- Back Up and Restore Deployments >
- Restore MongoDB Deployments >
- Restore a Sharded Cluster from a Backup
Restore a Sharded Cluster from a Backup¶
On this page
Overview¶
When you restore a cluster from a backup, Ops Manager provides you with a restore files for the selected restore point. For an overview of the restore process, please see Restore Overview.
Considerations¶
The BSON specification changed the
default subtype for the BSON binary datatype (BinData
) from 2
to 0
. Some binary data stored in a snapshot may be BinData
subtype 2. The Backup Agent automatically detects and converts snapshot
data in BinData
subtype 2 to BinData
subtype 0. If your
application code expects BinData
subtype 2, you must update your
application code to work with BinData
subtype 0.
See also
The notes on the BSON specification explain the particular specifics of this change.
The backup restore file now includes a metadata file, restoreInfo.txt
. This file captures the options the database used when the snapshot was taken. The database must be run with the listed options after it has been restored.
See also
The Considerations of
Seed a New Secondary from Snapshot
explains the use and contents of the restoreInfo.txt
metadata file.
Prerequisites¶
Client Requests During Restoration¶
You must ensure that the MongoDB deployment does not receive client requests during restoration. You must either:
- Restore to new systems with new hostnames and reconfigure your application code once the new deployment is running, or
- Ensure that the MongoDB deployment will not receive client requests while you restore data.
Snapshots when Agent Cannot Stop Balancer¶
Ops Manager displays a warning next to cluster snapshots taken while the balancer is enabled. If you restore from such a snapshot, you run the risk of lost or orphaned data. For more information, see Snapshots when Agent Cannot Stop Balancer.
Secure Copy (SCP
) Delivery¶
Important
SCP
requires you to
generate a key pair
before attempting to transfer files. SCP
provides faster file
delivery than HTTP
.
Note
Microsoft Windows computers do not include SCP
and require
additional setup outside the scope of this manual.
Important
When copying files individually, the MaxStartups
value in
sshd_config
should be increased to:
(4 × (number of shards + number of config servers)) + 10
SCP
is performed in parallel and, by default, Secure Shell Daemon
(sshd
) installations use a small number of concurrent connections.
Changing this setting in sshd_config
allows SCP
to support
sufficient connections to complete the restore.
Example
For a sharded cluster with 7 shards and 3 config servers, change
MaxStartups
to 50
:
MaxStartups 50
Automatic Restore¶
To have Ops Manager automatically restore the backup, perform the select the snapshot procedure.
Manual Restore¶
To restore the backup manually, perform the following:
- Select and Retrieve the Snapshot.
- Restore Primary Replica Set Member for Each Shard.
- Restore All Secondary Replica Set Members for Each Shard.
- Restore Each Config Server.
- Start the MongoDB Sharded Cluster.
Select and Retrieve the Snapshot¶
Select the Backup tab and then Overview page.¶
Click the deployment’s ellipsis icon and select Restore.¶
Select the restore point.¶
Choose the point from which you want to restore your backup.
Restore Type | Description | Action |
---|---|---|
Snapshot | Allows you to choose one stored snapshot. | Select an existing snapshot to restore. |
Point In Time | Creates a custom snapshot that includes all operations up to but not including the selected time. Example If you select |
Select a Date and Time. |
Click Next.
Select whether to have Ops Manager restore the snapshot.¶
Yes | Ops Manager restores the snapshot for you. If you select this option, Ops Manager prompts to choose the cluster to which to restore. You can select an existing cluster or create a new one. Follow the prompts in the wizard. Important You can skip the remainder of this page. |
No, I’ll do it myself | Ops Manager provides you with the restore files for you to perform the restore manually. |
Choose how to receive the restore files.¶
Select the restore method, format and destination.
Pull Via Secure HTTP |
Important You can skip the remainder of this procedure. |
Push Via Secure Copy | Direct Ops Manager to copy the restore files to your server via
Important
Note Microsoft Windows computers do not include |
Format | Select the format in which you want to receive the restore files:
|
SCP Host | Type the hostname of the server to receive the files. |
SCP Port | Type the port of the server to receive the files. |
SCP User | Type the username used to access to the server. |
Auth Method | Select whether to use a username and password or an SSH certificate to authenticate to the server. |
Password | Type the user password used to access to the server. |
Passphrase | Type the SSH passphrase used to access to the server. |
Target Directory | Type the absolute path to the directory on the server to which to copy the restore files. |
Click Finalize Request.
Retrieve the snapshot.¶
- If you selected Pull Via Secure HTTP:
Ops Manager creates links to the snapshot. By default, these links are available for an hour and can be used just once. To download the snapshot:
- Click Backup, then the Restore History tab.
- When the restore job completes, a download link appears for each shard and for one of the config servers.
- Click each link to download the restore archives.
- For the config server, copy its config restore archive to each of its replica set members.
- For each shard, copy its restore archive to each of its replica set members.
- Extract the shard restore archive to the shard’s working database path.
- If you selected Push Via Secure Copy:
Ops Manager copies the files to the server directory you specified. After the files transfer, you should:
- Verify that the files are complete, see how to validate a secure copy restore.
- Copy the config server restore archive to each of its replica set members.
- Copy each shard restore archive to each shard’s replica set members.
- Extract the shard restore archive to the shard’s working database path.
Restore Primary Replica Set Member for Each Shard¶
For all shards, restore the primary replica set member. You must have a copy of the snapshot on the server that host the primary replica set member.
Shut down the entire replica set.¶
Shut down the replica set’s mongod processes using one of the following methods, depending on your configuration:
Automated Deployment¶
If you use Ops Manager Automation to manage the replica set, you must shut down through the Ops Manager console. See Shut Down a MongoDB Process.
Non-Automated Deployment on MongoDB 2.6 or Later¶
Connect to each member of the set and issue the following:
Non-Automated Deployment on MongoDB 2.4 or earlier¶
Connect to each member of the set and issue the following:
Restore the snapshot data files to the primary.¶
Extract the data files to the location where the mongod
process will access them through the dbpath
setting. If you
are restoring to existing hardware, use a different data directory than
used previously. The following are example commands:
Start the primary with the new dbpath
.¶
For example:
Connect to the primary and initiate the replica set.¶
Connect using the mongo
shell and issue the rs.initiate()
command.
Restart the primary as a standalone, without the --replSet
option.¶
Shut down the process using one of the following methods:
Automated Deployment¶
Shut down through the Ops Manager console. See Shut Down a MongoDB Process.
Non-Automated Deployment on MongoDB 2.6 or Later¶
Non-Automated Deployment on MongoDB 2.4 or earlier¶
Restart the process as a standalone:
Run the seedSecondary
script on the primary.¶
Use the appropriate script for your operating system:
UNIX-based | seedSecondary.sh |
Windows | seedSecondary.bat |
This script recreates the oplog collection and seeds it with the timestamp of the snapshot’s creation. This tells the secondary to recreate all operations to the current time without requiring a full initial sync. Ops Manager customizes this script for this particular snapshot and includes it in the snapshot archive.
To run the script, issue the following command, where:
<alternatePort> |
The port of the mongod process |
<oplogSizeInGigabytes> |
The size of the replica set’s oplog |
<replicaSetName> |
The name of the replica set |
<primaryHost:primaryPort> |
The hostname:port combination for the replica set’s primary |
For UNIX-based systems:¶
For Windows-based systems:¶
Restart the primary as part of a replica set.¶
Use the following sequence:
Shut down the process using one of the following methods:
Automated Deployment:
Shut down through the Ops Manager console. See Shut Down a MongoDB Process.
Non-Automated Deployment on MongoDB 2.6 or Later:
Non-Automated Deployment on MongoDB 2.4 or earlier:
Restart the process as part of a replica set:
Restore All Secondary Replica Set Members for Each Shard¶
After you have restored the primary for a shard you can restore all secondary replica set members. You must have a copy of the snapshot on all servers that host a secondary replica set member.
Connect to the server where you will create the new secondary.¶
Restore the snapshot data files to the secondary.¶
Extract the data files to the location where the mongod
process will access them through the dbpath
setting. If you
are restoring to existing hardware, use a different data directory than
used previously. The following are example commands:
Start the secondary as a standalone, without the --replSet
option.¶
Shut down the process using one of the following methods:
Automated Deployment¶
Shut down through the Ops Manager console. See Shut Down a MongoDB Process.
Non-Automated Deployment on MongoDB 2.6 or Later¶
Non-Automated Deployment on MongoDB 2.4 or earlier¶
Restart the process as a standalone:
Run the seedSecondary.sh
script on the secondary.¶
Use the appropriate script for your operating system:
UNIX-based | seedSecondary.sh |
Windows | seedSecondary.bat |
This script recreates the oplog collection and seeds it with the timestamp of the snapshot’s creation. This tells the secondary to recreate all operations to the current time without requiring a full initial sync. Ops Manager customizes this script for this particular snapshot and includes it in the snapshot archive.
To run the script, issue the following command, where:
<alternatePort> |
The port of the mongod process |
<oplogSizeInGigabytes> |
The size of the replica set’s oplog |
<replicaSetName> |
The name of the replica set |
<primaryHost:primaryPort> |
The hostname:port combination for the replica set’s primary |
For UNIX-based systems:¶
For Windows-based systems:¶
Restart the secondary as part of the replica set.¶
Use the following sequence:
Shut down the process using one of the following methods:
Automated Deployment:
Shut down through the Ops Manager console. See Shut Down a MongoDB Process.
Non-Automated Deployment on MongoDB 2.6 or Later:
Non-Automated Deployment on MongoDB 2.4 or earlier:
Restart the process as part of a replica set:
Connect to the primary and add the secondary to the replica set.¶
Connect to the primary and use rs.add()
to add the secondary
to the replica set.
Repeat this operation for each member of the set.
Restore Each Config Server¶
Perform this procedure separately for each config server. Each config server must have a copy of the tar file with the config server data.
Use this procedure only for manual restores.
Restore the snapshot to the config server.¶
Extract the data files to the location where the config server’s
mongod process will access them. This is the location you
will specify as the storage.dbPath
when running mongod for
the config server.
Start the config server.¶
The following example starts the config server using the new data:
Update the sharded cluster metadata.¶
If the new shards do not have the same hostnames and ports as the original cluster, you must update the shard metadata. To do this, connect to each config server and update the data.
First connect to the config server with the mongo shell. For example:
Then access the shards
collection in the config database. For example:
The find()
method returns the documents in the shards
collection. The collection contains a document for each shard in the
cluster. The host
field for a shard displays the name of the
shard’s replica set and then the hostname and port of the shard. For
example:
To change a shard’s hostname and port, use the MongoDB
update()
command to modify the
documents in the shards
collection.
Start the mongos
process¶
Start the cluster’s mongos bound to your new config servers.