From 82d25a99c4869b6ad3a4f6e1d365755058acf38e Mon Sep 17 00:00:00 2001 From: "Michael C. Harris" Date: Sun, 17 Mar 2013 16:58:10 +1100 Subject: [PATCH 1/2] DOCS-1245: Improve backup strategy overview. Both filesystem snapshots and mongodump can be point in time. Clean up the language. Leave the details that will help people make a decision as to their best options in the linked detail pages. Put the links in the same order we've talked about the options. --- source/administration/backups.txt | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) diff --git a/source/administration/backups.txt b/source/administration/backups.txt index 4c6c67c2e4c..03903264d48 100644 --- a/source/administration/backups.txt +++ b/source/administration/backups.txt @@ -60,30 +60,25 @@ Approaches to Backing Up MongoDB Systems ---------------------------------------- There are two main methodologies for backing up MongoDB -instances. Creating binary "dumps" of the database using -:program:`mongodump` or creating filesystem level snapshots. Both -methodologies have advantages and disadvantages: +instances; creating binary "dumps" of the database using +:program:`mongodump` or creating filesystem level snapshots. - binary database dumps are comparatively small, because they don't - include index content or pre-allocated free space, and :ref:`record - padding `. However, it's impossible to - capture a copy of a running system that reflects a single moment in - time using a binary dump. + include index content, pre-allocated free space, or :ref:`record + padding `, and can reflect a single + moment in time using the :option:`--oplog ` option. -- filesystem snapshots, sometimes called block level backups, produce +- filesystem snapshots, sometimes called block-level backups, produce larger backup sizes, but complete quickly and can reflect a single moment in time on a running system. However, snapshot systems require filesystem and operating system support and tools. The best option depends on the requirements of your deployment and -disaster recovery needs. Typically, filesystem snapshots are because -of their accuracy and simplicity; however, :program:`mongodump` is a -viable option used often to generate backups of MongoDB systems. +disaster recovery needs. The following topics provide details and +procedures on the two approaches: -The following topics provide details and procedures on the two approaches: - -- :doc:`/tutorial/backup-databases-with-filesystem-snapshots`. - :doc:`/tutorial/backup-databases-with-binary-database-dumps`. +- :doc:`/tutorial/backup-databases-with-filesystem-snapshots`. In some cases, taking backups is difficult or impossible because of large data volumes, distributed architectures, and data transmission From b062884829021d1cce13dd70978255f827ac2d75 Mon Sep 17 00:00:00 2001 From: "Michael C. Harris" Date: Mon, 18 Mar 2013 23:04:38 +1100 Subject: [PATCH 2/2] DOCS-1245: Separate point in time discussion. --- source/administration/backups.txt | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) diff --git a/source/administration/backups.txt b/source/administration/backups.txt index 03903264d48..cd1f5ddb655 100644 --- a/source/administration/backups.txt +++ b/source/administration/backups.txt @@ -65,13 +65,12 @@ instances; creating binary "dumps" of the database using - binary database dumps are comparatively small, because they don't include index content, pre-allocated free space, or :ref:`record - padding `, and can reflect a single - moment in time using the :option:`--oplog ` option. + padding `, but can take significant + time to complete. -- filesystem snapshots, sometimes called block-level backups, produce - larger backup sizes, but complete quickly and can reflect a single - moment in time on a running system. However, snapshot systems - require filesystem and operating system support and tools. +- filesystem snapshots, sometimes called block-level backups, complete + quickly, but produce larger backup sizes and require filesystem and + operating system support and tools. The best option depends on the requirements of your deployment and disaster recovery needs. The following topics provide details and @@ -85,6 +84,20 @@ large data volumes, distributed architectures, and data transmission speeds. In these situations, increase the number of members in your replica set or sets. +Another issue to be considered is how closely a backup can represent a +single moment in time. This is possible with both binary dumps +[#oplog]_ and filesystem snapshots if a single node contains the +complete data set. However, the situation becomes more complex when a +single node does not contain the complete data set, as with +:term:`sharded clusters ` + +.. [#oplog] By default, the :program:`mongodump` command does not + include operations after execution has started. However, the + :option:`--oplog ` option causes the :term:`oplog` + to be included in the dump, which can then be restored using the + :program:`mongorestore` :option:`--oplogReplay ` option. + Backup Strategies for MongoDB Deployments -----------------------------------------