db/pgbackrest
1
0
Fork 0
mirror of https://github.com/pgbackrest/pgbackrest.git synced 2025-01-18 04:58:51 +02:00
Code Issues Releases Activity

Backup file bundling documentation.

Browse Source 
Make the feature user visible and add documentation to the user guide.
This commit is contained in:
David Steele 2022-05-06 10:21:20 -04:00 committed by GitHub
parent 4cc0d46d60
commit f405fc6ae2
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 71 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
doc/xml/release.xml
View File

@ -45,6 +45,33 @@
</release-bug-list>
<release-feature-list>
<release-item>
<commit subject="Remove redundant restoreFile() test and improve coverage."/>
<commit subject="Add limit parameter to ioCopyP()."/>
<commit subject="Optimize restore command for file bundling.">
<github-issue id="1149"/>
<github-pull-request id="1683"/>
</commit>
<commit subject="Add bundle logging to backup command."/>
<commit subject="Rename bundle-* options to repo-bundle-*.">
<github-pull-request id="1694"/>
</commit>
<commit subject="Disable repo-hardlink option when repo-bundle option is enabled."/>
<commit subject="Allow files that become zero-length after the backup manifest is built."/>
<commit subject="Backup file bundling documentation.">
<github-pull-request id="1740"/>
</commit>
<release-item-contributor-list>
<release-item-contributor id="david.steele"/>
<release-item-reviewer id="reid.thompson"/>
<release-item-reviewer id="stefan.fercot"/>
<release-item-reviewer id="chris.bandy"/>
</release-item-contributor-list>
<p>Backup file bundling for improved small file support.</p>
</release-item>
<release-item>
<commit subject="Remove dependency on pg_database.datlastsysoid.">
<github-pull-request id="1735"/>
@ -224,30 +251,6 @@
</release-improvement-list>
<release-development-list>
<release-item>
<commit subject="Remove redundant restoreFile() test and improve coverage."/>
<commit subject="Add limit parameter to ioCopyP()."/>
<commit subject="Optimize restore command for file bundling.">
<github-issue id="1149"/>
<github-pull-request id="1683"/>
</commit>
<commit subject="Add bundle logging to backup command."/>
<commit subject="Rename bundle-* options to repo-bundle-*.">
<github-pull-request id="1694"/>
</commit>
<commit subject="Disable repo-hardlink option when repo-bundle option is enabled."/>
<commit subject="Allow files that become zero-length after the backup manifest is built."/>
<release-item-contributor-list>
<release-item-contributor id="david.steele"/>
<release-item-reviewer id="reid.thompson"/>
<release-item-reviewer id="stefan.fercot"/>
<release-item-reviewer id="chris.bandy"/>
</release-item-contributor-list>
<p>Improve small file support.</p>
</release-item>
<release-item>
<commit subject="Add test for protocol greeting when a field is missing."/>
<commit subject="Return stats as a JSON string rather than a KeyValue object."/>

44
doc/xml/user-guide.xml
View File

@ -1526,6 +1526,50 @@
</section>
</section>
<!-- ======================================================================================================================= -->
<section id="backup" depend="quickstart/perform-backup">
<title>Backup</title>
<cmd-description key="backup"/>
<!-- =================================================================================================================== -->
<section id="bundle">
<title>File Bundling</title>
<p>Bundling files together in the repository saves time during the backup and some space in the repository. This is especially pronounced when the repository is stored on an object store such as <proper>S3</proper>. Per-file creation time on object stores is higher and very small files might cost as much to store as larger files.</p>
<p>The file bundling feature is enabled with the <br-option>repo-bundle</br-option> option.</p>
<backrest-config host="{[host-pg1]}" file="{[backrest-config-demo]}">
<title>Configure <br-option>repo1-bundle</br-option></title>
<backrest-config-option section="global" key="repo1-bundle">y</backrest-config-option>
</backrest-config>
<p>A full backup without file bundling will have 1000+ files in the backup path, but with bundling the total number of files is greatly reduced. An additional benefit is that zero-length files are not stored (except in the manifest), whereas in a normal backup each zero-length file is stored individually.</p>
<execute-list host="{[host-pg1]}">
<title>Perform a full backup</title>
<execute user="postgres">
<exe-cmd>{[project-exe]} {[dash]}-stanza={[postgres-cluster-demo]} --type=full backup</exe-cmd>
</execute>
</execute-list>
<execute-list host="{[host-pg1]}">
<title>Check file total</title>
<execute user="postgres" output="y">
<exe-cmd>find {[backrest-repo-path]}/backup/demo/latest/ -type f | wc -l</exe-cmd>
</execute>
</execute-list>
<p>The <br-option>repo-bundle-size</br-option> and <br-option>repo-bundle-limit</br-option> options can be used for tuning, though the defaults should be optimal in most cases.</p>
<p>While file bundling is generally more efficient, the downside is that it is more difficult to manually retrieve files from the repository. It may not be ideal for deduplicated storage since each full backup will arrange files in the bundles differently. Lastly, file bundles cannot be resumed, so be careful not to set <br-option>repo-bundle-size</br-option> too high.</p>
</section>
</section>
<!-- SECTION => RETENTION -->
<section id="retention" depend="quickstart/perform-backup">
<title>Retention</title>

2
src/build/config/config.yaml
View File

@ -1742,7 +1742,6 @@ option:
depend: repo-azure-account
repo-bundle:
internal: true
section: global
group: repo
type: boolean
@ -1753,7 +1752,6 @@ option:
main: {}
repo-bundle-size:
internal: true
section: global
group: repo
type: size