NoSlides example application

Example application for the talk at NoSlidesConf. This application is a sample application with some functionalities available on riak_core.

Erlang, Elixir, Rebar3 version

On this branch I start to use asdf and as you can find here I compile everything with:

erlang 19.3
elixir 1.4.2

I did a lot of problem to compile this because elixir now use rebar3 v 3.5.0 or greater, and it issue some problem especially with cuttlefish or other packages that want create and escript. Using an old version of rebar, the 3.2.0 for example everything works fine ...

There are some works on riak_core and riak_core_ng to move on Erlang 19.X, but when I wrote this docs it's not ready.

Clone the repository as usually, download the dependencies and compile the app in this way:

mix deps.get
mix compile

Run the unit test that executes the riak_core ring onseveral slave nodes in a cluster. Nodes are configured in config/test.exs. To prevent running RiakCore on the master node (which is harmless, but clutters the logs), invoke with:

mix test --no-start

Comments on Erlang/Elixir compatibility

(pmenhart, 2018-06-05) Hack to make the project working with erlang 20.3.6 and elixir 1.6.5:

• After mix deps.get, I had to comment out "warnings_as_errors" in deps/riak_ensemble/rebar.config and in deps/riak_core/rebar.config.
• Note this project is using {:riak_core, "~> 3.0.9", hex: :riak_core_ng}. Later riak_core_ng commits introduced gen_fsm_compat, which fails with newer Elixir+rebar3 because of "missing erl_vsn" issue: rebar_erl_vsn is a pre-compile hook in rebar.config of gen_fsm_compat and few other projects. Error seems to be caused by Mix not handling rebar3 hooks properly. See e.g. elixir-lang/elixir#7733 and Kyorai/riak_core#23
• This issue is not specific to erl_vsn. For example, the forked https://github.com/gpad/cuttlefish (as used in mix.exs here) differs from the official version only by rebar.config commenting out:

% {provider_hooks, [{post, [{compile, {default, escriptize}}]}]}.

(pmenhart, 2018-07-11) Rebar3 was fixed, use version 3.6.1 or up. {:riak_core, "~> 3.1.1", hex: :riak_core_ng} compiles (with gen_fsm_compat) without problems mentioned above. However, the forked version of cuttlefish is still needed.

How to start a single node

If you want run a single node you can execute in this way:

$ iex --name gpad@127.0.0.1 -S mix run

You can substitute gpad with the name that you prefer. You have to always run riak_core with full name using --name parameter.

How to start multinode on same machine

In this repo it's possibile execute 3 nodes on the same machine. You can execute it in this way:

# this is node 1
MIX_ENV=gpad_1 iex --name gpad_1@127.0.0.1 -S mix run

# this is node 2
MIX_ENV=gpad_2 iex --name gpad_2@127.0.0.1 -S mix run

# this is node 3
MIX_ENV=gpad_3 iex --name gpad_3@127.0.0.1 -S mix run

If you want add more nodes you can check the file in config directory.

Join/Leave node

When you run all the nodes they are running alone. If you want join two node together and create cluster you could go, for example, on console on node 2 and execute this command:

iex(gpad_2@127.0.0.1)1>  :riak_core.join('gpad_1@127.0.0.1')

In this way node 2, called gpad_2 is joined on node 1, called gpad_1. Now this two node are a cluster.

The same command could be executed on console of node 3:

iex(gpad_3@127.0.0.1)1>  :riak_core.join('gpad_1@127.0.0.1')

When the cluster is stable you can try to put some data in it using the put command (see below) and then you can remove a node from the cluster with the leave command. Try it in this way:

# removing node 2 from cluster
iex(gpad_2@127.0.0.1)1>  :riak_core.leave

How the cluster is composed is saved in disk so, if you stop and restart one or all nodes, they try to connect together, if you want destroy the cluster you can remove evry single node with leave command or deleted the various ring_data_dir*.

Print ring status

After or before the join you can check the status of the ring. You can do this executing this command on elixir shell:

iex(gpad_1@127.0.0.1)1> {:ok, ring} = :riak_core_ring_manager.get_my_ring
iex(gpad_1@127.0.0.1)1> :riak_core_ring.pretty_print(ring, [:legend])

Functionalities

In this example it's now possible execute a ping command with default value of 1 but it's possible pass also a different value to have more chance to change the destination node.

iex(gpad_1@127.0.0.1)1> NoSlides.Service.ping
iex(gpad_1@127.0.0.1)1> NoSlides.Service.ping(2)
#execute a lot of different ping
(1..20) |> Enum.each(fn v -> NoSlides.Service.ping(v) end)

It's also possible use this application as a simple KV memory store, in this way:

# on node 1
iex(gpad_1@127.0.0.1)1> NoSlides.Service.put(:key, 42)

On node execute the command to store the value 42 associated with key :key and, it's possible get this value from another node, in this way:

# on node 2
iex(gpad_2@127.0.0.1)1> NoSlides.Service.get(:key)

Depending on your configuration and how many nodes you are running, you can see some different node that respond to this requests. Try to get value before and after a join and a subsequent leave of a node.

More infos...

I'm writing some post about the use of riak_core from Elixir, and you can find it here.

Credits

Thanks to:

• basho for creating the library and the docs.
• Heinz N. Gies for creating riak_core_ng that can be used in Elixir.
• Ben Tyler for the inspiration talk at ElixirConf.EU.
• Mariano Guerra for this amazing book about riak_core.
• Ryan Zezeski for his blog about riak_core.

License

This code is released under Apache 2 License. Check LICENSE file for more information.