Cloud · Private Cloud

Private Cloud

Your own Proxmox VE cluster on dedicated hardware in a Canadian data centre, with Ceph storage spread across the nodes. We own and run the hardware, the network and the virtualization layer. You get a private cloud without running a data centre to have one.

See the three sizesWhich cloud's for me?
  • Dedicated hardware
  • Proxmox VE
  • Ceph storage
  • Billed in CAD
  • 3Facilities, all in Canada
  • 2Cities: Toronto and Vancouver
  • SOC 2 Type IIReport held
  • 1-866-821-8355Phone

Three sizes

One node to build on. Three to run on. Five when three is not enough.

Every private cloud is dedicated hardware running Proxmox VE. What changes between the sizes is how many nodes you get, and whether the cluster can lose one of them and carry on.

  1. One node

    Development and staging

    A single Proxmox VE server, yours alone, in one of our Canadian facilities. Build, test and stage on hardware nobody else touches. It is a server, not a cluster: there is no second node to fail over to and no Ceph, so it carries no availability commitment and it can go down for maintenance or a hardware fault at any time. That is the point of one node. Run work here that you can rebuild, not work that has to stay up.

  2. Three nodes

    The production minimum

    Three servers, clustered, with Ceph across all three. This is the smallest private cloud we will sell for production, and the next section says exactly why. A node can fail, or come out for patching, and your machines keep running on the two that are left.

  3. Five nodes

    When you outgrow three

    Two more nodes of the same hardware. More room for machines, more Ceph capacity, and more headroom while a node is out. Past five, the cluster grows a node at a time — the design conversation is the same one.

One node

Development and staging

From$1,099.95 CADper month

$1,299.95 CAD regular · up to 15% off

Choose your billing term next. Every term bills its full period up front.

Nodes
1
Per node
Dual Intel Xeon Gold 6210
Memory per node
256 GB DDR4 ECC
Storage per node
2 x 1.92 TB enterprise SSD, hardware RAID 1
IP block
/29 (5 usable)
Storage
Local to the node. No Ceph.
Availability
None. Development and staging only.
  • Cloudflare Magic Transit network DDoS protection included
  • Unmetered traffic
  • Full Proxmox VE cluster console
  • Managed Support available
  • 24/7 support

No availability commitment. One node is a server, not a cluster: no Ceph, no failover, no live migration. Run work here you can rebuild.

Order one node

Three nodes

Production

From$3,299.95 CADper month

$3,899.85 CAD regular · up to 15% off

Choose your billing term next. Every term bills its full period up front.

Nodes
3
Per node
Dual Intel Xeon Gold 6210
Memory per node
256 GB DDR4 ECC
Storage per node
2 x 1.92 TB enterprise SSD, hardware RAID 1
IP block
/28 (13 usable)
Storage
Ceph across all three nodes
Availability
High availability, automatic failover, live migration
  • Cloudflare Magic Transit network DDoS protection included
  • Unmetered traffic
  • Full Proxmox VE cluster console
  • Managed Support available
  • 24/7 support
Order three nodes

Five nodes

Production

From$5,524.95 CADper month

$6,499.95 CAD regular · up to 15% off

Choose your billing term next. Every term bills its full period up front.

Nodes
5
Per node
Dual Intel Xeon Gold 6210
Memory per node
256 GB DDR4 ECC
Storage per node
2 x 1.92 TB enterprise SSD, hardware RAID 1
IP block
/27 (29 usable)
Storage
Ceph across all five nodes
Availability
High availability, automatic failover, live migration
  • Cloudflare Magic Transit network DDoS protection included
  • Unmetered traffic
  • Full Proxmox VE cluster console
  • Managed Support available
  • 24/7 support
Order five nodes

Past five nodes the cluster grows a node at a time and the price stops being a list price. Tell sales what you run and you get a written quote in Canadian dollars.

Every size runs Proxmox VE, and only Proxmox VE. We no longer sell VMware or Hyper-V private cloud. If you are arriving from one of those pages, see Private Cloud after VMware.

Why three

Three nodes is the floor, and it is not a sales rule.

It comes from how Proxmox VE and Ceph work. Two nodes cannot vote, and two nodes cannot hold three copies of your data.

Quorum: the cluster decides by majority

A Proxmox cluster only acts while more than half of its nodes can see each other. That majority is called quorum. Lose it and the cluster drops to read-only: your machines keep running, but nothing can be started, moved or changed.

With two nodes and one failure, the survivor holds one vote out of two. One is not more than half, so a two-node cluster cannot recover from losing a node on its own. Proxmox says it plainly in its own cluster documentation: if you want high availability, you need at least three nodes for reliable quorum.

Ceph: your data is written to more than one node

Ceph is the storage. Instead of disks that live inside one server and fail with it, every write is copied to several nodes at once. A Proxmox storage pool keeps three copies by default and stays writable while two of them are reachable.

Three nodes is the smallest cluster that can put those three copies on three different machines, so losing one loses no data and stops nothing. Ceph's own monitors need three to hold their quorum as well. Proxmox's guide to a hyper-converged Ceph cluster opens on the same point: use at least three servers.

Put the two together and two nodes has no honest job in production. That is why the ladder goes one, three, five, and why we stopped selling a two-node private cloud.

Included with every cluster

What you get, whichever size you take.

The hardware is dedicated and the operations are ours. Everything below is part of the cluster, not an add-on with a line on the invoice.

  • Dedicated hardware

    Every node is yours. No other customer's workloads run on it, and nothing else competes for its CPU, memory or disks.

  • Ceph storage across the nodes

    Your data is written to more than one node, so losing a node loses nothing. Storage is not a box under one server that fails with it.

  • Automatic failover

    If a node stops, the machines that were running on it start again on the nodes that are still up. You do not have to be awake for it.

  • Live migration

    Move a running machine from one node to another without shutting it down, so patching a node is not a maintenance window for your users.

  • A cluster console

    One Proxmox VE web console for the whole cluster: create machines, resize them, snapshot them, watch what they are doing.

  • Unmetered traffic

    The port is part of the cluster and we do not bill by the gigabyte, so a busy month does not arrive as a surprise line on the invoice.

  • Our engineers run the platform

    We own and operate the hardware, the network and the virtualization layer. You run what is inside your machines, and Managed Support can take that on too.

  • Billed in Canadian dollars

    CACloud prices in CAD, full stop. No conversion, no exchange rate moving your invoice between months.

Failover and live migration are cluster features, so they apply from three nodes up — a single development node has nowhere to move a machine to. Uptime commitments and credits are on the Service Levels page. How traffic is filtered before it reaches a machine is on Security Architecture.

How an order runs

Order a size, then we build it with you.

The three sizes above are priced and orderable. What is not on a price list is the design conversation, and that still happens.

  1. Pick a size and a term

    One node to build on, three to run production, five when three is not enough. The term is chosen in the cart and bills its full period up front.

  2. We scope it with you

    Our engineers go through what you are moving, how the machines are sized and what has to stay isolated, before anything is racked. Nothing is guessed.

  3. We build it and hand you the console

    We rack the nodes, cluster them, bring up Ceph and give you the Proxmox console. Your first machines can be provisioned for you.

Need something the three sizes do not cover — more nodes, different memory or disk, a specific network? Tell sales what you run and the answer is a written quote, not a price list.

Not sure you need your own nodes?

Enterprise VM

Shared enterprise hosts, sold per virtual machine, with highly available storage and automatic host failover. Start with one machine and grow, with no monthly hardware commitment.

See Enterprise VM

Related

FAQ

Private cloud questions.

What is a Private Cloud?

Private Clouds are for organizations that want maximum control of their data plus the benefits of the cloud. A CACloud private cloud is dedicated hardware running Proxmox VE, with Ceph storage written across the nodes so your data lives on more than one machine. There are three sizes: one node for development and staging, three nodes as the production minimum, and five when three is not enough. All three are priced and orderable, and you choose the billing term in the cart. One node is a server rather than a cluster, so it carries no availability commitment. Past five nodes, our engineers scope and quote it.

Can I run Windows or Linux on my Cloud Server?

Yes. You can pick from preconfigured Windows and Linux templates, including Red Hat, Debian, Fedora and Ubuntu, right from the control panel. You can also one-click install leading applications such as WordPress, Drupal and Joomla.

What if I want to change my operating system?

At any time you can create a new Cloud Server from any of our templates, either by rebuilding your server or by adding a new one built for your new needs. Log in to your control panel, select “Manage”, click “Rebuild”, then pick your OS and distribution to start the rebuild.

More answers on the FAQs page.

Tell us what you run, and we will design the cluster around it.

Sales and our virtualization engineers scope the node count, the memory and the Ceph capacity with you, then send a written quote in Canadian dollars.

Email sales1-866-821-8355