LunaNode

What LunaNode actually is

LunaNode is a Canadian cloud provider — LunaNode Hosting Inc. — running an OpenStack-based KVM virtualization platform with hourly billing across two Canadian locations, Toronto and Montreal. It is a small company by hyperscaler standards: the same product has been sold since the early 2010s under the "Luna Node Dynamic" name, third-party pricing trackers date the founding to August 2013, and the company footer claims a copyright start of 2011. Either way, this is a long-running independent host, not a venture-fueled newcomer, and its shape is familiar to anyone who has shopped at the value end of the market: virtual machines, block storage volumes, snapshots, floating IPs, an API, and a control panel, sold at prices the big providers do not touch.

The pitch is granular billing and data residency. Every cloud service — virtual machines, image and volume storage, virtual networks — is billed hourly, with monthly-equivalent caps published on the pricing page, so a test environment spun up on Tuesday and destroyed on Thursday costs cents rather than a month of a reserved instance. For Canadian businesses with data-residency requirements, or developers who simply want their infrastructure in Canada, LunaNode is one of very few independent options left in that niche.

Pricing: what you really pay

The lineup is three families, all published with both a monthly figure and an hourly rate. Memory-optimized is the entry point: m.1s is $3.50 per month ($0.00486 per hour) for 1 GB RAM, 1 virtual core, 15 GB SSD storage and 1,000 GB bandwidth — the cheapest way in. m.2 is $7/month (2 GB), m.4 $14 (4 GB), m.8 $28 (8 GB), m.16 $56 (16 GB), m.32 $100 (32 GB), m.64 $200 (64 GB RAM, 8 vCPU, 500 GB SSD, 6,000 GB bandwidth).

General purpose flips the ratio toward CPU: s.half is $4.50/month (0.5 GB RAM, 1 vCPU, 15 GB SSD, 1,000 GB bandwidth), then s.1 at $7 (1 GB, 2 vCPU), s.2 at $14, s.4 at $28, s.8 at $56, s.16 at $100, s.32 at $200 (32 GB, 12 vCPU, 500 GB, 6,000 GB). Compute-optimized gets dedicated cores: c.1 is $20/month for 5 GB RAM and one dedicated core with 100 GB SSD and 3,000 GB bandwidth, scaling to c.16 at $320/month (80 GB RAM, 16 dedicated cores, 500 GB SSD, 10,000 GB bandwidth). That $3.50-to-$320 spread is the entire published compute catalog, and the price-per-resource ratios are competitive: the mid-range s.4 at $28 for 4 GB and six vCPUs is the kind of configuration that costs noticeably more at the big three.

All displayed prices are USD; CAD payments are converted to USD account credit at the rate on the day of payment, which is a small but honest piece of fine print for a Canadian vendor. Volumes, images, and virtual networks are billed hourly too — which cuts both ways: persistent volumes and snapshots cost money every hour they exist, so cleanup discipline matters. There is no free tier and no published trial credit that we could find.

Performance & infrastructure

The hardware description is modest and specific rather than breathless: pure SSD or SSD-cached RAID10 storage and Intel Xeon CPUs. Note the careful wording — the vendor's own pages say SSD, not NVMe, and we have not independently verified the underlying drive models; our database tags this tool as NVMe-backed and we flag that as unconfirmed rather than repeat it. There are no published benchmarks from LunaNode and none in our own 2026-04-01 test set, so we make no performance claims here beyond the spec sheet; our /benchmarks page covers the providers we have actually measured.

Location-wise it is Toronto and Montreal — and Montreal is the interesting one, launched in the OVH BHS facility, with contemporary trade coverage noting Montreal machines benefit from OVH's DDoS protection. That is the only DDoS mitigation claim we can source, and it is location-specific; we record it with that caveat. IPv6 is supported in Toronto: virtual machines get an address via SLAAC, and user-created virtual networks receive a /64 allocation, per the vendor's own FAQ. Snapshots and block volumes are first-class features (with dedicated feature pages), VMs can be resized via the panel, and there is a documented API plus DNS, email, and uptime-monitoring services offered alongside compute. Storage provisioning has a notable quirk worth knowing: shrinking a local-storage VM is a manual snapshot-volume-copy dance documented step by step in the FAQ — fine for a hobbyist, a small red flag for teams that resize constantly.

The day-to-day operating details live in a wiki that is functional rather than polished. Provisioning a VM can take up to five minutes depending on the operating system, with progress visible through a VNC console; template images come with conventional default users (ubuntu, debian, cloud-user and so on), while ISO-based installs are manual through VNC. Volumes attach cleanly to Linux guests using the virtio driver, and Windows users may need to shut the VM down before attaching or detaching storage — the kind of detail that separates a wiki written by people who run the platform from marketing copy. There is a shelving feature for pausing instances cheaply, plus floating IP addresses, private networking, security groups, and SSH key handling, which is a respectable cloud feature set for a provider this size. The API covers the same ground as the control panel, so deployments can be automated without clicking.

Two operational cautions stand out. First, resizing: growing a local-storage VM is straightforward, but shrinking one is a manual snapshot-and-copy procedure involving a rescue VM and filesystem tools — workable, but a reminder that this is a value platform with rough edges. Second, account funding is prepaid-credit style, and image and volume storage accrue hourly charges, so abandoned snapshots quietly spend money. On the positive side of the ledger, being small cuts the other way too: trade write-ups over the years describe direct, transparent communication from the operators, and hosting-forum users have repeatedly called the platform underrated. We weight that community reputation lightly but not zero — a small provider whose support tickets are answered by people who can fix the problem is worth a lot when something breaks at 2 a.m.

Where it falls short

Footprint is the obvious limitation: Canada only. No US, European, or Asian regions means latency for distant users and no geographic redundancy story beyond pairing the two Canadian cities. Second, this is a small operator — support is ticket-based, the wiki is functional but dated in places (VM templates referencing Ubuntu 14.04-era examples), and there is no enterprise support ladder or published uptime SLA that we could find. Third, the feature set stops where the hyperscalers begin: no managed Kubernetes, no serverless, no object-storage catalog we could confirm, no marketplace of one-click images. Fourth, the account model is prepaid-credit style, which suits tinkerers and agencies but may frustrate procurement teams used to invoicing and committed-use discounts. Finally, our database tags claim backups are included and that US datacenters exist; both are wrong or unconfirmed — backups are not a published free inclusion and the company is entirely Canadian.

Who should pick this

Our pick for Canadian developers, agencies, and small businesses that want their VMs in Canada with true hourly billing and do not need a hyperscaler's service catalog. It is a good fit for staging environments that live a few days at a time, for cost-conscious always-on small servers, and for anyone with a data-residency requirement inside Canada. It is a poor fit for multi-region architectures, for teams wanting managed Kubernetes or rich PaaS, and for organizations that need a contractual SLA and named support before signing. If you are comparing against Hetzner, Vultr, or OVH directly, LunaNode's edge is Canadian placement and per-hour granularity; its weakness is scale, polish, and geographic reach.

Key takeaways

• Canadian independent cloud (LunaNode Hosting Inc.), OpenStack/KVM, in business since roughly 2013, with Toronto and Montreal locations.
• True hourly billing across VMs, volumes, images and networks; entry VM m.1s is $3.50/month (1 GB RAM, 1 vCPU, 15 GB SSD, 1,000 GB bandwidth).
• Dedicated-core compute tops out at c.16 for $320/month; the full published catalog runs $3.50 to $320.
• IPv6 in Toronto (SLAAC plus /64 per virtual network); DDoS mitigation is documented only for Montreal via the OVH BHS facility.
• SSD/SSD-cached RAID10 storage per the vendor — the NVMe claim in our tags is unconfirmed.
• No free tier, no published SLA, Canada-only footprint; snapshot and volume storage costs accrue hourly.

Sources

• https://www.lunanode.com/pricing
• https://www.lunanode.com/about
• https://lunanode.com/features
• https://wiki.lunanode.com/index.php/Frequently_asked_questions
• https://lowendbox.com/blog/luna-node-now-available-in-montreal-openstack-kvm-from-3-60month
• https://vps-prices.com/provider/lunanode
• https://webvps.blogspot.com/2015/05/luna-node-dynamic-kvm-platform-with.html