OpenStack Frequently Asked Questions

Published on: November 4, 2016

What is the cloud?

The cloud is a pool of separate resources that can be provisioned on demand as needed. It provides a concept of Infrastructure-as-a-Service (IaaS). As a user of the cloud, you can provision resources through the Dashboard WebUI, command line utilities, or direct access to the RESTful service APIs.

How much do your Nebula and Singularity services cost?

At the end of the month, HostDime will only bill you for what you have actually used. The prices and price estimator for Nebula (OpenStack compute) can be found at the bottom of the Nebula product page. Singularity (Swift object storage) prices are listed at the bottom of the Singularity product page. This can be used if you have an idea of what resources you will be using for the month. Most services are priced on a per-hour basis. You will be able to see your up-to-the-day incurred costs in Core so you can keep an eye on how much you are spending.

What support do you offer?

Nebula and Singularity are both self-service and features basic management. We support the service APIs and the physical cluster networking (not a project’s virtual networking, but the bare metal infrastructure behind it). If you run into errors using the API or dashboard, please let us know. Otherwise, you will be responsible for the contents of your project. At this time, we offer no official support for how to use OpenStack nor do we offer assistance with instances or any other project resource. We have written resource articles to help you get started, but your best bet is to simply read the official OpenStack documentation.

Why can’t I access my instance?

This is a very common question. After creating your first instance, how do you access it? There are a few things you need to do to get networking setup for your first instance. We wrote a resource article that details how to manage OpenStack networks via the Horizon Dashboard. This resource covers creating your first network and assigning floating IPs for public internet access. It is recommended that you read the entire article as we cover network setup and firewall rules, both of which are often the reason you can not access your instance.

The Security Group and/or Firewall-as-a-Service (FWaaS) rules need to be configured to allow certain incoming and outgoing traffic. Information on how to set up Security Groups can be find in our Configuring Security Groups resource page. For testing, we recommended temporarily allowing all inbound and outbound connections.

What is the default user for the cloud images?

The “cloud-user” account is typically the default user. When you first create a Linux instance, you should be able to SSH into the instance as “cloud-user” with the keypair you provided.

Is the HostDime cloud highly available?

Very! We have designed Nebula so that all openstack services are highly available. If any of our physical server fails, the remaining servers pick up the slack with no API service downtime to the end-user. Our engineers will fix the problem in the background with no service downtime.

If the host compute node of an instance fails, all instances on that node are immediately migrated to the remaining compute nodes. This downtime is not covered by our SLA, but in the unlikely event that it happens, recovery should be automatic.

Each file that is uploaded to the block storage (Cinder/Glance) and object storage (Swift) services have multiple backups. This means that corruption or loss of data is near impossible due to hardware failure. We still recommend tertiary backups as a matter of best practice, but we are proud of the data safety that this redundancy provides.

What is the maximum bandwidth throughput of your cloud?

The OpenStack Nebula and Singularity services is built to achieve up to a 40 Gbp/s bandwidth channel as it is backed by multiple high-end fiber network devices. While you will likely never use this much, it is important that we be able to serve the needs of all tenants of the cloud without service degradation.

What is your service level agreement for OpenStack?

Our OpenStack Public Cloud service level agreement can be found under the Legal section of HostDime.com.

What is vertical and horizontal scaling?

Vertical scaling is when you make one server more powerful by giving it more processing power or RAM. Horizontal scaling is when additional servers are clustered to serve content in parallel. The cloud focuses on horizontal scaling as you can create as many servers as you need.

How do I make my application cloudy?

Applications should be abstracted as much as possible. This may be different depending on your actual use case but it is best explained by examples.

Multiple instances can be created with a basic Linux, Apache, MySQL, and PHP (LAMP) stack. A volume can then be created and attached to all instances as a single shared network file system. By putting your application on this shared volume, you can spin up instances on the fly, attach this same volume and be ready to horizontally scale your application on demand.

The Load-Balancing-as-a-Service (LBaaS) offered by Neutron can be used to automate load balancing between the instances. Others may look into putting all of their static website content into Swift for fast processing and its high scalability. Links to these files can be placed straight into your website’s code. By uploading files to Object Storage rather than the traditional “local uploads folder”, files that are unrelated to your core application will not be in the way of your scaling needs.

How should object storage be used?

Object storage provides a place to store static files. With HostDime’s Singularity (Swift) service, high-demand files can be placed here for fast and reliable transfer speeds. This also helps to offload processing power away from instances. This can be used for simple file backup or as a means of simply storing uploaded files that are unrelated to your core application.



Back To Top