerminateBackgroundWorker(<em class="parameter"><code>BackgroundWorkerHandle *</code></em>)</code>.
   <code class="function">GetBackgroundWorkerPid</code> can be used to poll the status of the
   worker: a return value of <code class="literal">BGWH_NOT_YET_STARTED</code> indicates that
   the worker has not yet been started by the postmaster;
   <code class="literal">BGWH_STOPPED</code> indicates that it has been started but is
   no longer running; and <code class="literal">BGWH_STARTED</code> indicates that it is
   currently running.  In this last case, the PID will also be returned via the
   second argument.
   <code class="function">TerminateBackgroundWorker</code> causes the postmaster to send
   <code class="literal">SIGTERM</code> to the worker if it is running, and to unregister it
   as soon as it is not.
  </p><p>
   In some cases, a process which registers a background worker may wish to
   wait for the worker to start up.  This can be accomplished by initializing
   <code class="structfield">bgw_notify_pid</code> to <code class="literal">MyProcPid</code> and
   then passing the <code class="type">BackgroundWorkerHandle *</code> obtained at
   registration time to
   <code class="function">WaitForBackgroundWorkerStartup(<em class="parameter"><code>BackgroundWorkerHandle
   *handle</code></em>, <em class="parameter"><code>pid_t *</code></em>)</code> function.
   This function will block until the postmaster has attempted to start the
   background worker, or until the postmaster dies.  If the background worker
   is running, the return value will be <code class="literal">BGWH_STARTED</code>, and
   the PID will be written to the provided address.  Otherwise, the return
   value will be <code class="literal">BGWH_STOPPED</code> or
   <code class="literal">BGWH_POSTMASTER_DIED</code>.
  </p><p>
   A process can also wait for a background worker to shut down, by using the
   <code class="function">WaitForBackgroundWorkerShutdown(<em class="parameter"><code>BackgroundWorkerHandle
   *handle</code></em>)</code> function and passing the
   <code class="type">BackgroundWorkerHandle *</code> obtained at registration. This
   function will block until the background worker exits, or postmaster dies.
   When the background worker exits, the return value is
   <code class="literal">BGWH_STOPPED</code>, if postmaster dies it will return
   <code class="literal">BGWH_POSTMASTER_DIED</code>.
  </p><p>
   Background workers can send asynchronous notification messages, either by
   using the <code class="command">NOTIFY</code> command via <acronym class="acronym">SPI</acronym>,
   or directly via <code class="function">Async_Notify()</code>.  Such notifications
   will be sent at transaction commit.
   Background workers should not register to receive asynchronous
   notifications with the <code class="command">LISTEN</code> command, as there is no
   infrastructure for a worker to consume such notifications.
  </p><p>
   The <code class="filename">src/test/modules/worker_spi</code> module
   contains a working example,
   which demonstrates some useful techniques.
  </p><p>
   The maximum number of registered background workers is limited by
   <a class="xref" href="runtime-config-resource.html#GUC-MAX-WORKER-PROCESSES">max_worker_processes</a>.
  </p></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="spi-spi-start-transaction.html" title="SPI_start_transaction">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="server-programming.html" title="Part V. Server Programming">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="logicaldecoding.html" title="Chapter 49. Logical Decoding">Next</a></td></tr><tr><td width="40%" align="left" valign="top">SPI_start_transaction </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="40%" align="right" valign="top"> Chapter 49. Logical Decoding</td></tr></table></div></body></html>