Servers

Adding GlusterFS

GlusterFS is a scalable network file-system (NAS), and it's easy to add to your application via the Cloud 66 Dashboard.

Adding GlusterFS

To add GlusterFS to your application:

  1. Open your application from the Dashboard.
  2. Click on Data Sources in the left-hand nav and then Add Source in the sub-nav
  3. Click the green + Add Data Source button and select GlusterFS
  4. A drawer will open from the left, with configuration options for the server.
  5. Choose a Replication Factor - this is how many servers will run the Gluster cluster. If you require additional redundancy (and to avoid split-brain) you should choose "3" here.
  6. Click Add Server to start the process

You can now watch the logs, as usual to see the progress of the process.

Configuring GlusterFS replica count

Replicas are exactly what they sound like - parallel copies of the filesystem kept in sync. They are useful for fault tolerance and load balancing. You can specify replica_count when you configure GlusterFS. If you are using a manifest to add GlusterFS, you can add replica_count to your manifest.yml.

Replica count cannot be changed

You cannot change replica_count after GlusterFS has been added to your application.

Why would I need GlusterFS?

Almost all applications have some sort of data storage needs. As your application grows and runs on multiple servers, you will need to be able to share this data storage between your servers. For example, a web application that allows its users to upload images, will need to store those images on a share storage accessible by all servers.

GlusterFS is one of the options you have to for a Network Attached Storage (NAS) or shared file-system. Other options include NFS (Linux Network File System) and Ceph as well as many other tools and open source projects.

GlusterFS gives you a shared storage space that is accessible from each server or container on your application and is resilient to faults with powerful access control features.

How can I use GlusterFS in my application?

Now that you have a share storage service provided by GlusterFS in your application, you can use it in your application like a normal disk volume. By default, Cloud 66 will create and mount a shared volume on /mnt/gfs/<volume-name> on every application server of your application.

To see how your shared file system works, you can SSH to one of you web servers and run the following commands:

$ cd /mnt/gfs/<volume-name>
$ touch hello.txt

Now SSH to another web server on your application and you should be able to see hello.txt under /mnt/gfs/<volume-name>.

Fine grained access control for your data

By default Cloud 66 builds a GlusterFS cluster for your application, creates a default mount point on it and mounts that onto every application server. This is great to start with and for many workloads.

Using the manifest file allows you to choose the servers on which you would like to have the volumes mounted.

Accessing your GlusterFS servers

GlusterFS servers are added to a new group called GlusterFS Cluster under your application. These servers are accessible via the usual GlusterFS tooling (available from GlusterFS website).

Every server in your application will have the following 3 environment variables available by default:

  • GLUSTERFS_<GROUP_NAME>_ADDRESS_INT The internal IP address of the master volume server in your GlusterFS cluster.
  • GLUSTERFS_<GROUP_NAME>_ADDRESS_EXT The external IP address of the master volume server in your GlusterFS cluster.
  • GLUSTERFS_<GROUP_NAME>_HOSTS The internal hostnames of all server in this GlusterFS cluster.
Previous
Adding & Scaling Elasticsearch