Skip to content

AppsFlyer/aerospike-clj

Repository files navigation

Aerospike-clj

An opinionated Clojure library wrapping Aerospike Java Client.

Clojars Project

Build Status

Docs:

Generated docs

Tutorial

here.

More advanced docs

Requirements

  • Java 8
  • Clojure 1.8
  • Aerospike server version >= 4.9.0
  • Clojure version >= 1.11.0

Features

  • Converts Java client's callback model into Java(8) CompletableFuture based API.
  • Expose passing functional (asynchronous) transcoders over payloads (both put/get).
  • Health-check utility.
  • Functions return Clojure records.

Maturity

  • Feature completeness: mostly near complete.
  • Stability: production ready. Actively and widely used in production.

Opinionated

  • Non blocking only: Expose only the non-blocking API. Block with deref if you like.
  • Futures instead of callbacks. Futures (and functional chaining) are more composable and less cluttered. If synchronous behaviour is still desired, the calling code can still deref (@) the returned future object. For a more sophisticated coordination, a variety of control mechanisms can be used by directly using Java's CompletableFuture API or the more Clojure friendly promesa (which is also used internally), or via the library using transcoders or hooks.
  • Follows the method names of the underlying Java APIs.
  • TTLs should be explicit, and developers should think about them. Forces passing a TTL and not use the cluster default (This can be still achieved by passing the special values -2,-1 or 0).
  • Minimal dependencies.
  • Single client per Aerospike namespace. Namespaces in Aerospike usually indicate different cluster configurations. In order to reduce overhead for clusters with more than a single namespace create 2 client instances and share an event loop between them.

Usage

user=> (require '[aerospike-clj.client :as aero])
nil
user=> (def c (aero/init-simple-aerospike-client
  #_=>          ["aerospike-001.com", "aerospik-002.com"] "my-ns" {:enable-logging true}))

It is possible to inject additional asynchronous user-defined behaviour. To do that add an implementation of the ClientEvents protocol during client initialization or per operation.
Some useful info is passed in-order to support metering and to read client configuration. op-start-time is (System/nanoTime). see more here.

(let [c (aero/init-simple-aerospike-client
          ["localhost"]
          "test"
          {:client-events (reify ClientEvents
                            (on-success [_ op-name op-result index op-start-time]
                                (println op-name "success!")))
                            (on-failure [_  op-name op-ex index op-start-time]
                                (println "oh-no" op-name "failed on index" index)))})]

  (get-single c "index" "set-name"))
; for better performance, a `deftype` might be preferred over `reify`, if possible.

Query/Put

For demo purposes we will use a docker based local DB:

$ sudo docker run -d --name aerospike -p 3000:3000 -p 3001:3001 -p 3002:3002 -p 3003:3003 aerospike

And connect to it:

user=> (def c (aero/init-simple-aerospike-client ["localhost"] "test"))
#'user/db
user=> (require '[promesa.core :as p])
nil
user=> (aero/put c "index" "set-name" 42 1000)
#object[java.util.concurrent.CompletableFuture 0x6264b083 "pending"]
user=> (def f (aero/get-single c "index" "set-name"))
#'user/f
user=> (p/chain (aero/get-single c "index" "set-name")
  #_=>          :ttl
  #_=>          aero/expiry-unix
  #_=>          #(java.time.Instant/ofEpochSecond %)
  #_=>          str
  #_=>          println)
2020-08-13T09:52:49Z
#object[java.util.concurrent.CompletableFuture 0x654830f5 "pending"]

We actually get back a record with the payload, the DB generation and the TTL (in an Aerospike style EPOCH format).

user=> @(aero/get-single c "index" "set-name")
#aerospike_clj.client.AerospikeRecord{:payload 42, :gen 1, :ttl 285167713}

Unix EPOCH TTL

Aerospike returns a TTL on the queried records that is epoch style, but with a different "beginning of time" which is "2010-01-01T00:00:00Z". Call expiry-unix with the returned TTL to get a TTL relative to the UNIX epoch.

Testing

Unit tests

Executed via running lein test.

Integration tests

Testing is performed against a local Aerospike docker container.

Mocking in application unit tests

For unit tests purposes you can use a mock client that implements the client protocols: MockClient.

Usage:

(ns com-example.app 
  (:require [clojure.test :refer [deftest use-fixtures]]
            [aerospike-clj.protocols :as pt]
            [aerospike-clj.mock-client :as mock])
  (:import [aerospike_clj.client SimpleAerospikeClient]))

(def ^:dynamic ^SimpleAerospikeClient client nil)

(defn- bind-client-to-mock [test-fn]
  (binding [client (mock/create-instance)]
    (test-fn)))

(use-fixtures :each bind-client-to-mock)

(deftest ...) ;; define your application unit tests as usual

The sample code executes on every test run. It initializes the mock with a proper type hint so you can just invoke all client protocol methods on it.

Note: If the production client is initiated using a state management framework, you would also need to stop and restart the state on each test run.

Contributing

PRs are welcome with these rules:

  • A PR should increment the project's version in project.clj according to Semantic Versioning.
  • A PR should have its above version set to SNAPSHOT, e.g. 1.0.2-SNAPSHOT. Once it will be merged into master this suffix would be trimmed before release.
  • All PRs would be linted and tested. Passing lint and tests is a reuirement for maintainers to review the PR.

License

Distributed under the Apache 2.0 License - found here.