The configuration of a Riak cluster requires configuring a node to listen on a non-local interface (i.e., not 127.0.0.1), and then joining nodes together.

Begin by editing your etc/app.config file. The app.config file will be located in your rel/riak/etc/ directory if you compiled from source, and /etc/riak/ if you used a binary install of Riak.

The commands below assume you are running from a source install, but if you have installed Riak with one of our binary installs you can substitute the usage of bin/riak with sudo /usr/sbin/riak and bin/riak-admin with sudo /usr/sbin/riak-admin.

Configure the First Node

First, stop your Riak node if it is currently running:

$ bin/riak stop

Change the default IP address located under http{} in the riak_core section of etc/app.config . (The port, 8098, does not need to be changed.) Let’s say our machine’s IP address is 192.168.1.10. (This is just an example IP address. Yours will be specific to your machine.)

{http, [ {"127.0.0.1", 8098 } ]},

becomes

{http, [ {"192.168.1.10", 8098 } ]},

Next edit the etc/vm.args file and change the -name to your new IP:

-name riak@127.0.0.1

becomes

-name riak@192.168.1.10

Start the Riak node:

$ bin/riak start

If you have started this Riak node before, you will need to “reip” the system so the node can take appropriate measures to update its ring file.

$ bin/riak-admin reip riak@127.0.0.1 riak@192.168.1.10

This node is now configured and ready to have a node join it in the Riak cluster.

Add a Second Node to Your Cluster

Repeat the steps above for another other host on the network. Once the node has started you will use the bin/riak-admin command to have it join the other node in the Riak cluster.

$ bin/riak-admin join riak@192.168.1.10
#Sent join request to riak@192.168.1.10

Your second Riak node is now part of the cluster and has begun syncing with your first node. There are two ways to see if your Riak cluster has been correctly configured.
Using riak-admin:

$ bin/riak-admin status | grep ring_members
#ring_members : ['riak@192.168.1.10','riak@192.168.1.11']

Using riak attach:

$ riak attach
1> {ok, R} = riak_core_ring_manager:get_my_ring().
{ok,{chstate,'riak@192.168.1.10',.........
(riak@192.168.52.129)2> riak_core_ring:all_members(R).
['riak@192.168.1.10','riak@192.168.1.11']

You can repeat the above steps to add more nodes to your cluster.

Failed: riak@10.0.1.156 has a different ring_creation_size

Running Multiple Nodes on One Host

If you have built Riak from source code, or if you are using the Mac OSX pre-built package, then you can easily run multiple Riak nodes on the same machine. The most common scenario for doing this is to experiment with running a Riak cluster. (Note: if you have installed the .deb or .rpm package, then you will need to download and build Riak source to follow the directions below.)

To run multiple nodes, make copies of the riak directory.

• If you have run make all rel, then this can be found in ./rel/riak under the root Riak source directory.
• If you are running Mac OSX, then this is the directory where you unzipped the .tar.gz file.

Now, assuming that you made copied ./rel/riak into ./rel/riak1, ./rel/riak2, ./rel/riak3, etc.:

• In riakN/etc/app.config, change handoff_port, pb_port, and the port specified in the http{} section to unique ports for each node.
• In riakN/etc/vm.args, change the line that says -name riak127.0.0.1@ to a unique name for each node, for example: -name riak1127.0.0.1@.

Now, start the nodes, changing path names and nodes as appropriate:

./rel/riak1/bin/riak start
./rel/riak2/bin/riak start
./rel/riak3/bin/riak start
(etc.)

Next, join the nodes into a cluster:

./rel/riak2/bin/riak-admin join riak1@127.0.0.1
./rel/riak3/bin/riak-admin join riak1@127.0.0.1
(etc.)

Alternatively, you can run make all devrel, which will create three copies of Riak under the directories ./dev/dev1, ./dev/dev2, and ./dev/dev3, with the configuration pre-set to allow you to run the nodes simultaneously. (Note: The Web port for the three nodes is 8091, 8092, and 8093, which is different from the default of 8098 that you get when you run “make all rel”.)

You can start each of them and join them together using the commands shown above, substituting directory names as appropriate.

Multiple Clusters on One Host

Using the above technique it is possible to run multiple clusters on one node. If a node hasn’t joined an existing cluster it will behave just as a cluster would. Running multiple clusters on one node is simply a matter of having two or more distinct nodes or groups of nodes joined together.