<p>Frequently Asked Questions are intended to provide details for specific questions that may or may not be covered in the User Guide, Configuration, or Command reference. If you are unable to find details for your specific issue here, remember that the <backrest/><linkurl="https://github.com/pgbackrest/pgbackrest/issues">Issues List in GitHub</link> is also a valuable resource.</p>
</section>
<sectionid="timeout">
<title>What if I get the <quote>could not find WAL segment</quote> error?</title>
<p>The cause of this error can be a result of many different issues, some of which may be:</p>
<list-item>network or permissions issue</list-item>
<list-item>third party product (e.g. S3, Swift or Minio) configuration issue</list-item>
<list-item>large amount of WAL queueing to be archived</list-item>
</list>
<p>It is advisable to:</p>
<list>
<list-item>check the archive_command in <postgres/></list-item>
<list-item>check the <backrest/> configuration settings on each host (e.g. pg* settings are set on the repository host and repo* settings on the pg host)</list-item>
<list-item>run the <cmd>check</cmd> command with <br-setting>{[dash]}-archive-timeout</br-setting> set to a higher value than in the <backrest/> configuration file (or default) to see if the WAL queue needs more time to clear. If the system is generating a lot of WAL, then consider configuring <linkurl="https://pgbackrest.org/user-guide.html#async-archiving">asynchronous archiving</link></list-item>
<p>A full backup set can be expired using the <br-setting>{[dash]}-set</br-setting> option as explained in <linkurl="https://pgbackrest.org/command.html#command-expire">Command Reference: Expire</link>.</p>
<title>How can I configure options independently for each command?</title>
<p><backrest/> has the ability to set options independently in the configuration file for each command. <linkurl="https://pgbackrest.org/user-guide.html#quickstart/configure-stanza">Configure Cluster Stanza</link> details this feature as well as option precedence.</p>
<p>For example, the <br-option>process-max</br-option> option can be optimized for each command:</p>
<code-block>
[global]
# used where not overridden
process-max=2
[global:backup]
# more cores for backup
process-max=4
[global:restore]
# all the cores for restore
process-max=8
[global:archive-push]
# more cores for archive-push
process-max=3
[global:archive-get]
# fewer cores for archive-get
process-max=1
</code-block>
</section>
<sectionid="s3-bucket">
<title>Can I use dots (periods) in my S3 bucket name?</title>
<p><proper>RFC-2818</proper> does not allow wildcards to match on a dot (.) so s3 bucket names must not contain dots. If there are dots in the S3 bucket name then an error such as <quote>unable to find hostname 'my.backup.bucket.s3.amazonaws.com' in certificate common name or subject alternative names</quote> will occur.</p>
<title>Where can I find packages for older versions of <backrest/>?</title>
<p>The <linkurl="https://apt.postgresql.org">apt.postgresql.org</link> repository maintains an <linkurl="https://atalia.postgresql.org/morgue/p/pgbackrest">archive of older versions</link>. Debian also maintains <linkurl="https://snapshot.debian.org/binary/pgbackrest/">snapshots</link> of all test builds.</p>
<title>Why does a backup attempt fail when <br-option>backup-standby=y</br-option> and the standby database is down?</title>
<p>Configuring backup from standby is generally intended to reduce load on the primary, so switching backups to the primary when the standby is down often defeats the point. Putting more load on the primary in a situation where there are already failures in the system is not recommended. Backups are not critical as long as you have one that is fairly recent -- the important thing is to keep up with WAL archiving. There is plenty of time to get a backup when the system is stable again.</p>
<p>If you really need a backup, the solution is to have more standbys or remove <br-option>backup-standby</br-option>. This can be overridden on the command line with <br-option>--no-backup-standby</br-option>, so there is no need to reconfigure for a one-off backup.</p>
<title>Should I setup my repository on a standby host?</title>
<p>No. When primary and standby databases are configured, the <backrest/> configuration files should be symmetric in order to seamlessly handle failovers. If they are not, the configurations will need to be changed on failover or further problems may result.</p>
<p>See the <linkurl="user-guide.html#repo-host">Dedicated Repository Host</link> section of the <proper>User Guide</proper> for more information.</p>
<p>The most common mistake when using time-based Point-in-Time Recovery is forgetting to choose a backup set that is before the target time. <backrest/> will attempt to discover a backup to play forward from the time specified by the <setting>--target=</setting> if the <setting>--set</setting> option is not specified. If a backup set cannot be found, then restore will default to the latest backup. However, if the latest backup is after the target time, then <setting>--target=</setting> is not considered valid by <postgres/> and is therefore ignored, resulting in WAL recovery to the latest time available.</p>
<p>To use the <setting>--set</setting> option, choose a backup set by running the <cmd>info</cmd> command and finding the backup with a timestamp stop that is before the target time. Then when running the restore, specify the option <setting>--set=BACKUP_LABEL</setting> where <id>BACKUP_LABEL</id> is the chosen backup set.</p>
<p>It is often desirable to restore the latest backup from a production server to a development server. In principal, the instructions are the same as in <linkurl="https://pgbackrest.org/user-guide.html#replication/hot-standby">setting up a hot standby</link> with a few exceptions.</p>
<p>NEED TO ELABORATE HERE: Need an example of the restore command - what settings are different? Would they be {[dash]}-target, {[dash]}-target-action=promote, {[dash]}-type=immediate on the command-line? What about in the POSTGRES (e.g. hot_standby = on / wal_level = hot_standby - these would be different, no?) and PGBACKREST (e.g. would recovery-option=standby_mode=on still be set?) config files</p>
</section>
<sectionid="minio">
<title>Setting up Minio for pgBackRest</title>
<p>Setting up Minio for pgBackRest https://github.com/pgbackrest/pgbackrest/issues/645/</p>