The Dell PowerEdge R640 VxRail (E560F) is not a general-purpose R640 variant. It is a purpose-built hyperconverged infrastructure (HCI) node designed specifically for VMware vSAN and VxRail environments. It ships from Dell in a form factor based on the R640 chassis but with hardware configuration locked to VxRail certification requirements. If you are building or expanding a VxRail cluster, this is the node. If you are looking for a general-purpose R640 for standard virtualization or storage workloads, one of the standard R640 configurations is the right call.

VxRail is VMware's jointly engineered hyperconverged appliance platform built on vSAN, managed through the VxRail Manager plugin in vCenter. The E560F is the all-flash node in the VxRail E-series: a 1U, 10-bay 2.5" chassis optimized for NVMe and SAS SSD in vSAN all-flash configurations. It is sold as a complete hardware-software stack, not a configurable build-your-own platform in the traditional sense. Refurbished VxRail nodes require careful consideration of VxRail software licensing, vSAN licensing, and cluster compatibility, and we cover that reality in full below before describing the hardware. Read the licensing section before requesting a quote if VxRail is your deployment target.

To configure a build, call 1-800-778-1545 or use the quote form below. Every refurbished unit ships under our 180-day warranty with 12+ hour burn-in testing, and volume pricing starts at 5 units. VxRail builds in particular benefit from a design conversation before the quote, so we recommend opening with a call rather than a form submission.

When the VxRail E560F Is the Right Node

The E560F earns its place when one of these design patterns applies: expanding an existing VxRail cluster where hardware compatibility with the current cluster version is confirmed and licensing scales cleanly, building a dev/test VxRail environment at meaningfully reduced capital cost vs new hardware list pricing, or organizations with existing VxRail entitlements that cover additional nodes through their existing contract. The common thread is that VxRail is already in the environment in some form (production cluster, existing entitlement, sandbox cluster) and the refurbished node is fitting into that existing context.

What does not belong on this node: greenfield vSAN deployments where the customer has no prior VxRail experience and the operational overhead of VxRail Manager is not justified (a standard vSAN ReadyNode on the 10-Bay NVMe is the simpler architecture), general-purpose virtualization without vSAN (any of the standard R640 variants are the cleaner answer), and non-VMware hypervisor environments (VxRail is VMware-only). We will tell you directly at quote time when VxRail is the wrong answer for your environment, even when refurbished VxRail hardware is what you initially asked about.

Storage - 10 All-Flash Bays (vSAN All-Flash Architecture)

Ten 2.5" hot-swap bays on a backplane supporting NVMe and SAS SSD, same backplane architecture as the R640 10-Bay NVMe variant. The E560F is designed for vSAN all-flash deployments: configurations where both cache tier and capacity tier are solid-state. This is meaningfully different from hybrid vSAN (SSD cache plus HDD capacity) in performance characteristics and cost profile, and the all-flash architecture is what defines the E560F vs other VxRail E-series nodes.

• Cache tier drives: High-endurance NVMe or SAS SSD, mixed-use or write-intensive (1 to 3 DWPD minimum). The cache tier absorbs writes before destaging to capacity. Do not use read-intensive drives here; the endurance mismatch will shorten drive life significantly under production vSAN write patterns.
• Capacity tier drives: NVMe or SAS SSD. Read-intensive drives are appropriate here because the capacity tier is predominantly read under normal vSAN operation after destaging.
• Disk group architecture - OSA vs ESA: vSAN ESA (Express Storage Architecture, available in vSAN 8.x) changes how disk groups work compared to vSAN OSA. ESA does not use the traditional cache plus capacity disk group model; all drives participate in a unified storage pool. If your VxRail deployment targets vSAN ESA, the disk configuration requirements differ. Confirm with your VMware account team which vSAN architecture applies to your target deployment before finalizing hardware.
• BOSS module for boot: Mandatory on every VxRail node. ESXi boots from the BOSS-mirrored M.2 SSDs; the front bays are reserved entirely for vSAN cache and capacity tier drives.

Storage Controllers

VxRail's vSAN-managed storage architecture means hardware RAID is not in the data path on the front bays. The NVMe drives bypass the PERC controller entirely; the SAS SSD drives present through an HBA in pass-through mode so vSAN manages them directly. The controller landscape on the E560F is shaped by that constraint:

• HBA330 (pass-through HBA): The standard controller on VxRail SAS-tier configurations. Pass-through to vSAN without hardware RAID abstraction. vSAN manages drive redundancy at the policy layer.
• NVMe direct attach (no controller): NVMe drives connect directly to CPU PCIe lanes. No controller in the data path; vSAN manages redundancy.
• PERC H730 (1 GB cache, battery-backed): The 13th-gen-era controller that Dell maintained Mini-PERC slot compatibility for on 14th gen. Appears on refurbished E560F units occasionally as a carryover from earlier deployments. Not in the vSAN data path (vSAN does not use hardware RAID), but may be present on the node managing rear-bay boot media or auxiliary storage. Generally not load-bearing on a VxRail configuration; flag at quote time so the customer knows whether their refurbished node ships with one and what role it plays in their specific configuration.
• PERC H730P (2 GB cache) and PERC H740P (8 GB NV cache): Similar to the H730 commentary; may be present on refurbished hardware but are not in the vSAN data path. Documented here for completeness when an auxiliary controller is part of the build.

Important VxRail-specific note: VxRail's Hardware Compatibility List (HCL) is strict about controller models. Refurbished hardware shipping with an unexpected controller can cause cluster validation issues. We verify the controller present on every refurbished E560F against the customer's target VxRail version before shipping.

VxRail Licensing - Critical Considerations for Refurbished Nodes

This section is why we recommend a call before a quote on VxRail. VxRail nodes require software licensing that is separate from the hardware purchase, and the licensing reality is more constrained than for standard R640 hardware:

• VxRail software subscription: VxRail Manager and the VxRail-specific vCenter integration require an active VxRail subscription from Dell. This subscription is not transferable with refurbished hardware. A new subscription is required for refurbished nodes added to a cluster or used to build a new cluster.
• vSAN license: VMware vSAN licensing is required separately. vSAN is licensed per CPU socket. For a dual-socket E560F in a new cluster, you need vSAN licenses for both sockets on every node.
• vSphere / ESXi license: VxRail runs on vSphere. ESXi licensing is required per host socket, separate from vSAN.
• Existing cluster expansion: Adding refurbished E560F nodes to an existing VxRail cluster requires compatibility validation against the cluster's current VxRail version and the VxRail HCL. Not all refurbished configurations will be compatible with all cluster versions. This validation must happen before purchase.
• VxRail version compatibility: VxRail versions tie hardware to specific firmware baselines, ESXi versions, and vSAN versions. Mixed-version clusters are constrained. Provide your current VxRail version, cluster model, and node count when requesting a quote so we can validate hardware compatibility before quoting.

Our recommendation: If you are expanding an existing VxRail cluster, the licensing path is usually straightforward: incremental nodes to your existing subscription. If you are building a new VxRail cluster from refurbished nodes, the licensing math may not favor VxRail over standard vSAN ReadyNodes, and we strongly recommend involving your VMware and Dell account teams in the design before procurement. We will tell you directly when a standard vSAN ReadyNode is the better economic answer for your specific situation.

Processors

CPU options on the E560F: Dual 2nd Generation Intel Xeon Scalable (Cascade Lake-SP, 2019). VxRail E-series supports Gold and Platinum tier processors within its certification matrix; Bronze and Silver are not typically certified on VxRail nodes. Socket LGA 3647 on the Intel C620-series chipset, same Purley platform as the rest of the R640 family but with the CPU options narrowed to the VxRail-certified subset.

Our SKU recommendations on this node: Gold 6230 (20 cores, 2.1 GHz, 125W) is the balanced default for general-purpose vSAN cluster workloads. Gold 6248 (20 cores, 2.5 GHz, 150W) is the right step up for VDI-on-VxRail clusters or high-VM-density production. For VxRail clusters carrying compute-intensive workloads (Oracle on vSAN, SAP on vSAN), Gold 6254 (18 cores, 3.1 GHz, 200W) delivers the per-core clock that those workloads benefit from. VxRail's certification matrix is the authoritative reference; we cross-check every CPU selection against the target VxRail version's HCL.

Heatsink requirement on top-bin CPUs: Any CPU above 150W TDP requires Dell's high-performance heatsink kit and high-performance fan kit. VxRail-shipped E560F units typically come with the correct kit for the CPU configured at factory, but refurbished units may have been re-CPUed in the field. We verify heatsink-to-CPU match on every refurbished E560F before shipping; it is one of the most common sources of "the node thermal-throttles under sustained load" issues we see on field-rebuilt VxRail hardware.

Dual-socket only: VxRail E560F nodes are dual-socket configurations. Single-socket is not a VxRail-supported design point on the E-series; the vSAN cache and capacity disk groups depend on the dual-socket PCIe lane budget for full bay enumeration.

Memory

Architecture: 24 DDR4 DIMM slots, 12 per CPU across 6 channels at 2 DIMMs per channel. Same Purley 6-channel layout. VxRail minimum memory requirements per node depend on cluster size, vSAN configuration (OSA vs ESA), and the per-host capacity provisioned; always size above the documented minimum to account for vSAN's reservation.

Supported DIMM types:

• RDIMM: Standard enterprise choice. Up to 64 GB per DIMM, 1.5 TB total at full population. The most common DIMM type on VxRail nodes.
• LRDIMM: Up to 128 GB per DIMM, 3 TB total. Common on high-VM-density VxRail builds where 3 TB of host memory backs many VMs per host.
• Intel Optane Persistent Memory (PMem): Cascade Lake L-series CPUs only. Supported on the E560F in specific configurations where the VxRail HCL validates the combination; Memory Mode is the more common Optane use case on VxRail for cost-effective memory pool expansion, App Direct mode for persistent storage tier extending alongside vSAN. Confirm the VxRail HCL allows your target Optane configuration before purchase.
• NVDIMM-N: Niche; rarely used on VxRail.

vSAN memory reservation: vSAN reserves a meaningful amount of host memory for caching, deduplication, compression, and metadata. On all-flash vSAN nodes the reservation is higher than on hybrid. The reservation grows with per-host capacity. Size the DIMM count to leave headroom for VMs after vSAN's reservation. We include this calculation in every VxRail node quote.

Memory speed by population: DDR4-2933 on Gold 6200 / 5222 SKUs at 1 DPC, DDR4-2666 on other Cascade Lake SKUs and at full 2 DPC. Full population is common on VxRail nodes because the workloads (VDI, high VM density, mixed enterprise virtualization) benefit from full memory bandwidth more than from the partial-population clock speed.

Mixing rules: Match ranks, capacity, and timing within a channel. VxRail's HCL is strict about DIMM consistency across nodes in a cluster; we cross-check this when expanding existing clusters.

Networking and PCIe Expansion

VxRail networking requirements are strict: 10 GbE is the minimum supported per-node networking for vSAN traffic; 25 GbE is strongly recommended for all-flash deployments to prevent the network from becoming the performance bottleneck. The E560F generates more storage traffic than a hybrid node because all-flash sustained throughput is meaningfully higher.

• 2x 25 GbE SFP28: The recommended baseline NDC for all-flash VxRail clusters. Pair with 25 GbE top-of-rack switching and a dedicated vSAN network. Most modern VxRail deployments land here.
• 4x 10 GbE SFP+: Acceptable for smaller VxRail clusters with modest VM density. Treat it as a transitional configuration where 25 GbE switching is not yet in place; not the long-term target for all-flash.
• 2x 25 GbE SFP28 plus add-in 100 GbE NIC: Common architecture for dense all-flash VxRail clusters. NDC carries management and VM traffic; the add-in NIC carries the vSAN storage fabric. Increasingly the right answer for production VxRail at scale.
• VxRail Manager network requirements: VxRail requires separated networks for management, vSAN, vMotion, and VM traffic. Plan the network design with these segments in mind; VxRail Manager validates this at cluster initialization.

PCIe expansion: Up to 3 PCIe Gen3 slots depending on riser configuration. The 10-Bay NVMe-capable backplane consumes meaningful PCIe lane budget; ten NVMe drives at x4 is 40 lanes from the front backplane alone, and PCIe slots and the NDC consume the remainder. We confirm lane allocation against the build at quote time. Common VxRail PCIe builds: 100 GbE add-in NIC for the vSAN storage fabric, occasional GPU add for VDI clusters carrying GPU-accelerated desktops, and Fibre Channel HBA when VxRail integrates with an external FC array (uncommon but supported).

GPU Support

GPU support on the E560F is constrained by the same 1U thermal envelope and PCIe lane budget as the other R640 variants. The typical VxRail GPU use case is VDI-on-VxRail with NVIDIA T4 cards (single-width, low-profile, 70W) for GPU-accelerated virtual desktops. Up to three T4 cards is structurally possible but may be limited by the lane budget when paired with full NVMe population; we validate this combination at quote time.

For VxRail clusters carrying AI inference workloads alongside general virtualization, the T4 configuration is the standard answer. For AI training workloads on VxRail (uncommon; usually a different platform is the right call), the R740-based VxRail nodes are the better fit. The E560F is a vSAN all-flash node first, a GPU compute platform a distant second.

Management - iDRAC9 and VxRail Manager

iDRAC9 Enterprise: Required for production deployment. Remote KVM, virtual media, predictive analytics, Group Manager, Quick Sync 2, and Silicon Root of Trust. iDRAC9 sits beneath VxRail Manager; iDRAC manages the hardware platform, VxRail Manager orchestrates the cluster-level lifecycle.

VxRail Manager: The cluster lifecycle management plane. Handles VxRail-specific node deployment, firmware compliance across the cluster, ESXi and vSAN version coordination, and the VxRail-aware upgrade process. The VxRail Manager experience is the operational differentiator vs running standalone vSAN; it is also the source of the additional licensing requirement.

Security baseline: Silicon Root of Trust anchors firmware verification in immutable silicon. System Lockdown mode prevents unauthorized firmware changes after deployment. TPM 2.0 module supported and recommended for any VxRail deployment subject to NIST 800-171, CMMC, FedRAMP, HIPAA, or PCI DSS compliance frameworks. VxRail's certified configurations include TPM as standard on most enterprise deployments.

Lifecycle Controller and OpenManage Enterprise: Same Dell management plane as the rest of the R640 family at the per-node hardware layer. VxRail Manager is the cluster-level orchestration on top.

Power and Cooling

All-flash VxRail nodes draw more power than equivalent compute-only configurations because the NVMe and SAS SSD drives consume sustained power under load (more consistently than spinning disks, which idle thermally). PSU recommendations for the E560F:

• Balanced (Gold 6230, full RAM, 10 mixed cache + capacity SSDs): 2x 1100W Platinum, peak draw approximately 620W
• Heavy (Gold 6248, full RAM, 10 NVMe drives, 25 GbE plus 100 GbE NICs): 2x 1100W Platinum, peak draw approximately 780W
• VDI-on-VxRail (Gold 6248, 3 TB LRDIMM, 10 SSDs plus T4 GPU): 2x 1100W Platinum or 2x 1600W Platinum for headroom, peak draw approximately 820W

Thermal: Eight hot-plug redundant fans standard. The high-performance fan kit is recommended on all-flash VxRail builds with Gold-tier CPUs because the sustained drive activity under vSAN load keeps thermal output elevated compared to compute-only nodes. ASHRAE A3 (40C) extended ambient support is achievable with the high-performance fan kit, though the operating margin on dense VxRail builds is tighter than on the standard 10-bay variants.

Physical Specs & Platform Notes

• Form factor: 1U rack server. 42.8mm H x 434mm W x approximately 735 to 760mm D depending on bezel and cable management options. Standard 19-inch rack mount with Dell ReadyRails II. Same physical footprint as the standard R640 10-Bay variants.
• PCIe expansion: Up to 3 PCIe Gen3 slots. Structurally identical to the standard 10-Bay NVMe chassis; the practical limit is the PCIe lane budget against the NVMe bay count, not slot count.
• Parts availability: Strong on the underlying R640 chassis components (PERC controllers, NDC cards, riser kits, fan modules, PSUs are widely available). The VxRail-specific firmware baseline and HCL constrain which exact part revisions are validated for a given VxRail version; we cross-check this on every refurbished E560F. VxRail-specific accessories (the VxRail-branded bezel, factory VxRail labels) may or may not be present on refurbished units; the underlying hardware is functionally identical with or without them.
• Accessories we recommend: Dell LCD bezel (P/N 521RX security bezel, 7M3F1 LCD bezel without security, 9NN24 with security; confirm part at quote time against your chassis revision and whether a VxRail-branded bezel is required for your environment), Dell ReadyRails II static or sliding rails, and the Dell cable management arm (CMA) for serviceability.
• Platform notes: VxRail firmware baseline must match the cluster's target VxRail version; this is the most consequential pre-purchase check on any refurbished VxRail hardware. NVMe bifurcation settings in BIOS must be configured correctly for drives to enumerate properly. CPU hot-plug is not supported. NDC swap requires powered-down access. Drive replacement in production must go through VxRail Manager rather than direct hardware swap to keep the cluster state consistent.

Our Assessment

Where it excels: Expanding existing VxRail E560F clusters where hardware compatibility with the cluster's current VxRail version is confirmed and licensing scales cleanly through the existing subscription. Building dev/test VxRail environments at meaningfully reduced capital cost vs new VxRail hardware list pricing, particularly for organizations running production VxRail elsewhere that want a matching dev/test platform. Organizations with existing VxRail enterprise agreements that cover additional nodes through their existing contract. vSAN all-flash workloads inside VxRail-managed environments where the operational benefits of VxRail Manager (cluster lifecycle, firmware orchestration, vSAN-aware upgrades) justify the licensing overhead. VDI-on-VxRail deployments where the all-flash architecture supports the random-I/O workload pattern that desktop pools generate.

Where to look instead: If you are building a vSAN cluster without prior VxRail experience and the operational overhead of VxRail Manager is not justified, the standard vSAN ReadyNode path on the R640 10-Bay NVMe gives you vSAN all-flash capability without the VxRail subscription requirement. You manage vSAN directly through vCenter rather than VxRail Manager. If your workload is general virtualization without vSAN, any of the standard R640 variants are the cleaner answer. If your hypervisor is not VMware, VxRail does not apply at all. If your workload needs PCIe Gen4 NVMe bandwidth, the VxRail E660F (R650-based, 15th gen) is the forward-generation step.

Bottom line: The E560F is specialized hardware for a specialized deployment. It delivers exactly what it is designed for, a validated and certified node for VMware vSAN all-flash HCI environments, but it carries more procurement complexity than a standard R640 configuration. The hardware is excellent. The licensing requirements are significant and non-negotiable. If you are considering this configuration, we strongly recommend a design conversation before a standard quote. VxRail deployments benefit from getting the configuration right before hardware ships; we would rather spend 30 minutes on the front end than ship hardware that creates licensing or compatibility issues on the back end.

Where the R640 VxRail Fits in 2026

The R640 family is 2 to 3 generations behind current Dell production. The 13th-gen step-down on the same workload profile is the Dell PowerEdge R630 10-Bay, which predates VxRail's mainstream cluster lifecycle tooling and is generally not appropriate for VxRail cluster expansion (Dell did not validate VxRail nodes on the R630 platform). The 15th-gen step-up is the Dell PowerEdge R650 platform, whose VxRail E-series equivalent is the E660F (Ice Lake-SP), and on the R660 platform the E660N (16th gen, Sapphire Rapids). The R640 10-Bay Standard page covers the generational ladder and support status for the base R640 family in full. VxRail-specifically: the E560F remains a strong fit in 2026 for cluster expansion where existing E560F nodes anchor the cluster's VxRail version, and for dev/test environments mirroring E560F production fleets. For greenfield VxRail deployments in 2026, the conversation about whether to land on E560F (14th gen), E660F (15th gen), or E660N (16th gen) depends on the deployment lifecycle expectation and the current price delta between generations; we walk through this at quote time when greenfield is the use case.

Honest Limitations

• VxRail subscription is not transferable. Refurbished hardware does not carry the original VxRail subscription. A new subscription is required, either through your existing VxRail enterprise agreement or as a standalone purchase from Dell. This is the most consequential procurement reality for refurbished VxRail.
• Cluster compatibility must be validated before purchase. Not every refurbished E560F configuration is compatible with every VxRail cluster version. We validate this against your current cluster before quoting; do not purchase without that validation.
• No hardware RAID on the vSAN data path. NVMe drives bypass the PERC controller; SAS SSD drives present through HBA pass-through. Redundancy is at the vSAN policy layer, not the controller. This is by design but is sometimes surprising to operators new to vSAN.
• VxRail Manager is a learned operational layer. If your team has not used VxRail Manager before, plan for the learning curve. The benefits (cluster lifecycle, firmware orchestration) are real but require operational familiarity.
• PCIe Gen3, not Gen4. NVMe drives are PCIe Gen3 x4. For workloads where per-drive Gen4 bandwidth matters, the VxRail E660F (R650-based) is the forward-generation step.
• 10 GbE is a floor, not a target on all-flash. All-flash VxRail generates enough storage traffic to make 10 GbE the bottleneck under load. 25 GbE is the appropriate target; 100 GbE is increasingly common on dense clusters.
• Mixed-version clusters are constrained. VxRail's cluster-version coordination is strict. Adding refurbished hardware running an older VxRail baseline to a current-version cluster may not be supported; the upgrade path is then "match the cluster first, add nodes second," which has its own operational implications.
• 14th gen, not current production. Dell's current production VxRail E-series is the E660N. The E560F represents strong refurbished value in 2026 but is not new hardware.

Workload Fit

Where to Look Instead

• Building a vSAN cluster without VxRail overhead? Standard vSAN ReadyNodes on the R640 10-Bay NVMe give you vSAN all-flash capability without VxRail subscription requirements. You manage vSAN directly through vCenter rather than VxRail Manager.
• General virtualization without vSAN? Any of the standard R640 variants support standard vSphere without VxRail or vSAN licensing complexity. The 10-Bay Standard chassis is the primary R640 build for general enterprise virtualization.
• Compute-first virtualization with SAN storage? The R640 8-Bay 2.5" is the compute-first chassis for vSphere hosts feeding centralized storage.
• Native NVMe storage with hardware RAID alternative? The 10-Bay NVMe is the NVMe-first chassis; the 10-Bay Standard with SAS SSDs is the hardware-RAID path for similar IOPS at lower acquisition cost.
• Need PCIe Gen4 VxRail? The VxRail E660F (R650-based, 15th gen) is the forward-generation E-series equivalent. Contact us for availability and pricing comparison.
• HPE HCI equivalent? HPE's HCI platform is SimpliVity, which uses a different architecture than VxRail and is not a direct one-to-one swap. For vSAN-specifically on HPE hardware, the HPE ProLiant DL360 Gen10 10-Bay as a vSAN ReadyNode is the closest analog (vSAN-certified but not HCI-appliance-packaged; the HPE cross-vendor companion to the R640 in 1U Purley configurations).

Ready to Configure?

VxRail configurations start with a conversation, not a quote form. Contact our account team with your current VxRail version (if expanding an existing cluster), node count, target cluster size, and workload profile. We will validate hardware compatibility against your cluster's VxRail HCL, confirm controller and DIMM consistency requirements, advise on licensing requirements (VxRail subscription, vSAN per-socket, vSphere per-socket), and provide a configuration recommendation before any pricing discussion. This is the right sequence for VxRail procurement; hardware selection without software validation creates expensive problems downstream. Every refurbished unit ships with the Wholesale Servers 180-day warranty and 12+ hour burn-in testing, and volume pricing starts at 5 units. Call 1-800-778-1545 to start the design conversation.