1PG_BASEBACKUP(1) PostgreSQL 12.6 Documentation PG_BASEBACKUP(1)
2
6 pg_basebackup - take a base backup of a PostgreSQL cluster
7
9 pg_basebackup [option...]
10
12 pg_basebackup is used to take base backups of a running PostgreSQL
13 database cluster. These are taken without affecting other clients to
14 the database, and can be used both for point-in-time recovery (see
15 Section 25.3) and as the starting point for a log shipping or streaming
16 replication standby servers (see Section 26.2).
17 pg_basebackup makes a binary copy of the database cluster files, while
19 making sure the system is put in and out of backup mode automatically.
20 Backups are always taken of the entire database cluster; it is not
21 possible to back up individual databases or database objects. For
22 individual database backups, a tool such as pg_dump(1) must be used.
23 The backup is made over a regular PostgreSQL connection, and uses the
25 replication protocol. The connection must be made with a superuser or a
26 user having REPLICATION permissions (see Section 21.2), and pg_hba.conf
27 must explicitly permit the replication connection. The server must also
28 be configured with max_wal_senders set high enough to leave at least
29 one session available for the backup and one for WAL streaming (if
30 used).
31 There can be multiple pg_basebackups running at the same time, but it
33 is better from a performance point of view to take only one backup, and
34 copy the result.
35 pg_basebackup can make a base backup from not only the master but also
37 the standby. To take a backup from the standby, set up the standby so
38 that it can accept replication connections (that is, set
39 max_wal_senders and hot_standby, and configure host-based
40 authentication). You will also need to enable full_page_writes on the
41 master.
42 Note that there are some limitations in an online backup from the
44 standby:
45 · The backup history file is not created in the database cluster
47 backed up.
48 · If you are using -X none, there is no guarantee that all WAL files
50 required for the backup are archived at the end of backup.
51 · If the standby is promoted to the master during online backup, the
53 backup fails.
54 · All WAL records required for the backup must contain sufficient
56 full-page writes, which requires you to enable full_page_writes on
57 the master and not to use a tool like pg_compresslog as
58 archive_command to remove full-page writes from WAL files.
59
62 The following command-line options control the location and format of
63 the output.
64 -D directory
66 --pgdata=directory
67 Directory to write the output to. pg_basebackup will create the
68 directory and any parent directories if necessary. The directory
69 may already exist, but it is an error if the directory already
70 exists and is not empty.
71 When the backup is in tar mode, and the directory is specified as -
73 (dash), the tar file will be written to stdout.
74 This option is required.
76 -F format
78 --format=format
79 Selects the format for the output. format can be one of the
80 following:
81 p
83 plain
84 Write the output as plain files, with the same layout as the
85 current data directory and tablespaces. When the cluster has no
86 additional tablespaces, the whole database will be placed in
87 the target directory. If the cluster contains additional
88 tablespaces, the main data directory will be placed in the
89 target directory, but all other tablespaces will be placed in
90 the same absolute path as they have on the server.
91 This is the default format.
93 t
95 tar
96 Write the output as tar files in the target directory. The main
97 data directory will be written to a file named base.tar, and
98 all other tablespaces will be named after the tablespace OID.
99 If the value - (dash) is specified as target directory, the tar
101 contents will be written to standard output, suitable for
102 piping to for example gzip. This is only possible if the
103 cluster has no additional tablespaces and WAL streaming is not
104 used.
105 -r rate
107 --max-rate=rate
108 The maximum transfer rate of data transferred from the server.
109 Values are in kilobytes per second. Use a suffix of M to indicate
110 megabytes per second. A suffix of k is also accepted, and has no
111 effect. Valid values are between 32 kilobytes per second and 1024
112 megabytes per second.
113 The purpose is to limit the impact of pg_basebackup on the running
115 server.
116 This option always affects transfer of the data directory. Transfer
118 of WAL files is only affected if the collection method is fetch.
119 -R
121 --write-recovery-conf
122 Create standby.signal and append connection settings to
123 postgresql.auto.conf in the output directory (or into the base
124 archive file when using tar format) to ease setting up a standby
125 server. The postgresql.auto.conf file will record the connection
126 settings and, if specified, the replication slot that pg_basebackup
127 is using, so that the streaming replication will use the same
128 settings later on.
129 -T olddir=newdir
131 --tablespace-mapping=olddir=newdir
132 Relocate the tablespace in directory olddir to newdir during the
133 backup. To be effective, olddir must exactly match the path
134 specification of the tablespace as it is currently defined. (But it
135 is not an error if there is no tablespace in olddir contained in
136 the backup.) Both olddir and newdir must be absolute paths. If a
137 path happens to contain a = sign, escape it with a backslash. This
138 option can be specified multiple times for multiple tablespaces.
139 See examples below.
140 If a tablespace is relocated in this way, the symbolic links inside
142 the main data directory are updated to point to the new location.
143 So the new data directory is ready to be used for a new server
144 instance with all tablespaces in the updated locations.
145 --waldir=waldir
147 Specifies the location for the write-ahead log directory. waldir
148 must be an absolute path. The write-ahead log directory can only be
149 specified when the backup is in plain mode.
150 -X method
152 --wal-method=method
153 Includes the required write-ahead log files (WAL files) in the
154 backup. This will include all write-ahead logs generated during the
155 backup. Unless the method none is specified, it is possible to
156 start a postmaster directly in the extracted directory without the
157 need to consult the log archive, thus making this a completely
158 standalone backup.
159 The following methods for collecting the write-ahead logs are
161 supported:
162 n
164 none
165 Don't include write-ahead log in the backup.
166 f
168 fetch
169 The write-ahead log files are collected at the end of the
170 backup. Therefore, it is necessary for the wal_keep_segments
171 parameter to be set high enough that the log is not removed
172 before the end of the backup. If the log has been rotated when
173 it's time to transfer it, the backup will fail and be unusable.
174 When tar format mode is used, the write-ahead log files will be
176 written to the base.tar file.
177 s
179 stream
180 Stream the write-ahead log while the backup is created. This
181 will open a second connection to the server and start streaming
182 the write-ahead log in parallel while running the backup.
183 Therefore, it will use up two connections configured by the
184 max_wal_senders parameter. As long as the client can keep up
185 with write-ahead log received, using this mode requires no
186 extra write-ahead logs to be saved on the master.
187 When tar format mode is used, the write-ahead log files will be
189 written to a separate file named pg_wal.tar (if the server is a
190 version earlier than 10, the file will be named pg_xlog.tar).
191 This value is the default.
193 -z
196 --gzip
197 Enables gzip compression of tar file output, with the default
198 compression level. Compression is only available when using the tar
199 format, and the suffix .gz will automatically be added to all tar
200 filenames.
201 -Z level
203 --compress=level
204 Enables gzip compression of tar file output, and specifies the
205 compression level (0 through 9, 0 being no compression and 9 being
206 best compression). Compression is only available when using the tar
207 format, and the suffix .gz will automatically be added to all tar
208 filenames.
209 The following command-line options control the generation of the backup
211 and the running of the program.
212 -c fast|spread
214 --checkpoint=fast|spread
215 Sets checkpoint mode to fast (immediate) or spread (default) (see
216 Section 25.3.3).
217 -C
219 --create-slot
220 This option causes creation of a replication slot named by the
221 --slot option before starting the backup. An error is raised if the
222 slot already exists.
223 -l label
225 --label=label
226 Sets the label for the backup. If none is specified, a default
227 value of “pg_basebackup base backup” will be used.
228 -n
230 --no-clean
231 By default, when pg_basebackup aborts with an error, it removes any
232 directories it might have created before discovering that it cannot
233 finish the job (for example, data directory and write-ahead log
234 directory). This option inhibits tidying-up and is thus useful for
235 debugging.
236 Note that tablespace directories are not cleaned up either way.
238 -N
240 --no-sync
241 By default, pg_basebackup will wait for all files to be written
242 safely to disk. This option causes pg_basebackup to return without
243 waiting, which is faster, but means that a subsequent operating
244 system crash can leave the base backup corrupt. Generally, this
245 option is useful for testing but should not be used when creating a
246 production installation.
247 -P
249 --progress
250 Enables progress reporting. Turning this on will deliver an
251 approximate progress report during the backup. Since the database
252 may change during the backup, this is only an approximation and may
253 not end at exactly 100%. In particular, when WAL log is included in
254 the backup, the total amount of data cannot be estimated in
255 advance, and in this case the estimated target size will increase
256 once it passes the total estimate without WAL.
257 When this is enabled, the backup will start by enumerating the size
259 of the entire database, and then go back and send the actual
260 contents. This may make the backup take slightly longer, and in
261 particular it will take longer before the first data is sent.
262 -S slotname
264 --slot=slotname
265 This option can only be used together with -X stream. It causes the
266 WAL streaming to use the specified replication slot. If the base
267 backup is intended to be used as a streaming replication standby
268 using replication slots, it should then use the same replication
269 slot name in primary_slot_name. That way, it is ensured that the
270 server does not remove any necessary WAL data in the time between
271 the end of the base backup and the start of streaming replication.
272 The specified replication slot has to exist unless the option -C is
274 also used.
275 If this option is not specified and the server supports temporary
277 replication slots (version 10 and later), then a temporary
278 replication slot is automatically used for WAL streaming.
279 -v
281 --verbose
282 Enables verbose mode. Will output some extra steps during startup
283 and shutdown, as well as show the exact file name that is currently
284 being processed if progress reporting is also enabled.
285 --no-slot
287 This option prevents the creation of a temporary replication slot
288 during the backup even if it's supported by the server.
289 Temporary replication slots are created by default if no slot name
291 is given with the option -S when using log streaming.
292 The main purpose of this option is to allow taking a base backup
294 when the server is out of free replication slots. Using replication
295 slots is almost always preferred, because it prevents needed WAL
296 from being removed by the server during the backup.
297 --no-verify-checksums
299 Disables verification of checksums, if they are enabled on the
300 server the base backup is taken from.
301 By default, checksums are verified and checksum failures will
303 result in a non-zero exit status. However, the base backup will not
304 be removed in such a case, as if the --no-clean option had been
305 used. Checksum verifications failures will also be reported in the
306 pg_stat_database view.
307 The following command-line options control the database connection
309 parameters.
310 -d connstr
312 --dbname=connstr
313 Specifies parameters used to connect to the server, as a connection
314 string; these will override any conflicting command line options.
315 The option is called --dbname for consistency with other client
317 applications, but because pg_basebackup doesn't connect to any
318 particular database in the cluster, database name in the connection
319 string will be ignored.
320 -h host
322 --host=host
323 Specifies the host name of the machine on which the server is
324 running. If the value begins with a slash, it is used as the
325 directory for the Unix domain socket. The default is taken from the
326 PGHOST environment variable, if set, else a Unix domain socket
327 connection is attempted.
328 -p port
330 --port=port
331 Specifies the TCP port or local Unix domain socket file extension
332 on which the server is listening for connections. Defaults to the
333 PGPORT environment variable, if set, or a compiled-in default.
334 -s interval
336 --status-interval=interval
337 Specifies the number of seconds between status packets sent back to
338 the server. This allows for easier monitoring of the progress from
339 server. A value of zero disables the periodic status updates
340 completely, although an update will still be sent when requested by
341 the server, to avoid timeout disconnect. The default value is 10
342 seconds.
343 -U username
345 --username=username
346 User name to connect as.
347 -w
349 --no-password
350 Never issue a password prompt. If the server requires password
351 authentication and a password is not available by other means such
352 as a .pgpass file, the connection attempt will fail. This option
353 can be useful in batch jobs and scripts where no user is present to
354 enter a password.
355 -W
357 --password
358 Force pg_basebackup to prompt for a password before connecting to a
359 database.
360 This option is never essential, since pg_basebackup will
362 automatically prompt for a password if the server demands password
363 authentication. However, pg_basebackup will waste a connection
364 attempt finding out that the server wants a password. In some cases
365 it is worth typing -W to avoid the extra connection attempt.
366 Other options are also available:
368 -V
370 --version
371 Print the pg_basebackup version and exit.
372 -?
374 --help
375 Show help about pg_basebackup command line arguments, and exit.
376
378 This utility, like most other PostgreSQL utilities, uses the
379 environment variables supported by libpq (see Section 33.14).
380 The environment variable PG_COLOR specifies whether to use color in
382 diagnostic messages. Possible values are always, auto and never.
383
385 At the beginning of the backup, a checkpoint needs to be written on the
386 server the backup is taken from. Especially if the option
387 --checkpoint=fast is not used, this can take some time during which
388 pg_basebackup will be appear to be idle.
389 The backup will include all files in the data directory and
391 tablespaces, including the configuration files and any additional files
392 placed in the directory by third parties, except certain temporary
393 files managed by PostgreSQL. But only regular files and directories are
394 copied, except that symbolic links used for tablespaces are preserved.
395 Symbolic links pointing to certain directories known to PostgreSQL are
396 copied as empty directories. Other symbolic links and special device
397 files are skipped. See Section 52.4 for the precise details.
398 Tablespaces will in plain format by default be backed up to the same
400 path they have on the server, unless the option --tablespace-mapping is
401 used. Without this option, running a plain format base backup on the
402 same host as the server will not work if tablespaces are in use,
403 because the backup would have to be written to the same directory
404 locations as the original tablespaces.
405 When tar format mode is used, it is the user's responsibility to unpack
407 each tar file before starting the PostgreSQL server. If there are
408 additional tablespaces, the tar files for them need to be unpacked in
409 the correct locations. In this case the symbolic links for those
410 tablespaces will be created by the server according to the contents of
411 the tablespace_map file that is included in the base.tar file.
412 pg_basebackup works with servers of the same or an older major version,
414 down to 9.1. However, WAL streaming mode (-X stream) only works with
415 server version 9.3 and later, and tar format mode (--format=tar) of the
416 current version only works with server version 9.5 or later.
417 pg_basebackup will preserve group permissions in both the plain and tar
419 formats if group permissions are enabled on the source cluster.
420
422 To create a base backup of the server at mydbserver and store it in the
423 local directory /usr/local/pgsql/data:
424 $ pg_basebackup -h mydbserver -D /usr/local/pgsql/data
426 To create a backup of the local server with one compressed tar file for
428 each tablespace, and store it in the directory backup, showing a
429 progress report while running:
430 $ pg_basebackup -D backup -Ft -z -P
432 To create a backup of a single-tablespace local database and compress
434 this with bzip2:
435 $ pg_basebackup -D - -Ft -X fetch | bzip2 > backup.tar.bz2
437 (This command will fail if there are multiple tablespaces in the
439 database.)
440 To create a backup of a local database where the tablespace in /opt/ts
442 is relocated to ./backup/ts:
443 $ pg_basebackup -D backup/data -T /opt/ts=$(pwd)/backup/ts
445
448 pg_dump(1)
449PostgreSQL 12.6 2021 PG_BASEBACKUP(1)