<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Continuous Archiving and Point-in-Time Recovery (PITR)</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 9.3.9 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Backup and Restore"
HREF="backup.html"><LINK
REL="PREVIOUS"
TITLE="File System Level Backup"
HREF="backup-file.html"><LINK
REL="NEXT"
TITLE="High Availability, Load Balancing, and Replication"
HREF="high-availability.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="2015-06-13T20:07:22"></HEAD
><BODY
CLASS="SECT1"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="5"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.3.9 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="File System Level Backup"
HREF="backup-file.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="backup.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Chapter 24. Backup and Restore</TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="High Availability, Load Balancing, and Replication"
HREF="high-availability.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="CONTINUOUS-ARCHIVING"
>24.3. Continuous Archiving and Point-in-Time Recovery (PITR)</A
></H1
><P
> At all times, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> maintains a
<I
CLASS="FIRSTTERM"
>write ahead log</I
> (WAL) in the <TT
CLASS="FILENAME"
>pg_xlog/</TT
>
subdirectory of the cluster's data directory. The log records
every change made to the database's data files. This log exists
primarily for crash-safety purposes: if the system crashes, the
database can be restored to consistency by <SPAN
CLASS="QUOTE"
>"replaying"</SPAN
> the
log entries made since the last checkpoint. However, the existence
of the log makes it possible to use a third strategy for backing up
databases: we can combine a file-system-level backup with backup of
the WAL files. If recovery is needed, we restore the file system backup and
then replay from the backed-up WAL files to bring the system to a
current state. This approach is more complex to administer than
either of the previous approaches, but it has some significant
benefits:
<P
></P
></P><UL
><LI
><P
> We do not need a perfectly consistent file system backup as the starting point.
Any internal inconsistency in the backup will be corrected by log
replay (this is not significantly different from what happens during
crash recovery). So we do not need a file system snapshot capability,
just <SPAN
CLASS="APPLICATION"
>tar</SPAN
> or a similar archiving tool.
</P
></LI
><LI
><P
> Since we can combine an indefinitely long sequence of WAL files
for replay, continuous backup can be achieved simply by continuing to archive
the WAL files. This is particularly valuable for large databases, where
it might not be convenient to take a full backup frequently.
</P
></LI
><LI
><P
> It is not necessary to replay the WAL entries all the
way to the end. We could stop the replay at any point and have a
consistent snapshot of the database as it was at that time. Thus,
this technique supports <I
CLASS="FIRSTTERM"
>point-in-time recovery</I
>: it is
possible to restore the database to its state at any time since your base
backup was taken.
</P
></LI
><LI
><P
> If we continuously feed the series of WAL files to another
machine that has been loaded with the same base backup file, we
have a <I
CLASS="FIRSTTERM"
>warm standby</I
> system: at any point we can bring up
the second machine and it will have a nearly-current copy of the
database.
</P
></LI
></UL
><P>
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> and
<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> do not produce file-system-level
backups and cannot be used as part of a continuous-archiving solution.
Such dumps are <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>logical</I
></SPAN
> and do not contain enough
information to be used by WAL replay.
</P
></BLOCKQUOTE
></DIV
><P
> As with the plain file-system-backup technique, this method can only
support restoration of an entire database cluster, not a subset.
Also, it requires a lot of archival storage: the base backup might be bulky,
and a busy system will generate many megabytes of WAL traffic that
have to be archived. Still, it is the preferred backup technique in
many situations where high reliability is needed.
</P
><P
> To recover successfully using continuous archiving (also called
<SPAN
CLASS="QUOTE"
>"online backup"</SPAN
> by many database vendors), you need a continuous
sequence of archived WAL files that extends back at least as far as the
start time of your backup. So to get started, you should set up and test
your procedure for archiving WAL files <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>before</I
></SPAN
> you take your
first base backup. Accordingly, we first discuss the mechanics of
archiving WAL files.
</P
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-ARCHIVING-WAL"
>24.3.1. Setting Up WAL Archiving</A
></H2
><P
> In an abstract sense, a running <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> system
produces an indefinitely long sequence of WAL records. The system
physically divides this sequence into WAL <I
CLASS="FIRSTTERM"
>segment
files</I
>, which are normally 16MB apiece (although the segment size
can be altered when building <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>). The segment
files are given numeric names that reflect their position in the
abstract WAL sequence. When not using WAL archiving, the system
normally creates just a few segment files and then
<SPAN
CLASS="QUOTE"
>"recycles"</SPAN
> them by renaming no-longer-needed segment files
to higher segment numbers. It's assumed that segment files whose
contents precede the checkpoint-before-last are no longer of
interest and can be recycled.
</P
><P
> When archiving WAL data, we need to capture the contents of each segment
file once it is filled, and save that data somewhere before the segment
file is recycled for reuse. Depending on the application and the
available hardware, there could be many different ways of <SPAN
CLASS="QUOTE"
>"saving
the data somewhere"</SPAN
>: we could copy the segment files to an NFS-mounted
directory on another machine, write them onto a tape drive (ensuring that
you have a way of identifying the original name of each file), or batch
them together and burn them onto CDs, or something else entirely. To
provide the database administrator with flexibility,
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> tries not to make any assumptions about how
the archiving will be done. Instead, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> lets
the administrator specify a shell command to be executed to copy a
completed segment file to wherever it needs to go. The command could be
as simple as a <TT
CLASS="LITERAL"
>cp</TT
>, or it could invoke a complex shell
script — it's all up to you.
</P
><P
> To enable WAL archiving, set the <A
HREF="runtime-config-wal.html#GUC-WAL-LEVEL"
>wal_level</A
>
configuration parameter to <TT
CLASS="LITERAL"
>archive</TT
> (or <TT
CLASS="LITERAL"
>hot_standby</TT
>),
<A
HREF="runtime-config-wal.html#GUC-ARCHIVE-MODE"
>archive_mode</A
> to <TT
CLASS="LITERAL"
>on</TT
>,
and specify the shell command to use in the <A
HREF="runtime-config-wal.html#GUC-ARCHIVE-COMMAND"
>archive_command</A
> configuration parameter. In practice
these settings will always be placed in the
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> file.
In <TT
CLASS="VARNAME"
>archive_command</TT
>,
<TT
CLASS="LITERAL"
>%p</TT
> is replaced by the path name of the file to
archive, while <TT
CLASS="LITERAL"
>%f</TT
> is replaced by only the file name.
(The path name is relative to the current working directory,
i.e., the cluster's data directory.)
Use <TT
CLASS="LITERAL"
>%%</TT
> if you need to embed an actual <TT
CLASS="LITERAL"
>%</TT
>
character in the command. The simplest useful command is something
like:
</P><PRE
CLASS="PROGRAMLISTING"
>archive_command = 'test ! -f /mnt/server/archivedir/%f && cp %p /mnt/server/archivedir/%f' # Unix
archive_command = 'copy "%p" "C:\\server\\archivedir\\%f"' # Windows</PRE
><P>
which will copy archivable WAL segments to the directory
<TT
CLASS="FILENAME"
>/mnt/server/archivedir</TT
>. (This is an example, not a
recommendation, and might not work on all platforms.) After the
<TT
CLASS="LITERAL"
>%p</TT
> and <TT
CLASS="LITERAL"
>%f</TT
> parameters have been replaced,
the actual command executed might look like this:
</P><PRE
CLASS="PROGRAMLISTING"
>test ! -f /mnt/server/archivedir/00000001000000A900000065 && cp pg_xlog/00000001000000A900000065 /mnt/server/archivedir/00000001000000A900000065</PRE
><P>
A similar command will be generated for each new file to be archived.
</P
><P
> The archive command will be executed under the ownership of the same
user that the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server is running as. Since
the series of WAL files being archived contains effectively everything
in your database, you will want to be sure that the archived data is
protected from prying eyes; for example, archive into a directory that
does not have group or world read access.
</P
><P
> It is important that the archive command return zero exit status if and
only if it succeeds. Upon getting a zero result,
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> will assume that the file has been
successfully archived, and will remove or recycle it. However, a nonzero
status tells <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> that the file was not archived;
it will try again periodically until it succeeds.
</P
><P
> The archive command should generally be designed to refuse to overwrite
any pre-existing archive file. This is an important safety feature to
preserve the integrity of your archive in case of administrator error
(such as sending the output of two different servers to the same archive
directory).
</P
><P
> It is advisable to test your proposed archive command to ensure that it
indeed does not overwrite an existing file, <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>and that it returns
nonzero status in this case</I
></SPAN
>.
The example command above for Unix ensures this by including a separate
<TT
CLASS="COMMAND"
>test</TT
> step. On some Unix platforms, <TT
CLASS="COMMAND"
>cp</TT
> has
switches such as <TT
CLASS="OPTION"
>-i</TT
> that can be used to do the same thing
less verbosely, but you should not rely on these without verifying that
the right exit status is returned. (In particular, GNU <TT
CLASS="COMMAND"
>cp</TT
>
will return status zero when <TT
CLASS="OPTION"
>-i</TT
> is used and the target file
already exists, which is <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
> the desired behavior.)
</P
><P
> While designing your archiving setup, consider what will happen if
the archive command fails repeatedly because some aspect requires
operator intervention or the archive runs out of space. For example, this
could occur if you write to tape without an autochanger; when the tape
fills, nothing further can be archived until the tape is swapped.
You should ensure that any error condition or request to a human operator
is reported appropriately so that the situation can be
resolved reasonably quickly. The <TT
CLASS="FILENAME"
>pg_xlog/</TT
> directory will
continue to fill with WAL segment files until the situation is resolved.
(If the file system containing <TT
CLASS="FILENAME"
>pg_xlog/</TT
> fills up,
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> will do a PANIC shutdown. No committed
transactions will be lost, but the database will remain offline until
you free some space.)
</P
><P
> The speed of the archiving command is unimportant as long as it can keep up
with the average rate at which your server generates WAL data. Normal
operation continues even if the archiving process falls a little behind.
If archiving falls significantly behind, this will increase the amount of
data that would be lost in the event of a disaster. It will also mean that
the <TT
CLASS="FILENAME"
>pg_xlog/</TT
> directory will contain large numbers of
not-yet-archived segment files, which could eventually exceed available
disk space. You are advised to monitor the archiving process to ensure that
it is working as you intend.
</P
><P
> In writing your archive command, you should assume that the file names to
be archived can be up to 64 characters long and can contain any
combination of ASCII letters, digits, and dots. It is not necessary to
preserve the original relative path (<TT
CLASS="LITERAL"
>%p</TT
>) but it is necessary to
preserve the file name (<TT
CLASS="LITERAL"
>%f</TT
>).
</P
><P
> Note that although WAL archiving will allow you to restore any
modifications made to the data in your <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> database,
it will not restore changes made to configuration files (that is,
<TT
CLASS="FILENAME"
>postgresql.conf</TT
>, <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> and
<TT
CLASS="FILENAME"
>pg_ident.conf</TT
>), since those are edited manually rather
than through SQL operations.
You might wish to keep the configuration files in a location that will
be backed up by your regular file system backup procedures. See
<A
HREF="runtime-config-file-locations.html"
>Section 18.2</A
> for how to relocate the
configuration files.
</P
><P
> The archive command is only invoked on completed WAL segments. Hence,
if your server generates only little WAL traffic (or has slack periods
where it does so), there could be a long delay between the completion
of a transaction and its safe recording in archive storage. To put
a limit on how old unarchived data can be, you can set
<A
HREF="runtime-config-wal.html#GUC-ARCHIVE-TIMEOUT"
>archive_timeout</A
> to force the server to switch
to a new WAL segment file at least that often. Note that archived
files that are archived early due to a forced switch are still the same
length as completely full files. It is therefore unwise to set a very
short <TT
CLASS="VARNAME"
>archive_timeout</TT
> — it will bloat your archive
storage. <TT
CLASS="VARNAME"
>archive_timeout</TT
> settings of a minute or so are
usually reasonable.
</P
><P
> Also, you can force a segment switch manually with
<CODE
CLASS="FUNCTION"
>pg_switch_xlog</CODE
> if you want to ensure that a
just-finished transaction is archived as soon as possible. Other utility
functions related to WAL management are listed in <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-BACKUP-TABLE"
>Table 9-61</A
>.
</P
><P
> When <TT
CLASS="VARNAME"
>wal_level</TT
> is <TT
CLASS="LITERAL"
>minimal</TT
> some SQL commands
are optimized to avoid WAL logging, as described in <A
HREF="populate.html#POPULATE-PITR"
>Section 14.4.7</A
>. If archiving or streaming replication were
turned on during execution of one of these statements, WAL would not
contain enough information for archive recovery. (Crash recovery is
unaffected.) For this reason, <TT
CLASS="VARNAME"
>wal_level</TT
> can only be changed at
server start. However, <TT
CLASS="VARNAME"
>archive_command</TT
> can be changed with a
configuration file reload. If you wish to temporarily stop archiving,
one way to do it is to set <TT
CLASS="VARNAME"
>archive_command</TT
> to the empty
string (<TT
CLASS="LITERAL"
>''</TT
>).
This will cause WAL files to accumulate in <TT
CLASS="FILENAME"
>pg_xlog/</TT
> until a
working <TT
CLASS="VARNAME"
>archive_command</TT
> is re-established.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-BASE-BACKUP"
>24.3.2. Making a Base Backup</A
></H2
><P
> The easiest way to perform a base backup is to use the
<A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
> tool. It can create
a base backup either as regular files or as a tar archive. If more
flexibility than <A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
> can provide is
required, you can also make a base backup using the low level API
(see <A
HREF="continuous-archiving.html#BACKUP-LOWLEVEL-BASE-BACKUP"
>Section 24.3.3</A
>).
</P
><P
> It is not necessary to be concerned about the amount of time it takes
to make a base backup. However, if you normally run the
server with <TT
CLASS="VARNAME"
>full_page_writes</TT
> disabled, you might notice a drop
in performance while the backup runs since <TT
CLASS="VARNAME"
>full_page_writes</TT
> is
effectively forced on during backup mode.
</P
><P
> To make use of the backup, you will need to keep all the WAL
segment files generated during and after the file system backup.
To aid you in doing this, the base backup process
creates a <I
CLASS="FIRSTTERM"
>backup history file</I
> that is immediately
stored into the WAL archive area. This file is named after the first
WAL segment file that you need for the file system backup.
For example, if the starting WAL file is
<TT
CLASS="LITERAL"
>0000000100001234000055CD</TT
> the backup history file will be
named something like
<TT
CLASS="LITERAL"
>0000000100001234000055CD.007C9330.backup</TT
>. (The second
part of the file name stands for an exact position within the WAL
file, and can ordinarily be ignored.) Once you have safely archived
the file system backup and the WAL segment files used during the
backup (as specified in the backup history file), all archived WAL
segments with names numerically less are no longer needed to recover
the file system backup and can be deleted. However, you should
consider keeping several backup sets to be absolutely certain that
you can recover your data.
</P
><P
> The backup history file is just a small text file. It contains the
label string you gave to <A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
>, as well as
the starting and ending times and WAL segments of the backup.
If you used the label to identify the associated dump file,
then the archived history file is enough to tell you which dump file to
restore.
</P
><P
> Since you have to keep around all the archived WAL files back to your
last base backup, the interval between base backups should usually be
chosen based on how much storage you want to expend on archived WAL
files. You should also consider how long you are prepared to spend
recovering, if recovery should be necessary — the system will have to
replay all those WAL segments, and that could take awhile if it has
been a long time since the last base backup.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-LOWLEVEL-BASE-BACKUP"
>24.3.3. Making a Base Backup Using the Low Level API</A
></H2
><P
> The procedure for making a base backup using the low level
APIs contains a few more steps than
the <A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
> method, but is relatively
simple. It is very important that these steps are executed in
sequence, and that the success of a step is verified before
proceeding to the next step.
<P
></P
></P><OL
TYPE="1"
><LI
><P
> Ensure that WAL archiving is enabled and working.
</P
></LI
><LI
><P
> Connect to the database as a superuser and issue the command:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT pg_start_backup('label');</PRE
><P>
where <TT
CLASS="LITERAL"
>label</TT
> is any string you want to use to uniquely
identify this backup operation. (One good practice is to use the
full path where you intend to put the backup dump file.)
<CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
> creates a <I
CLASS="FIRSTTERM"
>backup label</I
> file,
called <TT
CLASS="FILENAME"
>backup_label</TT
>, in the cluster directory with
information about your backup, including the start time and label
string. The file is critical to the integrity of the backup, should
you need to restore from it.
</P
><P
> It does not matter which database within the cluster you connect to to
issue this command. You can ignore the result returned by the function;
but if it reports an error, deal with that before proceeding.
</P
><P
> By default, <CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
> can take a long time to finish.
This is because it performs a checkpoint, and the I/O
required for the checkpoint will be spread out over a significant
period of time, by default half your inter-checkpoint interval
(see the configuration parameter
<A
HREF="runtime-config-wal.html#GUC-CHECKPOINT-COMPLETION-TARGET"
>checkpoint_completion_target</A
>). This is
usually what you want, because it minimizes the impact on query
processing. If you want to start the backup as soon as
possible, use:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT pg_start_backup('label', true);</PRE
><P>
This forces the checkpoint to be done as quickly as possible.
</P
></LI
><LI
><P
> Perform the backup, using any convenient file-system-backup tool
such as <SPAN
CLASS="APPLICATION"
>tar</SPAN
> or <SPAN
CLASS="APPLICATION"
>cpio</SPAN
> (not
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> or
<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
>). It is neither
necessary nor desirable to stop normal operation of the database
while you do this.
</P
></LI
><LI
><P
> Again connect to the database as a superuser, and issue the command:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT pg_stop_backup();</PRE
><P>
This terminates the backup mode and performs an automatic switch to
the next WAL segment. The reason for the switch is to arrange for
the last WAL segment file written during the backup interval to be
ready to archive.
</P
></LI
><LI
><P
> Once the WAL segment files active during the backup are archived, you are
done. The file identified by <CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
>'s result is
the last segment that is required to form a complete set of backup files.
If <TT
CLASS="VARNAME"
>archive_mode</TT
> is enabled,
<CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
> does not return until the last segment has
been archived.
Archiving of these files happens automatically since you have
already configured <TT
CLASS="VARNAME"
>archive_command</TT
>. In most cases this
happens quickly, but you are advised to monitor your archive
system to ensure there are no delays.
If the archive process has fallen behind
because of failures of the archive command, it will keep retrying
until the archive succeeds and the backup is complete.
If you wish to place a time limit on the execution of
<CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
>, set an appropriate
<TT
CLASS="VARNAME"
>statement_timeout</TT
> value.
</P
></LI
></OL
><P>
</P
><P
> Some file system backup tools emit warnings or errors
if the files they are trying to copy change while the copy proceeds.
When taking a base backup of an active database, this situation is normal
and not an error. However, you need to ensure that you can distinguish
complaints of this sort from real errors. For example, some versions
of <SPAN
CLASS="APPLICATION"
>rsync</SPAN
> return a separate exit code for
<SPAN
CLASS="QUOTE"
>"vanished source files"</SPAN
>, and you can write a driver script to
accept this exit code as a non-error case. Also, some versions of
GNU <SPAN
CLASS="APPLICATION"
>tar</SPAN
> return an error code indistinguishable from
a fatal error if a file was truncated while <SPAN
CLASS="APPLICATION"
>tar</SPAN
> was
copying it. Fortunately, GNU <SPAN
CLASS="APPLICATION"
>tar</SPAN
> versions 1.16 and
later exit with 1 if a file was changed during the backup,
and 2 for other errors. With GNU <SPAN
CLASS="APPLICATION"
>tar</SPAN
> version 1.23 and
later, you can use the warning options <TT
CLASS="LITERAL"
>--warning=no-file-changed
--warning=no-file-removed</TT
> to hide the related warning messages.
</P
><P
> Be certain that your backup dump includes all of the files under
the database cluster directory (e.g., <TT
CLASS="FILENAME"
>/usr/local/pgsql/data</TT
>).
If you are using tablespaces that do not reside underneath this directory,
be careful to include them as well (and be sure that your backup dump
archives symbolic links as links, otherwise the restore will corrupt
your tablespaces).
</P
><P
> You can, however, omit from the backup dump the files within the
cluster's <TT
CLASS="FILENAME"
>pg_xlog/</TT
> subdirectory. This
slight adjustment is worthwhile because it reduces the risk
of mistakes when restoring. This is easy to arrange if
<TT
CLASS="FILENAME"
>pg_xlog/</TT
> is a symbolic link pointing to someplace outside
the cluster directory, which is a common setup anyway for performance
reasons. You might also want to exclude <TT
CLASS="FILENAME"
>postmaster.pid</TT
>
and <TT
CLASS="FILENAME"
>postmaster.opts</TT
>, which record information
about the running <SPAN
CLASS="APPLICATION"
>postmaster</SPAN
>, not about the
<SPAN
CLASS="APPLICATION"
>postmaster</SPAN
> which will eventually use this backup.
(These files can confuse <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
>.)
</P
><P
> It's also worth noting that the <CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
> function
makes a file named <TT
CLASS="FILENAME"
>backup_label</TT
> in the database cluster
directory, which is removed by <CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
>.
This file will of course be archived as a part of your backup dump file.
The backup label file includes the label string you gave to
<CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
>, as well as the time at which
<CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
> was run, and the name of the starting WAL
file. In case of confusion it is therefore possible to look inside a
backup dump file and determine exactly which backup session the dump file
came from. However, this file is not merely for your information; its
presence and contents are critical to the proper operation of the system's
recovery process.
</P
><P
> It is also possible to make a backup dump while the server is
stopped. In this case, you obviously cannot use
<CODE
CLASS="FUNCTION"
>pg_start_backup</CODE
> or <CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
>, and
you will therefore be left to your own devices to keep track of which
backup dump is which and how far back the associated WAL files go.
It is generally better to follow the continuous archiving procedure above.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-PITR-RECOVERY"
>24.3.4. Recovering Using a Continuous Archive Backup</A
></H2
><P
> Okay, the worst has happened and you need to recover from your backup.
Here is the procedure:
<P
></P
></P><OL
TYPE="1"
><LI
><P
> Stop the server, if it's running.
</P
></LI
><LI
><P
> If you have the space to do so,
copy the whole cluster data directory and any tablespaces to a temporary
location in case you need them later. Note that this precaution will
require that you have enough free space on your system to hold two
copies of your existing database. If you do not have enough space,
you should at least save the contents of the cluster's <TT
CLASS="FILENAME"
>pg_xlog</TT
>
subdirectory, as it might contain logs which
were not archived before the system went down.
</P
></LI
><LI
><P
> Remove all existing files and subdirectories under the cluster data
directory and under the root directories of any tablespaces you are using.
</P
></LI
><LI
><P
> Restore the database files from your file system backup. Be sure that they
are restored with the right ownership (the database system user, not
<TT
CLASS="LITERAL"
>root</TT
>!) and with the right permissions. If you are using
tablespaces,
you should verify that the symbolic links in <TT
CLASS="FILENAME"
>pg_tblspc/</TT
>
were correctly restored.
</P
></LI
><LI
><P
> Remove any files present in <TT
CLASS="FILENAME"
>pg_xlog/</TT
>; these came from the
file system backup and are therefore probably obsolete rather than current.
If you didn't archive <TT
CLASS="FILENAME"
>pg_xlog/</TT
> at all, then recreate
it with proper permissions,
being careful to ensure that you re-establish it as a symbolic link
if you had it set up that way before.
</P
></LI
><LI
><P
> If you have unarchived WAL segment files that you saved in step 2,
copy them into <TT
CLASS="FILENAME"
>pg_xlog/</TT
>. (It is best to copy them,
not move them, so you still have the unmodified files if a
problem occurs and you have to start over.)
</P
></LI
><LI
><P
> Create a recovery command file <TT
CLASS="FILENAME"
>recovery.conf</TT
> in the cluster
data directory (see <A
HREF="recovery-config.html"
>Chapter 26</A
>). You might
also want to temporarily modify <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> to prevent
ordinary users from connecting until you are sure the recovery was successful.
</P
></LI
><LI
><P
> Start the server. The server will go into recovery mode and
proceed to read through the archived WAL files it needs. Should the
recovery be terminated because of an external error, the server can
simply be restarted and it will continue recovery. Upon completion
of the recovery process, the server will rename
<TT
CLASS="FILENAME"
>recovery.conf</TT
> to <TT
CLASS="FILENAME"
>recovery.done</TT
> (to prevent
accidentally re-entering recovery mode later) and then
commence normal database operations.
</P
></LI
><LI
><P
> Inspect the contents of the database to ensure you have recovered to
the desired state. If not, return to step 1. If all is well,
allow your users to connect by restoring <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> to normal.
</P
></LI
></OL
><P>
</P
><P
> The key part of all this is to set up a recovery configuration file that
describes how you want to recover and how far the recovery should
run. You can use <TT
CLASS="FILENAME"
>recovery.conf.sample</TT
> (normally
located in the installation's <TT
CLASS="FILENAME"
>share/</TT
> directory) as a
prototype. The one thing that you absolutely must specify in
<TT
CLASS="FILENAME"
>recovery.conf</TT
> is the <TT
CLASS="VARNAME"
>restore_command</TT
>,
which tells <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> how to retrieve archived
WAL file segments. Like the <TT
CLASS="VARNAME"
>archive_command</TT
>, this is
a shell command string. It can contain <TT
CLASS="LITERAL"
>%f</TT
>, which is
replaced by the name of the desired log file, and <TT
CLASS="LITERAL"
>%p</TT
>,
which is replaced by the path name to copy the log file to.
(The path name is relative to the current working directory,
i.e., the cluster's data directory.)
Write <TT
CLASS="LITERAL"
>%%</TT
> if you need to embed an actual <TT
CLASS="LITERAL"
>%</TT
>
character in the command. The simplest useful command is
something like:
</P><PRE
CLASS="PROGRAMLISTING"
>restore_command = 'cp /mnt/server/archivedir/%f %p'</PRE
><P>
which will copy previously archived WAL segments from the directory
<TT
CLASS="FILENAME"
>/mnt/server/archivedir</TT
>. Of course, you can use something
much more complicated, perhaps even a shell script that requests the
operator to mount an appropriate tape.
</P
><P
> It is important that the command return nonzero exit status on failure.
The command <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>will</I
></SPAN
> be called requesting files that are not present
in the archive; it must return nonzero when so asked. This is not an
error condition. Not all of the requested files will be WAL segment
files; you should also expect requests for files with a suffix of
<TT
CLASS="LITERAL"
>.backup</TT
> or <TT
CLASS="LITERAL"
>.history</TT
>. Also be aware that
the base name of the <TT
CLASS="LITERAL"
>%p</TT
> path will be different from
<TT
CLASS="LITERAL"
>%f</TT
>; do not expect them to be interchangeable.
</P
><P
> WAL segments that cannot be found in the archive will be sought in
<TT
CLASS="FILENAME"
>pg_xlog/</TT
>; this allows use of recent un-archived segments.
However, segments that are available from the archive will be used in
preference to files in <TT
CLASS="FILENAME"
>pg_xlog/</TT
>.
</P
><P
> Normally, recovery will proceed through all available WAL segments,
thereby restoring the database to the current point in time (or as
close as possible given the available WAL segments). Therefore, a normal
recovery will end with a <SPAN
CLASS="QUOTE"
>"file not found"</SPAN
> message, the exact text
of the error message depending upon your choice of
<TT
CLASS="VARNAME"
>restore_command</TT
>. You may also see an error message
at the start of recovery for a file named something like
<TT
CLASS="FILENAME"
>00000001.history</TT
>. This is also normal and does not
indicate a problem in simple recovery situations; see
<A
HREF="continuous-archiving.html#BACKUP-TIMELINES"
>Section 24.3.5</A
> for discussion.
</P
><P
> If you want to recover to some previous point in time (say, right before
the junior DBA dropped your main transaction table), just specify the
required stopping point in <TT
CLASS="FILENAME"
>recovery.conf</TT
>. You can specify
the stop point, known as the <SPAN
CLASS="QUOTE"
>"recovery target"</SPAN
>, either by
date/time, named restore point or by completion of a specific transaction
ID. As of this writing only the date/time and named restore point options
are very usable, since there are no tools to help you identify with any
accuracy which transaction ID to use.
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> The stop point must be after the ending time of the base backup, i.e.,
the end time of <CODE
CLASS="FUNCTION"
>pg_stop_backup</CODE
>. You cannot use a base backup
to recover to a time when that backup was in progress. (To
recover to such a time, you must go back to your previous base backup
and roll forward from there.)
</P
></BLOCKQUOTE
></DIV
><P
> If recovery finds corrupted WAL data, recovery will
halt at that point and the server will not start. In such a case the
recovery process could be re-run from the beginning, specifying a
<SPAN
CLASS="QUOTE"
>"recovery target"</SPAN
> before the point of corruption so that recovery
can complete normally.
If recovery fails for an external reason, such as a system crash or
if the WAL archive has become inaccessible, then the recovery can simply
be restarted and it will restart almost from where it failed.
Recovery restart works much like checkpointing in normal operation:
the server periodically forces all its state to disk, and then updates
the <TT
CLASS="FILENAME"
>pg_control</TT
> file to indicate that the already-processed
WAL data need not be scanned again.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-TIMELINES"
>24.3.5. Timelines</A
></H2
><P
> The ability to restore the database to a previous point in time creates
some complexities that are akin to science-fiction stories about time
travel and parallel universes. For example, in the original history of the database,
suppose you dropped a critical table at 5:15PM on Tuesday evening, but
didn't realize your mistake until Wednesday noon.
Unfazed, you get out your backup, restore to the point-in-time 5:14PM
Tuesday evening, and are up and running. In <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>this</I
></SPAN
> history of
the database universe, you never dropped the table. But suppose
you later realize this wasn't such a great idea, and would like
to return to sometime Wednesday morning in the original history.
You won't be able
to if, while your database was up-and-running, it overwrote some of the
WAL segment files that led up to the time you now wish you
could get back to. Thus, to avoid this, you need to distinguish the series of
WAL records generated after you've done a point-in-time recovery from
those that were generated in the original database history.
</P
><P
> To deal with this problem, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> has a notion
of <I
CLASS="FIRSTTERM"
>timelines</I
>. Whenever an archive recovery completes,
a new timeline is created to identify the series of WAL records
generated after that recovery. The timeline
ID number is part of WAL segment file names so a new timeline does
not overwrite the WAL data generated by previous timelines. It is
in fact possible to archive many different timelines. While that might
seem like a useless feature, it's often a lifesaver. Consider the
situation where you aren't quite sure what point-in-time to recover to,
and so have to do several point-in-time recoveries by trial and error
until you find the best place to branch off from the old history. Without
timelines this process would soon generate an unmanageable mess. With
timelines, you can recover to <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>any</I
></SPAN
> prior state, including
states in timeline branches that you abandoned earlier.
</P
><P
> Every time a new timeline is created, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> creates
a <SPAN
CLASS="QUOTE"
>"timeline history"</SPAN
> file that shows which timeline it branched
off from and when. These history files are necessary to allow the system
to pick the right WAL segment files when recovering from an archive that
contains multiple timelines. Therefore, they are archived into the WAL
archive area just like WAL segment files. The history files are just
small text files, so it's cheap and appropriate to keep them around
indefinitely (unlike the segment files which are large). You can, if
you like, add comments to a history file to record your own notes about
how and why this particular timeline was created. Such comments will be
especially valuable when you have a thicket of different timelines as
a result of experimentation.
</P
><P
> The default behavior of recovery is to recover along the same timeline
that was current when the base backup was taken. If you wish to recover
into some child timeline (that is, you want to return to some state that
was itself generated after a recovery attempt), you need to specify the
target timeline ID in <TT
CLASS="FILENAME"
>recovery.conf</TT
>. You cannot recover into
timelines that branched off earlier than the base backup.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="BACKUP-TIPS"
>24.3.6. Tips and Examples</A
></H2
><P
> Some tips for configuring continuous archiving are given here.
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="BACKUP-STANDALONE"
>24.3.6.1. Standalone Hot Backups</A
></H3
><P
> It is possible to use <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>'s backup facilities to
produce standalone hot backups. These are backups that cannot be used
for point-in-time recovery, yet are typically much faster to backup and
restore than <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> dumps. (They are also much larger
than <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> dumps, so in some cases the speed advantage
might be negated.)
</P
><P
> As with base backups, the easiest way to produce a standalone
hot backup is to use the <A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
>
tool. If you include the <TT
CLASS="LITERAL"
>-X</TT
> parameter when calling
it, all the transaction log required to use the backup will be
included in the backup automatically, and no special action is
required to restore the backup.
</P
><P
> If more flexibility in copying the backup files is needed, a lower
level process can be used for standalone hot backups as well.
To prepare for low level standalone hot backups, set <TT
CLASS="VARNAME"
>wal_level</TT
> to
<TT
CLASS="LITERAL"
>archive</TT
> (or <TT
CLASS="LITERAL"
>hot_standby</TT
>), <TT
CLASS="VARNAME"
>archive_mode</TT
> to
<TT
CLASS="LITERAL"
>on</TT
>, and set up an <TT
CLASS="VARNAME"
>archive_command</TT
> that performs
archiving only when a <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>switch file</I
></SPAN
> exists. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>archive_command = 'test ! -f /var/lib/pgsql/backup_in_progress || (test ! -f /var/lib/pgsql/archive/%f && cp %p /var/lib/pgsql/archive/%f)'</PRE
><P>
This command will perform archiving when
<TT
CLASS="FILENAME"
>/var/lib/pgsql/backup_in_progress</TT
> exists, and otherwise
silently return zero exit status (allowing <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
to recycle the unwanted WAL file).
</P
><P
> With this preparation, a backup can be taken using a script like the
following:
</P><PRE
CLASS="PROGRAMLISTING"
>touch /var/lib/pgsql/backup_in_progress
psql -c "select pg_start_backup('hot_backup');"
tar -cf /var/lib/pgsql/backup.tar /var/lib/pgsql/data/
psql -c "select pg_stop_backup();"
rm /var/lib/pgsql/backup_in_progress
tar -rf /var/lib/pgsql/backup.tar /var/lib/pgsql/archive/</PRE
><P>
The switch file <TT
CLASS="FILENAME"
>/var/lib/pgsql/backup_in_progress</TT
> is
created first, enabling archiving of completed WAL files to occur.
After the backup the switch file is removed. Archived WAL files are
then added to the backup so that both base backup and all required
WAL files are part of the same <SPAN
CLASS="APPLICATION"
>tar</SPAN
> file.
Please remember to add error handling to your backup scripts.
</P
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="COMPRESSED-ARCHIVE-LOGS"
>24.3.6.2. Compressed Archive Logs</A
></H3
><P
> If archive storage size is a concern, you can use
<SPAN
CLASS="APPLICATION"
>gzip</SPAN
> to compress the archive files:
</P><PRE
CLASS="PROGRAMLISTING"
>archive_command = 'gzip < %p > /var/lib/pgsql/archive/%f'</PRE
><P>
You will then need to use <SPAN
CLASS="APPLICATION"
>gunzip</SPAN
> during recovery:
</P><PRE
CLASS="PROGRAMLISTING"
>restore_command = 'gunzip < /mnt/server/archivedir/%f > %p'</PRE
><P>
</P
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="BACKUP-SCRIPTS"
>24.3.6.3. <TT
CLASS="VARNAME"
>archive_command</TT
> Scripts</A
></H3
><P
> Many people choose to use scripts to define their
<TT
CLASS="VARNAME"
>archive_command</TT
>, so that their
<TT
CLASS="FILENAME"
>postgresql.conf</TT
> entry looks very simple:
</P><PRE
CLASS="PROGRAMLISTING"
>archive_command = 'local_backup_script.sh "%p" "%f"'</PRE
><P>
Using a separate script file is advisable any time you want to use
more than a single command in the archiving process.
This allows all complexity to be managed within the script, which
can be written in a popular scripting language such as
<SPAN
CLASS="APPLICATION"
>bash</SPAN
> or <SPAN
CLASS="APPLICATION"
>perl</SPAN
>.
</P
><P
> Examples of requirements that might be solved within a script include:
<P
></P
></P><UL
><LI
><P
> Copying data to secure off-site data storage
</P
></LI
><LI
><P
> Batching WAL files so that they are transferred every three hours,
rather than one at a time
</P
></LI
><LI
><P
> Interfacing with other backup and recovery software
</P
></LI
><LI
><P
> Interfacing with monitoring software to report errors
</P
></LI
></UL
><P>
</P
><DIV
CLASS="TIP"
><BLOCKQUOTE
CLASS="TIP"
><P
><B
>Tip: </B
> When using an <TT
CLASS="VARNAME"
>archive_command</TT
> script, it's desirable
to enable <A
HREF="runtime-config-logging.html#GUC-LOGGING-COLLECTOR"
>logging_collector</A
>.
Any messages written to <SPAN
CLASS="SYSTEMITEM"
>stderr</SPAN
> from the script will then
appear in the database server log, allowing complex configurations to
be diagnosed easily if they fail.
</P
></BLOCKQUOTE
></DIV
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="CONTINUOUS-ARCHIVING-CAVEATS"
>24.3.7. Caveats</A
></H2
><P
> At this writing, there are several limitations of the continuous archiving
technique. These will probably be fixed in future releases:
<P
></P
></P><UL
><LI
><P
> Operations on hash indexes are not presently WAL-logged, so
replay will not update these indexes. This will mean that any new inserts
will be ignored by the index, updated rows will apparently disappear and
deleted rows will still retain pointers. In other words, if you modify a
table with a hash index on it then you will get incorrect query results
on a standby server. When recovery completes it is recommended that you
manually <A
HREF="sql-reindex.html"
>REINDEX</A
>
each such index after completing a recovery operation.
</P
></LI
><LI
><P
> If a <A
HREF="sql-createdatabase.html"
>CREATE DATABASE</A
>
command is executed while a base backup is being taken, and then
the template database that the <TT
CLASS="COMMAND"
>CREATE DATABASE</TT
> copied
is modified while the base backup is still in progress, it is
possible that recovery will cause those modifications to be
propagated into the created database as well. This is of course
undesirable. To avoid this risk, it is best not to modify any
template databases while taking a base backup.
</P
></LI
><LI
><P
> <A
HREF="sql-createtablespace.html"
>CREATE TABLESPACE</A
>
commands are WAL-logged with the literal absolute path, and will
therefore be replayed as tablespace creations with the same
absolute path. This might be undesirable if the log is being
replayed on a different machine. It can be dangerous even if the
log is being replayed on the same machine, but into a new data
directory: the replay will still overwrite the contents of the
original tablespace. To avoid potential gotchas of this sort,
the best practice is to take a new base backup after creating or
dropping tablespaces.
</P
></LI
></UL
><P>
</P
><P
> It should also be noted that the default <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>
format is fairly bulky since it includes many disk page snapshots.
These page snapshots are designed to support crash recovery, since
we might need to fix partially-written disk pages. Depending on
your system hardware and software, the risk of partial writes might
be small enough to ignore, in which case you can significantly
reduce the total volume of archived logs by turning off page
snapshots using the <A
HREF="runtime-config-wal.html#GUC-FULL-PAGE-WRITES"
>full_page_writes</A
>
parameter. (Read the notes and warnings in <A
HREF="wal.html"
>Chapter 29</A
>
before you do so.) Turning off page snapshots does not prevent
use of the logs for PITR operations. An area for future
development is to compress archived WAL data by removing
unnecessary page copies even when <TT
CLASS="VARNAME"
>full_page_writes</TT
> is
on. In the meantime, administrators might wish to reduce the number
of page snapshots included in WAL by increasing the checkpoint
interval parameters as much as feasible.
</P
></DIV
></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="backup-file.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="high-availability.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>File System Level Backup</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="backup.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>High Availability, Load Balancing, and Replication</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
