Service status & availability Crypto-only billing · No-KYC signup

Dedicated workload

Storage server rental

Rent 192 TB or 480 TB of raw HDD capacity with separate NVMe devices. Plan the usable space, backup strategy and recovery time before choosing the larger number.

Hardware priorities

  • raw capacity — verify it against the application, dataset and recovery objective.
  • RAID planning — verify it against the application, dataset and recovery objective.
  • off-site replication — verify it against the application, dataset and recovery objective.

Raw and usable capacity are different

192 TB raw comes from 12 × 16 TB HDDs; a single conventional RAID 6 group with no spare yields 160 TB before filesystem overhead. 480 TB raw comes from 24 × 20 TB HDDs; a single conventional RAID 6 group with no spare yields 440 TB before filesystem overhead. These are arithmetic examples, not recommended layouts. The separate NVMe boot devices are excluded.

Deployment path

  1. Select a configuration and US region.
  2. Choose TrueNAS or attach a custom ISO.
  3. Add an SSH key and fund the USD ledger with crypto.
  4. Confirm the installation window and wait for completed delivery before moving production traffic.

Shortlist

Configurations for this workload.

ProcessorCoresMemoryStorageWeeklyMonthlyAction
Xeon Silver 4310Storage · 192 TB raw
12c/24t
64 GB ECC
12×16 TB HDD + 2×480 GB NVMe
$105per week
$349per month
Configure →
Xeon Gold 6338Storage · 480 TB raw
32c/64t
128 GB ECC
24×20 TB HDD + 2×960 GB NVMe
$195per week
$649per month
Configure →

Before you choose

Plan capacity, recovery and operating cost.

4 min guide

Choose bulk capacity for the right workload

Consider dedicated HDD storage for backup repositories, archives, media libraries and large datasets. Start with retention, growth and retrieval requirements. Stored capacity and daily change volume are different sizing inputs.

If the primary requirement is low-latency random access for transactional databases or many busy VM disks, compare the NVMe range as well. HDD capacity can be economical for the dataset while still being unsuitable for the access pattern. Keep compute, memory and storage behavior in the same purchasing decision.

Two capacity classes, two different commitments

Xeon Silver 4310 offers 192 TB of raw HDD space from $349 per thirty-day term. Xeon Gold 6338 increases capacity to 480 TB and changes compute and memory too. Compare complete specifications and prices; it is not a disk-only upgrade.

Both configurations list separate NVMe devices. Decide how those devices will be used for the operating system and supporting services; they are not automatically a cache, a mirrored boot volume or extra HDD-pool capacity. Confirm the controller mode and disk presentation required by your chosen storage software.

Translate raw capacity into a storage plan

Raw TB add together the manufacturers’ decimal disk capacities. TiB use binary units, so the numerical value shown by an operating system can be lower before any redundancy is applied. Parity, mirrors, hot spares, metadata and a working free-space allowance further change what applications can store.

As arithmetic examples only, putting all HDDs in one conventional RAID 6 group with no spare leaves 160 TB or 440 TB before filesystem overhead. These figures are not a recommended layout or a guarantee of usable space. Multiple groups reserve parity within each group; a different topology changes the calculation and failure behavior. For TrueNAS, evaluate the actual vdev layout and its capacity calculation.

Size the recovery window as well as the repository

Ask how much data must be read or restored during an incident, not just how much can fit on the disks. Both storage offers include a 1 Gbps unmetered port. Port speed, disk access, encryption, remote storage and the receiving system can each limit a restore. An advertised capacity does not establish a recovery-time commitment.

A 10 Gbps upgrade is quote-based and must be confirmed for the intended configuration and region. Measure a representative restore before relying on a deadline. Keep a copy outside this machine and test that the access credentials and restoration process are available when the original environment is unavailable.

Agree the storage and maintenance responsibilities

RAID can provide redundancy against certain drive failures; it does not replace a backup against deletion, compromised credentials or loss of the whole host. Before ordering, define your required disk-replacement process, monitoring responsibilities and the evidence needed when a device appears unhealthy. Do not assume a rebuild duration from capacity alone.

TrueNAS and Linux are available system choices. Confirm the release, hardware compatibility and intended filesystem before installation. Hardware RAID and off-site backup are separately quoted options, not an implied part of a software-defined storage design. You remain responsible for shares, access controls, application data and verified recovery copies.

Turn the capacity requirement into an order

Prepare the region, retained data size, change rate, restore target, operating system and redundancy. Compare the weekly or thirty-day price and quoted options. Supported cryptocurrencies fund the USD balance used for service charges.

Specify what the machine must store and how it will be recovered. Include initial access, an independent backup location and eventual data export in the plan.

Direct answers

Questions before you order

Are the advertised TB figures usable filesystem capacity?

No. They are raw HDD capacity. Redundancy, spares, filesystem overhead, free-space reserves and TB/TiB presentation all affect the final result.

Is RAID or a backup service automatically configured?

Do not assume a layout or off-site copy. Agree the installation design separately. Hardware RAID and off-site backup are quote-based options; your recovery plan must identify what is actually included.

Can I use a storage server for Proxmox VM disks?

It depends on the VM workload and storage design. Assess random I/O, latency, memory and failure handling, and compare NVMe configurations when active guest storage is the main requirement.