<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>pg_upgrade</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK
REV="MADE"
HREF="mailto:pgsql-docs@postgresql.org"><LINK
REL="HOME"
TITLE="PostgreSQL 9.6.4 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="PostgreSQL Server Applications"
HREF="reference-server.html"><LINK
REL="PREVIOUS"
TITLE="pg_test_timing"
HREF="pgtesttiming.html"><LINK
REL="NEXT"
TITLE="pg_xlogdump"
HREF="pgxlogdump.html"><LINK
REL="STYLESHEET"
TYPE="text/css"
HREF="stylesheet.css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=ISO-8859-1"><META
NAME="creation"
CONTENT="2017-08-11T02:27:18"></HEAD
><BODY
CLASS="REFENTRY"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="4"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.6.4 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="pg_test_timing"
HREF="pgtesttiming.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="reference-server.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
></TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="pg_xlogdump"
HREF="pgxlogdump.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><H1
><A
NAME="PGUPGRADE"
></A
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></H1
><DIV
CLASS="REFNAMEDIV"
><A
NAME="AEN102972"
></A
><H2
>Name</H2
>pg_upgrade -- upgrade a <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server instance</DIV
><DIV
CLASS="REFSYNOPSISDIV"
><A
NAME="AEN102976"
></A
><H2
>Synopsis</H2
><P
><TT
CLASS="COMMAND"
>pg_upgrade</TT
> <TT
CLASS="OPTION"
>-b</TT
> <TT
CLASS="REPLACEABLE"
><I
>oldbindir</I
></TT
> <TT
CLASS="OPTION"
>-B</TT
> <TT
CLASS="REPLACEABLE"
><I
>newbindir</I
></TT
> <TT
CLASS="OPTION"
>-d</TT
> <TT
CLASS="REPLACEABLE"
><I
>olddatadir</I
></TT
> <TT
CLASS="OPTION"
>-D</TT
> <TT
CLASS="REPLACEABLE"
><I
>newdatadir</I
></TT
> [<TT
CLASS="REPLACEABLE"
><I
>option</I
></TT
>...]</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN102997"
></A
><H2
>Description</H2
><P
> <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> (formerly called <SPAN
CLASS="APPLICATION"
>pg_migrator</SPAN
>) allows data
stored in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> data files to be upgraded to a later <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
major version without the data dump/reload typically required for
major version upgrades, e.g. from 8.4.7 to the current major release
of <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>. It is not required for minor version upgrades, e.g. from
9.0.1 to 9.0.4.
</P
><P
> Major PostgreSQL releases regularly add new features that often
change the layout of the system tables, but the internal data storage
format rarely changes. <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> uses this fact
to perform rapid upgrades by creating new system tables and simply
reusing the old user data files. If a future major release ever
changes the data storage format in a way that makes the old data
format unreadable, <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> will not be usable
for such upgrades. (The community will attempt to avoid such
situations.)
</P
><P
> <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> does its best to
make sure the old and new clusters are binary-compatible, e.g. by
checking for compatible compile-time settings, including 32/64-bit
binaries. It is important that
any external modules are also binary compatible, though this cannot
be checked by <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>.
</P
><P
> pg_upgrade supports upgrades from 8.4.X and later to the current
major release of <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>, including snapshot and alpha releases.
</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN103013"
></A
><H2
>Options</H2
><P
> <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> accepts the following command-line arguments:
<P
></P
></P><DIV
CLASS="VARIABLELIST"
><DL
><DT
><TT
CLASS="OPTION"
>-b</TT
> <TT
CLASS="REPLACEABLE"
><I
>bindir</I
></TT
><BR><TT
CLASS="OPTION"
>--old-bindir=</TT
><TT
CLASS="REPLACEABLE"
><I
>bindir</I
></TT
></DT
><DD
><P
>the old PostgreSQL executable directory;
environment variable <TT
CLASS="ENVAR"
>PGBINOLD</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-B</TT
> <TT
CLASS="REPLACEABLE"
><I
>bindir</I
></TT
><BR><TT
CLASS="OPTION"
>--new-bindir=</TT
><TT
CLASS="REPLACEABLE"
><I
>bindir</I
></TT
></DT
><DD
><P
>the new PostgreSQL executable directory;
environment variable <TT
CLASS="ENVAR"
>PGBINNEW</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-c</TT
><BR><TT
CLASS="OPTION"
>--check</TT
></DT
><DD
><P
>check clusters only, don't change any data</P
></DD
><DT
><TT
CLASS="OPTION"
>-d</TT
> <TT
CLASS="REPLACEABLE"
><I
>datadir</I
></TT
><BR><TT
CLASS="OPTION"
>--old-datadir=</TT
><TT
CLASS="REPLACEABLE"
><I
>datadir</I
></TT
></DT
><DD
><P
>the old cluster data directory; environment
variable <TT
CLASS="ENVAR"
>PGDATAOLD</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-D</TT
> <TT
CLASS="REPLACEABLE"
><I
>datadir</I
></TT
><BR><TT
CLASS="OPTION"
>--new-datadir=</TT
><TT
CLASS="REPLACEABLE"
><I
>datadir</I
></TT
></DT
><DD
><P
>the new cluster data directory; environment
variable <TT
CLASS="ENVAR"
>PGDATANEW</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-j</TT
><BR><TT
CLASS="OPTION"
>--jobs</TT
></DT
><DD
><P
>number of simultaneous processes or threads to use
</P
></DD
><DT
><TT
CLASS="OPTION"
>-k</TT
><BR><TT
CLASS="OPTION"
>--link</TT
></DT
><DD
><P
>use hard links instead of copying files to the new
cluster</P
></DD
><DT
><TT
CLASS="OPTION"
>-o</TT
> <TT
CLASS="REPLACEABLE"
><I
>options</I
></TT
><BR><TT
CLASS="OPTION"
>--old-options</TT
> <TT
CLASS="REPLACEABLE"
><I
>options</I
></TT
></DT
><DD
><P
>options to be passed directly to the
old <TT
CLASS="COMMAND"
>postgres</TT
> command; multiple
option invocations are appended</P
></DD
><DT
><TT
CLASS="OPTION"
>-O</TT
> <TT
CLASS="REPLACEABLE"
><I
>options</I
></TT
><BR><TT
CLASS="OPTION"
>--new-options</TT
> <TT
CLASS="REPLACEABLE"
><I
>options</I
></TT
></DT
><DD
><P
>options to be passed directly to the
new <TT
CLASS="COMMAND"
>postgres</TT
> command; multiple
option invocations are appended</P
></DD
><DT
><TT
CLASS="OPTION"
>-p</TT
> <TT
CLASS="REPLACEABLE"
><I
>port</I
></TT
><BR><TT
CLASS="OPTION"
>--old-port=</TT
><TT
CLASS="REPLACEABLE"
><I
>port</I
></TT
></DT
><DD
><P
>the old cluster port number; environment
variable <TT
CLASS="ENVAR"
>PGPORTOLD</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-P</TT
> <TT
CLASS="REPLACEABLE"
><I
>port</I
></TT
><BR><TT
CLASS="OPTION"
>--new-port=</TT
><TT
CLASS="REPLACEABLE"
><I
>port</I
></TT
></DT
><DD
><P
>the new cluster port number; environment
variable <TT
CLASS="ENVAR"
>PGPORTNEW</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-r</TT
><BR><TT
CLASS="OPTION"
>--retain</TT
></DT
><DD
><P
>retain SQL and log files even after successful completion
</P
></DD
><DT
><TT
CLASS="OPTION"
>-U</TT
> <TT
CLASS="REPLACEABLE"
><I
>username</I
></TT
><BR><TT
CLASS="OPTION"
>--username=</TT
><TT
CLASS="REPLACEABLE"
><I
>username</I
></TT
></DT
><DD
><P
>cluster's install user name; environment
variable <TT
CLASS="ENVAR"
>PGUSER</TT
></P
></DD
><DT
><TT
CLASS="OPTION"
>-v</TT
><BR><TT
CLASS="OPTION"
>--verbose</TT
></DT
><DD
><P
>enable verbose internal logging</P
></DD
><DT
><TT
CLASS="OPTION"
>-V</TT
><BR><TT
CLASS="OPTION"
>--version</TT
></DT
><DD
><P
>display version information, then exit</P
></DD
><DT
><TT
CLASS="OPTION"
>-?</TT
><BR><TT
CLASS="OPTION"
>--help</TT
></DT
><DD
><P
>show help, then exit</P
></DD
></DL
></DIV
><P>
</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN103157"
></A
><H2
>Usage</H2
><P
> These are the steps to perform an upgrade
with <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>:
</P
><DIV
CLASS="PROCEDURE"
><OL
TYPE="1"
><LI
CLASS="STEP"
><P
><B
>Optionally move the old cluster</B
></P
><P
> If you are using a version-specific installation directory, e.g.
<TT
CLASS="FILENAME"
>/opt/PostgreSQL/9.1</TT
>, you do not need to move the old cluster. The
graphical installers all use version-specific installation directories.
</P
><P
> If your installation directory is not version-specific, e.g.
<TT
CLASS="FILENAME"
>/usr/local/pgsql</TT
>, it is necessary to move the current PostgreSQL install
directory so it does not interfere with the new <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> installation.
Once the current <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server is shut down, it is safe to rename the
PostgreSQL installation directory; assuming the old directory is
<TT
CLASS="FILENAME"
>/usr/local/pgsql</TT
>, you can do:
</P><PRE
CLASS="PROGRAMLISTING"
>mv /usr/local/pgsql /usr/local/pgsql.old</PRE
><P>
to rename the directory.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>For source installs, build the new version</B
></P
><P
> Build the new PostgreSQL source with <TT
CLASS="COMMAND"
>configure</TT
> flags that are compatible
with the old cluster. <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> will check <TT
CLASS="COMMAND"
>pg_controldata</TT
> to make
sure all settings are compatible before starting the upgrade.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Install the new PostgreSQL binaries</B
></P
><P
> Install the new server's binaries and support
files. <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> is included in a default installation.
</P
><P
> For source installs, if you wish to install the new server in a custom
location, use the <TT
CLASS="LITERAL"
>prefix</TT
> variable:
</P><PRE
CLASS="PROGRAMLISTING"
>make prefix=/usr/local/pgsql.new install</PRE
><P></P
></LI
><LI
CLASS="STEP"
><P
><B
>Initialize the new PostgreSQL cluster</B
></P
><P
> Initialize the new cluster using <TT
CLASS="COMMAND"
>initdb</TT
>.
Again, use compatible <TT
CLASS="COMMAND"
>initdb</TT
>
flags that match the old cluster. Many
prebuilt installers do this step automatically. There is no need to
start the new cluster.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Install custom shared object files</B
></P
><P
> Install any custom shared object files (or DLLs) used by the old cluster
into the new cluster, e.g. <TT
CLASS="FILENAME"
>pgcrypto.so</TT
>,
whether they are from <TT
CLASS="FILENAME"
>contrib</TT
>
or some other source. Do not install the schema definitions, e.g.
<TT
CLASS="COMMAND"
>CREATE EXTENSION pgcrypto</TT
>, because these will be upgraded
from the old cluster.
Also, any custom full text search files (dictionary, synonym,
thesaurus, stop words) must also be copied to the new cluster.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Adjust authentication</B
></P
><P
> <TT
CLASS="COMMAND"
>pg_upgrade</TT
> will connect to the old and new servers several
times, so you might want to set authentication to <TT
CLASS="LITERAL"
>peer</TT
>
in <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> or use a <TT
CLASS="FILENAME"
>~/.pgpass</TT
> file
(see <A
HREF="libpq-pgpass.html"
>Section 32.15</A
>).
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Stop both servers</B
></P
><P
> Make sure both database servers are stopped using, on Unix, e.g.:
</P><PRE
CLASS="PROGRAMLISTING"
>pg_ctl -D /opt/PostgreSQL/8.4 stop
pg_ctl -D /opt/PostgreSQL/9.0 stop</PRE
><P>
or on Windows, using the proper service names:
</P><PRE
CLASS="PROGRAMLISTING"
>NET STOP postgresql-8.4
NET STOP postgresql-9.0</PRE
><P>
</P
><P
> Streaming replication and log-shipping standby servers can
remain running until a later step.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Prepare for standby server upgrades</B
></P
><P
> If you are upgrading standby servers (as outlined in section <A
HREF="pgupgrade.html#PGUPGRADE-STEP-REPLICAS"
>step 10</A
>), verify that the old standby
servers are caught up by running <SPAN
CLASS="APPLICATION"
>pg_controldata</SPAN
>
against the old primary and standby clusters. Verify that the
<SPAN
CLASS="QUOTE"
>"Latest checkpoint location"</SPAN
> values match in all clusters.
(There will be a mismatch if old standby servers were shut down
before the old primary.)
</P
><P
> Also, if upgrading standby servers, change <TT
CLASS="VARNAME"
>wal_level</TT
>
to <TT
CLASS="LITERAL"
>replica</TT
> in the <TT
CLASS="FILENAME"
>postgresql.conf</TT
> file on
the new master cluster.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Run <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></B
></P
><P
> Always run the <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> binary of the new server, not the old one.
<SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> requires the specification of the old and new cluster's
data and executable (<TT
CLASS="FILENAME"
>bin</TT
>) directories. You can also specify
user and port values, and whether you want the data linked instead of
copied (the default).
</P
><P
> If you use link mode, the upgrade will be much faster (no file
copying) and use less disk space, but you will not be able to access
your old cluster
once you start the new cluster after the upgrade. Link mode also
requires that the old and new cluster data directories be in the
same file system. (Tablespaces and <TT
CLASS="FILENAME"
>pg_xlog</TT
> can be on
different file systems.) See <TT
CLASS="LITERAL"
>pg_upgrade --help</TT
> for a full
list of options.
</P
><P
> The <TT
CLASS="OPTION"
>--jobs</TT
> option allows multiple CPU cores to be used
for copying/linking of files and to dump and reload database schemas
in parallel; a good place to start is the maximum of the number of
CPU cores and tablespaces. This option can dramatically reduce the
time to upgrade a multi-database server running on a multiprocessor
machine.
</P
><P
> For Windows users, you must be logged into an administrative account, and
then start a shell as the <TT
CLASS="LITERAL"
>postgres</TT
> user and set the proper path:
</P><PRE
CLASS="PROGRAMLISTING"
>RUNAS /USER:postgres "CMD.EXE"
SET PATH=%PATH%;C:\Program Files\PostgreSQL\9.0\bin;</PRE
><P>
and then run <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> with quoted directories, e.g.:
</P><PRE
CLASS="PROGRAMLISTING"
>pg_upgrade.exe
--old-datadir "C:/Program Files/PostgreSQL/8.4/data"
--new-datadir "C:/Program Files/PostgreSQL/9.0/data"
--old-bindir "C:/Program Files/PostgreSQL/8.4/bin"
--new-bindir "C:/Program Files/PostgreSQL/9.0/bin"</PRE
><P>
Once started, <TT
CLASS="COMMAND"
>pg_upgrade</TT
> will verify the two clusters are compatible
and then do the upgrade. You can use <TT
CLASS="COMMAND"
>pg_upgrade --check</TT
>
to perform only the checks, even if the old server is still
running. <TT
CLASS="COMMAND"
>pg_upgrade --check</TT
> will also outline any
manual adjustments you will need to make after the upgrade. If you
are going to be using link mode, you should use the <TT
CLASS="OPTION"
>--link</TT
>
option with <TT
CLASS="OPTION"
>--check</TT
> to enable link-mode-specific checks.
<TT
CLASS="COMMAND"
>pg_upgrade</TT
> requires write permission in the current directory.
</P
><P
> Obviously, no one should be accessing the clusters during the
upgrade. <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> defaults to running servers
on port 50432 to avoid unintended client connections.
You can use the same port number for both clusters when doing an
upgrade because the old and new clusters will not be running at the
same time. However, when checking an old running server, the old
and new port numbers must be different.
</P
><P
> If an error occurs while restoring the database schema, <TT
CLASS="COMMAND"
>pg_upgrade</TT
> will
exit and you will have to revert to the old cluster as outlined in <A
HREF="pgupgrade.html#PGUPGRADE-STEP-REVERT"
>step 16</A
>
below. To try <TT
CLASS="COMMAND"
>pg_upgrade</TT
> again, you will need to modify the old
cluster so the pg_upgrade schema restore succeeds. If the problem is a
<TT
CLASS="FILENAME"
>contrib</TT
> module, you might need to uninstall the <TT
CLASS="FILENAME"
>contrib</TT
> module from
the old cluster and install it in the new cluster after the upgrade,
assuming the module is not being used to store user data.
</P
></LI
><LI
CLASS="STEP"
><A
NAME="PGUPGRADE-STEP-REPLICAS"
></A
><P
><B
>Upgrade Streaming Replication and Log-Shipping standby servers</B
></P
><P
> If you have Streaming Replication (see <A
HREF="warm-standby.html#STREAMING-REPLICATION"
>Section 26.2.5</A
>) or Log-Shipping (see <A
HREF="warm-standby.html"
>Section 26.2</A
>) standby servers, follow these steps to
upgrade them. You will not be running <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>
on the standby servers, but rather <SPAN
CLASS="APPLICATION"
>rsync</SPAN
>. Do not
start any servers yet.
</P
><OL
CLASS="SUBSTEPS"
TYPE="a"
><LI
CLASS="STEP"
><P
><B
>Install the new PostgreSQL binaries on standby servers</B
></P
><P
> Make sure the new binaries and support files are installed on all
standby servers.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Make sure the new standby data directories do <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
> exist</B
></P
><P
> Make sure the new standby data directories do <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
>
exist or are empty. If <SPAN
CLASS="APPLICATION"
>initdb</SPAN
> was run, delete
the standby server data directories.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Install custom shared object files</B
></P
><P
> Install the same custom shared object files on the new standbys
that you installed in the new master cluster.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Stop standby servers</B
></P
><P
> If the standby servers are still running, stop them now using the
above instructions.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Save configuration files</B
></P
><P
> Save any configuration files from the standbys you need to keep,
e.g. <TT
CLASS="FILENAME"
>postgresql.conf</TT
>, <TT
CLASS="LITERAL"
>recovery.conf</TT
>,
as these will be overwritten or removed in the next step.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Run <SPAN
CLASS="APPLICATION"
>rsync</SPAN
></B
></P
><P
> From a directory that is above the old and new database cluster
directories, run this for each slave:
</P><PRE
CLASS="PROGRAMLISTING"
>rsync --archive --delete --hard-links --size-only old_pgdata new_pgdata remote_dir</PRE
><P>
where <TT
CLASS="OPTION"
>old_pgdata</TT
> and <TT
CLASS="OPTION"
>new_pgdata</TT
> are relative
to the current directory, and <TT
CLASS="OPTION"
>remote_dir</TT
> is
<SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>above</I
></SPAN
> the old and new cluster directories on
the standby server. The old and new relative cluster paths
must match on the master and standby server. Consult the
<SPAN
CLASS="APPLICATION"
>rsync</SPAN
> manual page for details on specifying the
remote directory, e.g. <TT
CLASS="LITERAL"
>standbyhost:/opt/PostgreSQL/</TT
>.
<SPAN
CLASS="APPLICATION"
>rsync</SPAN
> will be fast when <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>'s
<TT
CLASS="OPTION"
>--link</TT
> mode is used because it will create hard links
on the remote server rather than transferring user data.
Unfortunately, <SPAN
CLASS="APPLICATION"
>rsync</SPAN
> needlessly copies the
files associated with temporary and unlogged tables.
</P
><P
> If you have tablespaces, you will need to run a similar
<SPAN
CLASS="APPLICATION"
>rsync</SPAN
> command for each tablespace directory. If you
have relocated <TT
CLASS="FILENAME"
>pg_xlog</TT
> outside the data directories,
<SPAN
CLASS="APPLICATION"
>rsync</SPAN
> must be run on those directories too.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Configure streaming replication and log-shipping standby servers</B
></P
><P
> Configure the servers for log shipping. (You do not need to run
<CODE
CLASS="FUNCTION"
>pg_start_backup()</CODE
> and <CODE
CLASS="FUNCTION"
>pg_stop_backup()</CODE
>
or take a file system backup as the slaves are still synchronized
with the master.)
</P
></LI
></OL
></LI
><LI
CLASS="STEP"
><P
><B
>Restore <TT
CLASS="FILENAME"
>pg_hba.conf</TT
></B
></P
><P
> If you modified <TT
CLASS="FILENAME"
>pg_hba.conf</TT
>, restore its original settings.
It might also be necessary to adjust other configuration files in the new
cluster to match the old cluster, e.g. <TT
CLASS="FILENAME"
>postgresql.conf</TT
>.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Start the new server</B
></P
><P
> The new server can now be safely started, and then any
<SPAN
CLASS="APPLICATION"
>rsync</SPAN
>'ed standby servers.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Post-Upgrade processing</B
></P
><P
> If any post-upgrade processing is required, pg_upgrade will issue
warnings as it completes. It will also generate script files that must
be run by the administrator. The script files will connect to each
database that needs post-upgrade processing. Each script should be
run using:
</P><PRE
CLASS="PROGRAMLISTING"
>psql --username postgres --file script.sql postgres</PRE
><P>
The scripts can be run in any order and can be deleted once they have
been run.
</P
><DIV
CLASS="CAUTION"
><P
></P
><TABLE
CLASS="CAUTION"
BORDER="1"
WIDTH="100%"
><TR
><TD
ALIGN="CENTER"
><B
>Caution</B
></TD
></TR
><TR
><TD
ALIGN="LEFT"
><P
> In general it is unsafe to access tables referenced in rebuild scripts
until the rebuild scripts have run to completion; doing so could yield
incorrect results or poor performance. Tables not referenced in rebuild
scripts can be accessed immediately.
</P
></TD
></TR
></TABLE
></DIV
></LI
><LI
CLASS="STEP"
><P
><B
>Statistics</B
></P
><P
> Because optimizer statistics are not transferred by <TT
CLASS="COMMAND"
>pg_upgrade</TT
>, you will
be instructed to run a command to regenerate that information at the end
of the upgrade. You might need to set connection parameters to
match your new cluster.
</P
></LI
><LI
CLASS="STEP"
><P
><B
>Delete old cluster</B
></P
><P
> Once you are satisfied with the upgrade, you can delete the old
cluster's data directories by running the script mentioned when
<TT
CLASS="COMMAND"
>pg_upgrade</TT
> completes. (Automatic deletion is not
possible if you have user-defined tablespaces inside the old data
directory.) You can also delete the old installation directories
(e.g. <TT
CLASS="FILENAME"
>bin</TT
>, <TT
CLASS="FILENAME"
>share</TT
>).
</P
></LI
><LI
CLASS="STEP"
><A
NAME="PGUPGRADE-STEP-REVERT"
></A
><P
><B
>Reverting to old cluster</B
></P
><P
> If, after running <TT
CLASS="COMMAND"
>pg_upgrade</TT
>, you wish to revert to the old cluster,
there are several options:
<P
></P
></P><UL
><LI
><P
> If you ran <TT
CLASS="COMMAND"
>pg_upgrade</TT
>
with <TT
CLASS="OPTION"
>--check</TT
>, no modifications were made to the old
cluster and you can re-use it anytime.
</P
></LI
><LI
><P
> If you ran <TT
CLASS="COMMAND"
>pg_upgrade</TT
>
with <TT
CLASS="OPTION"
>--link</TT
>, the data files are shared between the
old and new cluster. If you started the new cluster, the new
server has written to those shared files and it is unsafe to
use the old cluster.
</P
></LI
><LI
><P
> If you ran <TT
CLASS="COMMAND"
>pg_upgrade</TT
> <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>without</I
></SPAN
>
<TT
CLASS="OPTION"
>--link</TT
> or did not start the new server, the
old cluster was not modified except that, if linking
started, a <TT
CLASS="LITERAL"
>.old</TT
> suffix was appended to
<TT
CLASS="FILENAME"
>$PGDATA/global/pg_control</TT
>. To reuse the old
cluster, possibly remove the <TT
CLASS="FILENAME"
>.old</TT
> suffix from
<TT
CLASS="FILENAME"
>$PGDATA/global/pg_control</TT
>; you can then restart the
old cluster.
</P
></LI
></UL
><P>
</P
></LI
></OL
></DIV
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN103351"
></A
><H2
>Notes</H2
><P
> <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> does not support upgrading of databases
containing these <TT
CLASS="TYPE"
>reg*</TT
> OID-referencing system data types:
<TT
CLASS="TYPE"
>regproc</TT
>, <TT
CLASS="TYPE"
>regprocedure</TT
>, <TT
CLASS="TYPE"
>regoper</TT
>,
<TT
CLASS="TYPE"
>regoperator</TT
>, <TT
CLASS="TYPE"
>regconfig</TT
>, and
<TT
CLASS="TYPE"
>regdictionary</TT
>. (<TT
CLASS="TYPE"
>regtype</TT
> can be upgraded.)
</P
><P
> All failure, rebuild, and reindex cases will be reported by
<SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> if they affect your installation;
post-upgrade scripts to rebuild tables and indexes will be
generated automatically. If you are trying to automate the upgrade
of many clusters, you should find that clusters with identical database
schemas require the same post-upgrade steps for all cluster upgrades;
this is because the post-upgrade steps are based on the database
schemas, and not user data.
</P
><P
> For deployment testing, create a schema-only copy of the old cluster,
insert dummy data, and upgrade that.
</P
><P
> If you are upgrading a pre-<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.2 cluster
that uses a configuration-file-only directory, you must pass the
real data directory location to <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>, and
pass the configuration directory location to the server, e.g.
<TT
CLASS="LITERAL"
>-d /real-data-directory -o '-D /configuration-directory'</TT
>.
</P
><P
> If using a pre-9.1 old server that is using a non-default Unix-domain
socket directory or a default that differs from the default of the
new cluster, set <TT
CLASS="ENVAR"
>PGHOST</TT
> to point to the old server's socket
location. (This is not relevant on Windows.)
</P
><P
> If you want to use link mode and you do not want your old cluster
to be modified when the new cluster is started, make a copy of the
old cluster and upgrade that in link mode. To make a valid copy
of the old cluster, use <TT
CLASS="COMMAND"
>rsync</TT
> to create a dirty
copy of the old cluster while the server is running, then shut down
the old server and run <TT
CLASS="COMMAND"
>rsync --checksum</TT
> again to update the
copy with any changes to make it consistent. (<TT
CLASS="OPTION"
>--checksum</TT
>
is necessary because <TT
CLASS="COMMAND"
>rsync</TT
> only has file modification-time
granularity of one second.) You might want to exclude some
files, e.g. <TT
CLASS="FILENAME"
>postmaster.pid</TT
>, as documented in <A
HREF="continuous-archiving.html#BACKUP-LOWLEVEL-BASE-BACKUP"
>Section 25.3.3</A
>. If your file system supports
file system snapshots or copy-on-write file copies, you can use that
to make a backup of the old cluster and tablespaces, though the snapshot
and copies must be created simultaneously or while the database server
is down.
</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN103379"
></A
><H2
>See Also</H2
><A
HREF="app-initdb.html"
>initdb</A
>, <A
HREF="app-pg-ctl.html"
><SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
></A
>, <A
HREF="app-pgdump.html"
>pg_dump</A
>, <A
HREF="app-postgres.html"
><SPAN
CLASS="APPLICATION"
>postgres</SPAN
></A
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="pgtesttiming.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="pgxlogdump.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><SPAN
CLASS="APPLICATION"
>pg_test_timing</SPAN
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="reference-server.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><SPAN
CLASS="APPLICATION"
>pg_xlogdump</SPAN
></TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
