This is the client that we use at Voxer. It's been tested very thoroughly in our environment and for our usage patterns, but there may be glaring holes in functionality that we don't use.
It uses the poolee module to do load balancing and failure detection so you don't necessarily need a load balancer between node and Riak.
The library has support for timing each request, resolving siblings, and conditional retry.
var RiakClient = require("riak"), client;
client = new RiakClient(["host1:port1", "host2:port2", ... "hostn:portn"], "client_id", "pool_name");
The RiakClient constructor takes a list of host:port pairs, each of which are equivalent nodes in a Riak cluster.
Try running the example and you should see something like this:
ranney-imac:~/work/node_riak (master)$ curl -X DELETE localhost:8098/riak/bucket_1/key_1
ranney-imac:~/work/node_riak (master)$ node example.js
metric: counter, riak_retry_filter|404_GET=1
metric: histogram, LB_Pool_pool_name|GET|bucket_1=51
metric: histogram, LB_Pool_pool_name|PUT|bucket_1=7
204: { counter: 1 }
ranney-imac:~/work/node_riak (master)$ node example.js
metric: histogram, LB_Pool_pool_name|GET|bucket_1=9
metric: histogram, LB_Pool_pool_name|PUT|bucket_1=5
204: { counter: 2 }
ranney-imac:~/work/node_riak (master)$ node example.js
metric: histogram, LB_Pool_pool_name|GET|bucket_1=9
metric: histogram, LB_Pool_pool_name|PUT|bucket_1=4
204: { counter: 3 }
var RiakClient = require("riak");
// list of riak servers you'd like to load balance over (poolee handles this).
var servers = ["127.0.0.1:8098"]
// should be unique, used by riak if you don't supply a vector clock
// default value: random integer
var client_id = "docs-client"
// informative name for logging purposes etc
var pool_name = "docs-pool"
var client = new RiakClient(servers, client_id, pool_name);From here on we will refer to client as an instance of RiakClient.
// shows an activity trace.
client.debug_mode = false;TODO: more info on how this is collected and what it means, and what types of metrics are collected.
client.on("metrics", function (type, key, val) {
// `type` is either "histogram" or "counter".
// Information is gleaned from requests made by poolee, and tells you about
// downed nodes, retries, and request duration.
console.log("metric: " + type + ", " + key + "=" + val);
});bucket: which bucket you want to look in, "/riak/" is automatically prepended
for you (all the functions do this).
key: the name of the thing you want
callback: mandatory callback function, invoked ala callback(error, response, object)
options: let caller specify http headers that riak may care about such as
X-Riak-Vclock and Content-Type.
When options.return_body = true, it which will return the body and a status
code of 200 instead of 204.
Content-Type is assumed to be application/json by default and will be stored as stringified json unless otherwise specified.
The default options are
var options = {
http_headers: {},
mime_types: [], // list of content-types. TODO: never referenced in the code?
return_body: false,
// more supported options:
// r_val: <number>, // default is whatever Riak's default is, usually basic quorum
// w_val: <number>, // default is whatever Riak's default is, usually basic quorum
// retry: <bool>, // default = true, will retry gets with exponential backoff when recieving a 404
// parse: <bool>, // default = true, will parse riak response assuming it is json
// resolver: <fn>, // no default = used to resolve sibling values
}
client.get(bucket, key, options, callback)Useful for inserting something when you don't care what the key is. Riak
responds with a location header that tells you the key that was generated for
you:
var value = { riak: "is fun" }
bucket = "bucket_1",
url = "/riak/" + bucket,
options = {};
client.post(url, JSON.stringify(value), function(err, res, obj) {
assert.ok(!err);
assert.ok(res.headers.location !== null);
var uri = res.headers.location.split("/"),
key = uri[uri.length - 1];
client.get(bucket, key, options, function(err, res, obj) {
assert.deepEqual(obj, value);
});
});This mutator function does a simple increment of prop2 with no error checking. It does a GET and then a PUT with your modifications.
client.modify("bucket_1", "key_1", function mutator(old, done) {
var newobj = old || {};
newobj.counter = newobj.counter || 0;
newobj.counter++;
done(newobj);
}, {}, function (err, res, obj) {
console.error(res.statusCode + ": ", obj);
process.exit();
});The value stored at key must be an array. Attempts to append new_val to the
end of the array. If new_val is already in the array, it will not be added
again.
Internally it uses client.modify with a mutator function to accomplish this.
Make a request to riak with a url of the form
/solr/<bucket>/select?q=<query>&wt=json&rows=<limit>
Copyright (c) 2012 Matthew Ranney, http://ranney.com/
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.