TNS
VOXPOP
You’re most productive when…
A recent TNS post discussed the factors that make developers productive. You code best when:
The work is interesting to me.
0%
I get lots of uninterrupted work time.
0%
I am well-supported by a good toolset.
0%
I understand the entire code base.
0%
All of the above.
0%
I am equally productive all the time.
0%
Cloud Native Ecosystem

Sharded Deployments with MongoDB and Brooklyn, a Framework for Scaling

Apr 28th, 2015 6:33am by
Featued image for: Sharded Deployments with MongoDB and Brooklyn, a Framework for Scaling
This post originally appeared on the Cloudsoft blog to demonstrate how to create services with the Apache Brooklyn Plugin for Cloud Foundry. Apache Brooklyn is a platform for integrating services across multiple data centers. It is the fourth post in a series of that covers how Brooklyn works with a wide range of services. Part one explored how the Cloud Foundry service broker works with Brookyn. Part two in the series covered how a user can deploy the Brooklyn plugin to integrate Cloud Foundry. Part three demonstrated how Apache Ambari’s deployment integration technology fits with Apache Brooklyn.

The nature of distributed architecture requires the user to think through how apps and services may run across multiple cloud services and data centers. Apps are one thing, but running a database in these types of environments carries a different level of complexity.

In this example, we are going to create a more complicated service involving MongoDB and sharding. We will then show you how you can use effectors to scale this service out or in.

MongoDB offers some pretty advanced options for scaling, both horizontally and vertically. So, we’re going to look at how to create a sharded deployment (like that described in the figure above) with the Brooklyn plugin.

Create a Sharded Deployment Service

First, taking advantage of our new ability to specify blueprints in the application manifest, under the Brooklyn section, create a MongoDBShardedDeployment service:


Notice with the blueprint, you can also specify provisioning properties and configure the sharded deployment. We have specified a minimum of 16GB of RAM with one router, five shards and three nodes per shard in this blueprint. Simply push this blueprint …


and you have a sharded deployment of MongoDB.

Examine the Sensors

As we saw in part two, you can view the sensors associated with a running service with the sensors command, so let’s do that with this service:


In this example, there are lots of entities and many more sensors, but a few are worth pointing out: service.state shows whether the entity is running; host.address and mongodb.server.port show how to connect to the individual MongoDB servers, should you need to; and mongodb.server.network.bytesIn and mongodb.server.network.bytesOut show how much traffic is coming in and out.

Scale it Back Using Effectors

So, looking at the sensors, suppose we decide we would like to scale back the number of shards to one. To find out which effector to use for that, we use the effectors command:


The one we need is cfgFazAk:resize so let’s invoke that:

Scale it Back Out

And if we need to scale back out again, we simply run the same command again but with the parameter changed:


What about increased replication? Just look at the effectors to find the cluster you want to increase and run the same command:

Video

To wrap up this post, once again, here is a video of our plugin in action:

Next Time

In previous posts, we saw the basics of using the Cloud Foundry Service Broker and Plugin to manage Brooklyn services, from creating services in the catalog to adding new items to the catalog; from specifying catalog items in the manifest to specifying new types of services in the catalog. And now we’ve seen a bigger service that we can control using sensors and effectors.

But Brooklyn has more, and in the next part we are going to show you how to deploy an application to Cloud Foundry with multi-cloud, polyglot persistence provided by Apache Brooklyn. See you then.

Cloudsoft is a sponsor of The New Stack.

Feature image via Flickr Creative Commons.

Group Created with Sketch.
TNS owner Insight Partners is an investor in: Simply.
TNS DAILY NEWSLETTER Receive a free roundup of the most recent TNS articles in your inbox each day.