Backup

Introduction#

Backing up a remote Redis instance can be achieved with replication. This is useful if you want to take a snapshot of a dataset prior to upgrading, deleting or changing a Redis database.

Backup of a remote Redis instance to a local instance

On the machine where you’d like to make the backup, jump to the Redis CLI:

redis-cli

Password?

If your master Redis DB (the one you want to replicate) has a password:

config set masterauth <password>

Start replication

Run the following to begin replication:

SLAVEOF <host> <port>

To check the replication is underway run:

INFO replication

And you should see output like this:

# Replication
role:slave
master_host:some-host.compute-1.amazonaws.com
master_port:6519
master_link_status:up
master_last_io_seconds_ago:3
master_sync_in_progress:0
slave_repl_offset:35492914
slave_priority:100
slave_read_only:1
connected_slaves:0
master_repl_offset:0
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0

Note the master_link_status should be up.

Checking sync progress

When the sync is complete, the INFO replication should show:

master_sync_in_progress:0

To check the dataset has been synced you could compare the size of the database:

DBSIZE

Saving a data dump to disk

To save the DB to disk asynchronously:

BGSAVE
CONFIG GET dir

Then you should find a dump.rdb file in the directory listed by the config command.

Halting replication

You can stop replication with:

SLAVEOF NO ONE

Reference: Redis replication guide