1PG_UPGRADE(1) PostgreSQL 9.2.24 Documentation PG_UPGRADE(1)
2
6 pg_upgrade - upgrade a PostgreSQL server instance
7
9 pg_upgrade -b oldbindir -B newbindir -d olddatadir -D newdatadir
10 [option...]
11
13 pg_upgrade (formerly called pg_migrator) allows data stored in
14 PostgreSQL data files to be upgraded to a later PostgreSQL major
15 version without the data dump/reload typically required for major
16 version upgrades, e.g. from 8.4.7 to the current major release of
17 PostgreSQL. It is not required for minor version upgrades, e.g. from
18 9.0.1 to 9.0.4.
19 Major PostgreSQL releases regularly add new features that often change
21 the layout of the system tables, but the internal data storage format
22 rarely changes. pg_upgrade uses this fact to perform rapid upgrades by
23 creating new system tables and simply reusing the old user data files.
24 If a future major release ever changes the data storage format in a way
25 that makes the old data format unreadable, pg_upgrade will not be
26 usable for such upgrades. (The community will attempt to avoid such
27 situations.)
28 pg_upgrade does its best to make sure the old and new clusters are
30 binary-compatible, e.g. by checking for compatible compile-time
31 settings, including 32/64-bit binaries. It is important that any
32 external modules are also binary compatible, though this cannot be
33 checked by pg_upgrade.
34 pg_upgrade supports upgrades from 8.3.X and later to the current major
36 release of PostgreSQL, including snapshot and alpha releases.
37
39 pg_upgrade accepts the following command-line arguments:
40 -b old_bindir, --old-bindir=old_bindir
42 the old cluster executable directory; environment variable PGBINOLD
43 -B new_bindir, --new-bindir=new_bindir
45 the new cluster executable directory; environment variable PGBINNEW
46 -c, --check
48 check clusters only, don't change any data
49 -d old_datadir, --old-datadir=old_datadir
51 the old cluster data directory; environment variable PGDATAOLD
52 -D new_datadir, --new-datadir=new_datadir
54 the new cluster data directory; environment variable PGDATANEW
55 -k, --link
57 use hard links instead of copying files to the new cluster
58 -o options, --old-options options
60 options to be passed directly to the old postgres command
61 -O options, --new-options options
63 options to be passed directly to the new postgres command
64 -p old_port_number, --old-port=old_portnum
66 the old cluster port number; environment variable PGPORTOLD
67 -P new_port_number, --new-port=new_portnum
69 the new cluster port number; environment variable PGPORTNEW
70 -r, --retain
72 retain SQL and log files even after successful completion
73 -u user_name, --user=user_name
75 cluster's super user name; environment variable PGUSER
76 -v, --verbose
78 enable verbose internal logging
79 -V, --version
81 display version information, then exit
82 -?, -h, --help
84 show help, then exit
85
87 These are the steps to perform an upgrade with pg_upgrade:
88 1. Optionally move the old cluster: If you are using a
90 version-specific installation directory, e.g. /opt/PostgreSQL/9.1,
91 you do not need to move the old cluster. The graphical installers
92 all use version-specific installation directories.
93 If your installation directory is not version-specific, e.g.
95 /usr/local/pgsql, it is necessary to move the current PostgreSQL
96 install directory so it does not interfere with the new PostgreSQL
97 installation. Once the current PostgreSQL server is shut down, it
98 is safe to rename the PostgreSQL installation directory; assuming
99 the old directory is /usr/local/pgsql, you can do:
100 mv /usr/local/pgsql /usr/local/pgsql.old
102 to rename the directory.
104 2. For source installs, build the new version: Build the new
106 PostgreSQL source with configure flags that are compatible with the
107 old cluster. pg_upgrade will check pg_controldata to make sure all
108 settings are compatible before starting the upgrade.
109 3. Install the new PostgreSQL binaries: Install the new server's
111 binaries and support files.
112 For source installs, if you wish to install the new server in a
114 custom location, use the prefix variable:
115 gmake prefix=/usr/local/pgsql.new install
117 4. Install pg_upgrade and pg_upgrade_support: Install the pg_upgrade
119 binary and pg_upgrade_support library in the new PostgreSQL
120 cluster.
121 5. Initialize the new PostgreSQL cluster: Initialize the new cluster
123 using initdb. Again, use compatible initdb flags that match the old
124 cluster. Many prebuilt installers do this step automatically. There
125 is no need to start the new cluster.
126 6. Install custom shared object files: Install any custom shared
128 object files (or DLLs) used by the old cluster into the new
129 cluster, e.g. pgcrypto.so, whether they are from contrib or some
130 other source. Do not install the schema definitions, e.g.
131 pgcrypto.sql, because these will be upgraded from the old cluster.
132 7. Adjust authentication: pg_upgrade will connect to the old and new
134 servers several times, so you might want to set authentication to
135 peer in pg_hba.conf or use a ~/.pgpass file (see Section 31.15,
136 “The Password File”, in the documentation).
137 8. Stop both servers: Make sure both database servers are stopped
139 using, on Unix, e.g.:
140 pg_ctl -D /opt/PostgreSQL/8.4 stop
142 pg_ctl -D /opt/PostgreSQL/9.0 stop
143 or on Windows, using the proper service names:
145 NET STOP postgresql-8.4
147 NET STOP postgresql-9.0
148 or
150 NET STOP pgsql-8.3 (PostgreSQL 8.3 and older used a different service name)
152 9. Run pg_upgrade: Always run the pg_upgrade binary of the new server,
154 not the old one. pg_upgrade requires the specification of the old
155 and new cluster's data and executable (bin) directories. You can
156 also specify user and port values, and whether you want the data
157 linked instead of copied (the default).
158 If you use link mode, the upgrade will be much faster (no file
160 copying), but you will not be able to access your old cluster once
161 you start the new cluster after the upgrade. Link mode also
162 requires that the old and new cluster data directories be in the
163 same file system. See pg_upgrade --help for a full list of options.
164 For Windows users, you must be logged into an administrative
166 account, and then start a shell as the postgres user and set the
167 proper path:
168 RUNAS /USER:postgres "CMD.EXE"
170 SET PATH=%PATH%;C:\Program Files\PostgreSQL\9.0\bin;
171 and then run pg_upgrade with quoted directories, e.g.:
173 pg_upgrade.exe
175 --old-datadir "C:/Program Files/PostgreSQL/8.4/data"
176 --new-datadir "C:/Program Files/PostgreSQL/9.0/data"
177 --old-bindir "C:/Program Files/PostgreSQL/8.4/bin"
178 --new-bindir "C:/Program Files/PostgreSQL/9.0/bin"
179 Once started, pg_upgrade will verify the two clusters are
181 compatible and then do the upgrade. You can use pg_upgrade --check
182 to perform only the checks, even if the old server is still
183 running. pg_upgrade --check will also outline any manual
184 adjustments you will need to make after the upgrade. If you are
185 going to be using link mode, you should use the --link option with
186 --check to enable link-mode-specific checks. pg_upgrade requires
187 write permission in the current directory.
188 Obviously, no one should be accessing the clusters during the
190 upgrade. pg_upgrade defaults to running servers on port 50432 to
191 avoid unintended client connections. You can use the same port
192 number for both clusters when doing an upgrade because the old and
193 new clusters will not be running at the same time. However, when
194 checking an old running server, the old and new port numbers must
195 be different.
196 If an error occurs while restoring the database schema, pg_upgrade
198 will exit and you will have to revert to the old cluster as
199 outlined in Step 14 below. To try pg_upgrade again, you will need
200 to modify the old cluster so the pg_upgrade schema restore
201 succeeds. If the problem is a contrib module, you might need to
202 uninstall the contrib module from the old cluster and install it in
203 the new cluster after the upgrade, assuming the module is not being
204 used to store user data.
205 10. Restore pg_hba.conf: If you modified pg_hba.conf, restore its
207 original settings. It might also be necessary to adjust other
208 configuration files in the new cluster to match the old cluster,
209 e.g. postgresql.conf.
210 11. Post-Upgrade processing: If any post-upgrade processing is
212 required, pg_upgrade will issue warnings as it completes. It will
213 also generate script files that must be run by the administrator.
214 The script files will connect to each database that needs
215 post-upgrade processing. Each script should be run using:
216 psql --username postgres --file script.sql postgres
218 The scripts can be run in any order and can be deleted once they
220 have been run.
221 Caution
223 In general it is unsafe to access tables referenced in rebuild
224 scripts until the rebuild scripts have run to completion; doing
225 so could yield incorrect results or poor performance. Tables
226 not referenced in rebuild scripts can be accessed immediately.
227 12. Statistics: Because optimizer statistics are not transferred by
229 pg_upgrade, you will be instructed to run a command to regenerate
230 that information at the end of the upgrade.
231 13. Delete old cluster: Once you are satisfied with the upgrade, you
233 can delete the old cluster's data directories by running the script
234 mentioned when pg_upgrade completes. You can also delete the old
235 installation directories (e.g. bin, share).
236 14. Reverting to old cluster: If, after running pg_upgrade, you wish
238 to revert to the old cluster, there are several options:
239 · If you ran pg_upgrade with --check, no modifications were made
241 to the old cluster and you can re-use it anytime.
242 · If you ran pg_upgrade with --link, the data files are shared
244 between the old and new cluster. If you started the new
245 cluster, the new server has written to those shared files and
246 it is unsafe to use the old cluster.
247 · If you ran pg_upgradewithout--link or did not start the new
249 server, the old cluster was not modified except that, if
250 linking started, a .old suffix was appended to
251 $PGDATA/global/pg_control. To reuse the old cluster, possibly
252 remove the .old suffix from $PGDATA/global/pg_control; you can
253 then restart the old cluster.
254
257 pg_upgrade does not support upgrading of databases containing these
258 reg* OID-referencing system data types: regproc, regprocedure, regoper,
259 regoperator, regconfig, and regdictionary. (regtype can be upgraded.)
260 All failure, rebuild, and reindex cases will be reported by pg_upgrade
262 if they affect your installation; post-upgrade scripts to rebuild
263 tables and indexes will be generated automatically.
264 For deployment testing, create a schema-only copy of the old cluster,
266 insert dummy data, and upgrade that.
267 If you are upgrading a pre-PostgreSQL 9.2 cluster that uses a
269 configuration-file-only directory, you must pass the real data
270 directory location to pg_upgrade, and pass the configuration directory
271 location to the server, e.g. -d /real-data-directory -o '-D
272 /configuration-directory'.
273 If using a pre-9.1 old server that is using a non-default Unix-domain
275 socket directory or a default that differs from the default of the new
276 cluster, set PGHOST to point to the old server's socket location. (This
277 is not relevant on Windows.)
278 A Log-Shipping Standby Server (Section 25.2, “Log-Shipping Standby
280 Servers”, in the documentation) cannot be upgraded because the server
281 must allow writes. The simplest way is to upgrade the primary and use
282 rsync to rebuild the standbys. You can run rsync while the primary is
283 down, or as part of a base backup (Section 24.3.2, “Making a Base
284 Backup”, in the documentation) which overwrites the old standby
285 cluster.
286 If you want to use link mode and you do not want your old cluster to be
288 modified when the new cluster is started, make a copy of the old
289 cluster and upgrade that in link mode. To make a valid copy of the old
290 cluster, use rsync to create a dirty copy of the old cluster while the
291 server is running, then shut down the old server and run rsync again to
292 update the copy with any changes to make it consistent. You might want
293 to exclude some files, e.g. postmaster.pid, as documented in Section
294 24.3.3, “Making a Base Backup Using the Low Level API”, in the
295 documentation.
296 Limitations in Upgrading from PostgreSQL 8.3
298 Upgrading from PostgreSQL 8.3 has additional restrictions not present
299 when upgrading from later PostgreSQL releases. For example, pg_upgrade
300 will not work for upgrading from 8.3 if a user column is defined as:
301 · a tsquery data type
303 · data type name and is not the first column
305 You must drop any such columns and upgrade them manually.
307 pg_upgrade will not work if the ltree contrib module is installed in a
309 database.
310 pg_upgrade will require a table rebuild if:
312 · a user column is of data type tsvector
314 pg_upgrade will require a reindex if:
316 · an index is of type hash or GIN
318 · an index uses bpchar_pattern_ops
320 Also, the default datetime storage format changed to integer after
322 PostgreSQL 8.3. pg_upgrade will check that the datetime storage format
323 used by the old and new clusters match. Make sure your new cluster is
324 built with the configure flag --disable-integer-datetimes.
325 For Windows users, note that due to different integer datetimes
327 settings used by the graphical installer and the MSI installer, it is
328 only possible to upgrade from version 8.3 of the installer distribution
329 to version 8.4 or later of the installer distribution. It is not
330 possible to upgrade from the MSI installer to the new graphical
331 installer.
332
334 initdb(1), pg_ctl(1), pg_dump(1), postgres(1)
335PostgreSQL 9.2.24 2017-11-06 PG_UPGRADE(1)