HPE Proliant DL380 G9 8-Bay 2.5"
Configure Your System:
1 Processor Required
Step 1: Choose a series
Intel Xeon E5-2600 v3 (Haswell) is the value play. Mature, plentiful, and the lowest cost per core we sell. Memory runs at 2133 MT/s.
Intel Xeon E5-2600 v4 (Broadwell) is the same socket a die shrink later: more cores at the top of the range, better performance per watt, and 2400 MT/s memory support. The price gap has narrowed enough that v4 is our default recommendation; take v3 when the budget is the whole story.
Whichever series you pick, populate both sockets. A single-CPU DL380 loses half its memory channels, half its DIMM slots, and the PCIe slots that route through the second processor.
Step 2: Choose your CPU
Running Windows Server 2022? Two 8-core CPUs land exactly on the 16-core base license. Our pick is the E5-2667 v4 (8 cores, 3.2 GHz): high clocks per core are exactly what Windows workloads reward, and the license covers the pair with nothing wasted.
Virtualizing on vSphere? VMware licenses a 16-core minimum per socket, so the 8-core strategy buys you nothing there. Take core count instead: the E5-2680 v4 (14 cores) is the workhorse pick, and the E5-2698 v4 (20 cores) is where dense consolidation lands.
General compute, Proxmox, and Linux duty: mid-range v4 parts in the 2650 to 2680 range cover the widest span of workloads per dollar.
Fair warning: a Gen9 platform is a value play, not a performance play. If your workload is licensed per core or single-thread bound, newer silicon can cost less in licenses than it adds in hardware, and we will tell you so at quote time if that is your situation.
2 Heat Sink Optional
What still matters is airflow discipline. Keep every drive bay blank and internal baffle installed; the DL380 cools front to back, and an open path lets air skip the components that need it. HPE's thermal engineering on this chassis is genuinely good, one of the reasons the Gen9 DL380 became the default enterprise workhorse of its era, and it holds up exactly as long as the airflow path stays intact.
It was engineered for datacenter intake air; it will run in a well-ventilated closet, but fan and drive life is better in a real cold aisle. If your build includes GPUs, cooling stops being simple; call 1-800-778-1545 before checkout and we will confirm the whole thermal picture.
3 Memory (RAM) 24 DIMM slots, modules added in sets of 2 Required
The E5-2600 platform runs four memory channels per socket. Populate up to two DIMMs per channel (16 slots total) and memory runs at full speed: 2400 MT/s with v4 processors, 2133 with v3. Load the third DIMM on each channel to reach all 24 slots and the whole bank drops a speed tier. That is the trade: 16 slots for maximum bandwidth, 24 slots for maximum capacity per dollar.
Our default is 16 DIMMs at full speed. Virtualization, databases, and general compute feel memory bandwidth more than they feel the last few slots.
Go to all 24 when capacity is the whole point: a budget consolidation host where 50 percent more RAM at a lower clock beats less RAM at a higher one. That is a real workload profile, and this platform serves it cheaply.
Whichever count you choose, buy it as one matched set, now. The instinct to leave slots open for a later upgrade rarely plays out well: a year from now you are hunting for DIMMs that match the rank and speed of what is in the box, and a mismatch either clocks the whole server down or refuses to train. The server is in production by then, so adding memory means a maintenance window you have to schedule and justify.
Common landing spots: 128 GB (16x 8 GB) for general duty, 256 GB (16x 16 GB) for a serious virtualization host at full speed, 384 GB (24x 16 GB) when capacity outranks bandwidth, and 512 GB (16x 32 GB) for database work.
4 RAID Controllers Optional
Smart Array P440ar (2 GB flash-backed cache) is the configurator default and our pick for most builds. The cache is what keeps write performance respectable on parity RAID, and on a loaded chassis it is the difference between an array that performs and one that merely functions.
Smart Array P840ar (4 GB flash-backed cache) is the step up for write-heavy duty on a full backplane: big parity arrays, database volumes, and the 24-bay chassis doing real work. Double the cache earns its price exactly there.
Smart Array H240ar is the no-cache option with a genuinely useful trick: it runs in HBA mode, passing drives straight through. That makes it the correct pick for two very different builds, a simple boot mirror on a SAN-attached host, and software-defined storage, vSAN, Storage Spaces Direct, Ceph, and ZFS all want the direct drive access HBA mode provides. Do not put it in front of a parity array you care about; that is what the cached controllers are for.
Array guidance regardless of controller: RAID 10 where write latency matters, RAID 6 over RAID 5 on large spinning drives, and a hot spare on any array whose rebuild you would rather not race.
5 Storage Drives Select up to 8 drives (0/8 Slots Used) Optional
The 2.5" DL380 Gen9 takes SAS and SATA drives in both flavors: solid state and spinning. Our guidance by workload: SSDs for anything latency-sensitive (virtualization datastores, databases, application volumes), 10K SAS spinning drives where capacity per dollar matters more than IOPS, and a mix is often the honest answer, fast tier up front, bulk behind it.
Buy new. We steer you away from refurbished on drives specifically, and we say that as a company that sells refurbished servers all day. The chassis, board, and power supplies in a refurbished DL380 have proven themselves through our burn-in; a used drive with unknown hours is a different bet, and a rebuild is exactly when you do not want to lose a second one.
The Gen9 platform predates a dedicated boot device, so plan your boot volume out of your data bays: a small RAID 1 pair of SSDs in two bays is the standard pattern. It costs you two bays and it is worth it; a boot failure and a data failure should never be the same event.
Fair warning on capacity math: RAID overhead, hot spares, and filesystem headroom mean usable space lands well under the sum of the labels. Size for the usable number, and when in doubt, one drive size up is cheap insurance against a forklift upgrade later.
6 Remote Access Required
Every unit includes the integrated iLO controller with the standard feature set: health monitoring, remote power control, and basic management. The dividing line is the Advanced license, which unlocks the full remote console and virtual media, meaning you can reach the server's screen, mount an ISO, and rebuild an OS from anywhere, even when the operating system is down. Without it, any problem below the OS level means walking to the rack.
Our pick is the Advanced license, and on a Gen9 server the case is stronger, not weaker, than on new hardware. These servers typically go into branch racks, colos, labs, and secondary roles, places nobody visits on a schedule, and the license costs a fraction of one site trip.
Standard iLO is the right call only when the server sits down the hall and someone can physically reach it without ceremony.
One practical note: iLO 4's interface is a product of its era. It works well, we use it daily, and later firmware brought a solid HTML5 console, so keep the firmware current and it serves fine.
If you are deploying 5 or more units, take the Advanced license on all of them. Fleet management without remote console access does not work in practice.
7 Power Supply 2x included in build Required
2x 500W is the configurator default and covers a real share of DL380 builds: dual mid-range CPUs, a moderate memory load, and a bay of drives land inside it. HPE's Gen9 supplies are famously efficient at partial load, which is part of why this modest pair serves so many builds honestly.
2x 800W is our pick the moment the build fills out: higher-TDP processors, a full backplane, or heavy memory. The price step is small and it buys you the 50 percent load target, where a PSU runs at peak efficiency, runs cooler, and lasts longer than one working near its limit. On the 24-bay chassis or the 12-bay spinning-disk build, start here; spinning drives pull their peak current at spin-up, exactly when the server is busiest bringing everything else online.
The 1400W option is for the heaviest corner cases and GPU configurations; if accelerators are in your plans, talk to us before checkout, because power and airflow change together.
Watch the estimated TDP counter at the bottom of the page as you add components to see where your build lands, and when it sits between two tiers, size up; the price difference is a rounding error against the headroom you get back.
8 Network Cards Required
Quad 1 GbE is the default and covers file, print, domain, and light application duty. If the server's clients are desktops on gigabit switching, this is honestly all you need, and it is the cheapest option on the list.
10 Gb becomes the floor the moment this server hosts a busy hypervisor, shared storage, or backup traffic on a deadline, and most DL380s are exactly that server. The dual 10 Gb options are the sweet spot: production on the fast pair, management on the embedded ports.
Then port type, and this is the one that catches people. SFP+ ports take fiber or DAC cables and talk to datacenter-grade switches. BASE-T ports are 10 Gb copper on ordinary RJ45. Neither is better; they must simply match your switch. Order SFP+ against a copper-only switch and the server racks up with nothing to plug into.
Need more ports than the FlexibleLOM provides? PCIe network cards stack on top of whatever you choose here. Call 1-800-778-1545 and we will spec the whole path, switch ports included.
9 Operating System Required
First, an honest compatibility note for this platform. Windows Server 2025 is not certified for this processor generation, and current VMware ESXi releases no longer support it. A Gen9 server is a value platform, and the right OS choices lean the same way.
Windows Server 2022 is the version to buy. Standard licensing is per core with a 16-core base, which is why our processor guidance steers Windows buyers to two 8-core CPUs: the 16-Core license covers that build exactly. Past 16 total cores, you need the 24-Core option. On CALs, count your users and round up.
Fair warning on Evaluation Editions: free and fully functional for 180 days, then they expire. Right for labs and proof-of-concept, wrong for production.
VMware: we will install ESXi on request for hosts joining an existing licensed cluster, but we do not recommend starting a new VMware deployment on this platform given the support status.
Proxmox VE is our quiet favorite for the DL380 Gen9: open-source KVM virtualization with clustering and live migration, no per-core license math, and full current support on this hardware. It turns a value chassis into a genuinely modern hypervisor.
Ubuntu Server LTS covers Linux application duty with five years of free updates and no licensing conversation at all.
No OS is there for teams imaging from their own deployment infrastructure, including TrueNAS builds, which the 12-bay chassis with the H240ar in HBA mode handles exceptionally well.
Not sure which fits? Tell us the workload at quote time, or call 1-800-778-1545.
WServer Warranty
HPE ProLiant DL380 Gen9 8-Bay 2.5" Drives Detailed Review
The refurbished HPE ProLiant DL380 Gen9 8-Bay 2.5" is the compute-focused member of the DL380 Gen9 family - the eight-bay variant of HPE's Gen9 dual-socket 2U mainstream platform, with eight 2.5" SFF hot-swap bays in the standard chassis. It runs the same Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) processors on the Grantley platform with the C610 chipset, the same 24 DDR4 DIMM slots and 3 TB memory ceiling, the same modular Smart Array storage controllers, the same FlexibleLOM networking, and the same iLO 4 management as the rest of the Gen9 line. What is different is the storage footprint: eight bays is the sweet spot for compute-driven workloads where primary data lives on SAN, NFS, or a distributed file system and local storage handles the OS, application binaries, and modest hot-data staging.
Within the family, the sixteen-bay configuration is the mainstream SFF default; this eight-bay build is for deployments that do not need that much local storage and would rather not pay for bays that sit empty. For family-level positioning and the cross-vendor comparison, the DL380 Gen9 16-Bay 2.5" is the primary page. This page carries the full platform detail in its own right and focuses on what is specific to the eight-bay variant: when eight SFF bays is the right tool, the bay-count-driven workload patterns, and the storage decisions that change at eight versus sixteen or twenty-four.
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.
When 8 SFF Bays Is the Right Default
The 8-Bay DL380 Gen9 fits most production dual-socket Gen9 workloads where storage is networked rather than local. The bay-count decision framework:
- VM cluster nodes with SAN-backed datastores. Boot plus a small local cache or VM tier lives on the eight bays; primary datastores sit on shared FC or iSCSI SAN reached through an FC HBA in PCIe expansion.
- Application servers with modest local data. The OS plus four to six SSDs for application data, logs, or staging, with the database backend on the network. Web, app, and middleware tiers where local storage is supplementary.
- Mid-tier SQL Server or Oracle hosts with networked primary storage. Local SSDs cover the OS, tempdb or Oracle Grid binaries, and transaction logs while the datafiles live on SAN.
- Domain controllers, file servers, and infrastructure services. Modest local capacity is sufficient, delivered at materially lower acquisition cost than the 16- or 24-bay builds.
- Branch office and ROBO deployments. Dual-socket compute with modest local storage for branch file services, AD/DNS, and application hosting; eight bays matches the typical ROBO footprint.
- Lab, dev, and staging environments. Mirroring Gen9 production at lower per-node cost, where eight bays is sufficient for non-production patterns.
If the workload genuinely needs more than eight bays of local storage (VDI hosts, HCI nodes, database hosts with local primary storage, mid-size file servers), step to the DL380 Gen9 16-Bay 2.5" for the SFF sweet spot or the DL380 Gen9 24-Bay 2.5" for maximum SFF density. If eight is sufficient, the 8-Bay delivers the same Gen9 platform at meaningful cost savings.
Storage - 8 SFF Bays
Eight 2.5" SAS/SATA hot-swap bays in the standard chassis configuration (Box 1 populated). The DL380 Gen9 chassis is upgradeable to 16, 18, or 24 SFF in the field via additional drive-cage kits (Box 2 and Box 3 plus the appropriate backplane), but the eight-bay starting point is the right answer when the workload does not anticipate needing more local storage. The optional Universal Media Bay (PN 724865-B21) is available on the eight-bay chassis and provides front VGA, 2x USB 2.0, and optional 2 SFF rear bays plus optical-drive support - useful for deployments that need front-panel management ports.
Drive options span the full Gen9 SFF portfolio: SAS SSDs in mixed-use and read-intensive endurance tiers (200 GB through 3.84 TB at Gen9 launch, larger capacities on later firmware), SATA SSDs for cost-optimized OS and bulk-storage roles, 10K and 15K SAS HDDs for moderate-IOPS data (300 GB through 2.4 TB at SFF), self-encrypting drive variants for compliance, and NVMe SSDs via the Express Bay option in Box 1 (consuming bay count for NVMe positions).
Common DL380 Gen9 8-Bay storage profiles:
- VMware cluster node, SAN datastore primary. 2x SSDs RAID 1 ESXi boot, six bays for a vSAN cache tier or local-VM datastore. Primary VM storage on shared FC/iSCSI SAN.
- Hyper-V cluster node with CSV cache. 2x SSDs for Windows Server, six SSDs for a CSV cache tier. Primary VM storage on SOFS or SAN.
- Application server with local logs/staging. 2x SSDs RAID 1 OS, 4-6 SSDs in RAID 5/10 for app data and logs. Primary data backend on a networked database.
- SQL Server with networked datafiles. 2x SSDs OS, 2x SSDs tempdb mirror, 4x SSDs log files in RAID 10. Primary database files on a SAN datastore.
- Domain controller / infrastructure services. 2x SSDs RAID 1 for OS and AD/DNS/DHCP roles, the remaining bays for supplementary storage or left unused.
- Branch office multi-role server. 2x SSDs OS, six SAS HDDs in RAID 6 for branch file shares. Cost-optimized branch deployment with single-server compute and storage.
Boot Drives
On the eight-bay configuration, boot-drive placement matters more than on 16- or 24-bay builds because consuming two of eight bays for the OS is a meaningful 25% of the storage budget. Three options:
- Rear-bay 2 SFF kit. Preserves all eight front bays for data; the OS lives in the rear bays. The right pattern when front-bay capacity matters.
- M.2 SATA SSDs via the HPE M.2 enablement card. M.2 boot in a PCIe slot, freeing all front bays for data. Consumes a PCIe slot but preserves the storage budget.
- 2x front-bay SSDs in RAID 1. The simplest configuration; consumes two of eight front bays. Acceptable when the six remaining bays cover the workload's data requirement.
We default to the rear-bay 2 SFF kit on 8-Bay DL380 Gen9 quotes when front-bay capacity is at all constrained, and to the standard front-bay RAID 1 pattern when the workload comfortably fits in six bays of data.
Storage Controllers
The controller decision does not change with bay count - only the cache-sizing tradeoff does. The same modular Smart Array "ar" controllers used across the family apply here:
- Smart Array P440ar (2 GB FBWC). The mainstream production controller for the eight-bay build. 2 GB of flash-backed write cache is comfortably sized for eight SSDs under most workload patterns, and it is the right pick for traditional hardware RAID in production.
- Smart Array H241 (HBA mode, PCIe plug-in). For software-defined storage (vSAN, S2D, Ceph, ZFS). Clean SAS pass-through with no hardware RAID.
- Smart Array P840ar (4 GB FBWC). The premium controller. Rarely needed at eight bays - the P440ar's 2 GB cache is sufficient for this storage scale.
- Dynamic Smart Array B140i (embedded software RAID). Acceptable for OS boot mirroring only; not appropriate for production data.
The HPE Smart Storage Battery is required with any P-series controller. The Gen9 FBWC battery is a wear item with a documented 5-7 year service life; we disclose battery state on every quote and replace cache modules that are past spec as part of build prep.
Processors
1 or 2 sockets of Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the C610 Grantley chipset. Mixing v3 and v4 is not supported - all installed CPUs must be the same generation, though a field upgrade from v3 to v4 (replacing both at once) is supported. Single-socket builds cut DIMM slots in half (12 instead of 24) and PCIe to three slots, so 2-socket is the production standard. The eight-bay chassis shares the same 2U thermal envelope as the higher-bay variants, and its lower drive count leaves a little more power and airflow headroom for top-bin CPUs.
- E5-2680 v4 (14 cores, 120W, DDR4-2400). The Gen9 production mainstream - 28 cores at 2S, balanced TDP, standard heatsink.
- E5-2690 v4 (14 cores, 135W, 2.6 GHz). Higher base frequency for single-thread-sensitive workloads within the same core budget.
- E5-2699 v4 (22 cores, 145W). Top-bin Broadwell-EP - 44 cores at 2S, the platform maximum. Requires the high-performance heatsink (auto-included for 120W+ CPUs).
- E5-2650 v4 (12 cores, 105W). Mid-tier production at modest TDP and lower acquisition cost - a good fit for general virtualization and application servers.
- E5-2620 v4 (8 cores, 85W) and E5-2667 v4 (8 cores, 135W, 3.2 GHz). Entry-tier and high-frequency specialty SKUs; the 2667 v4 is the per-core-licensing pick for Oracle and SQL Server Enterprise. Haswell-EP v3 equivalents are available at lower cost with a DDR4-2133 cap.
Memory
24 DDR4 DIMM slots (12 per CPU; only 12 are available with a single CPU). RDIMM and LRDIMM are supported but cannot be mixed in one server; the maximum is 3 TB with 128 GB LRDIMMs across all 24 slots on v4 CPUs. HPE DDR4 Smart Memory is required for rated speeds - third-party DDR4 drops to lower speeds, documented HPE behavior across Gen9.
Memory speed depends on CPU generation and population: v3 caps at DDR4-2133, v4 at DDR4-2400, and full 24-DIMM population drops to DDR4-1866 or DDR4-1600 depending on rank. For maximum bandwidth, populate at 1 DPC (12 DIMMs at 2S). HPE Persistent Memory (NVDIMM-N, 8 GB and 16 GB) is supported on v4 CPUs for DRAM-class latency with battery-backed persistence - uncommon, but available for SQL Server transaction logs and in-memory database WAL.
Networking and PCIe Expansion
The embedded HPE 4-port 1 GbE 331i adapter is standard and consumes no slot. The optional FlexibleLOM mezzanine supports 10 GbE SFP+ (530FLR/534FLR), 10 GBASE-T, 25 GbE SFP28, and converged FlexFabric. Unlike the DL580 Gen9, Wake-on-LAN works on both the embedded 1 GbE and the FlexibleLOM here. PCIe expansion is three PCIe Gen3 slots with one CPU, expanding to six with both CPUs populated; the +3-slot secondary riser requires the second processor. All slots are PCIe Gen3 and accept cards up to 150W, higher with the supplemental power-cable kit. On a SAN-backed eight-bay build, an FC HBA or a 10/25 GbE FlexibleLOM is usually the first expansion priority.
GPU Support
GPU and accelerator support is bounded by the PCIe Gen3 generation and the 2U thermal envelope. The eight-bay build's lower drive count can leave a little more PSU headroom for accelerators, but the slot and thermal limits are the same as the rest of the family:
- Single-width accelerators. Cards like the NVIDIA Tesla T4 (70W, single-slot, passive) for inference, transcoding, or VDI graphics offload. They fit standard riser positions and need no GPU power-cable kit.
- Double-width GPUs. Passively cooled Gen9-era cards (NVIDIA M40, M60, K80-class). These require the high-performance heatsink and an additional GPU power-cable kit (PN 669777-B21); plan for up to two, subject to PSU sizing.
- Thermal envelope. GPU builds require performance heatsinks and the high-performance fan kit, and ASHRAE A3/A4 ambient headroom is reduced with double-wide cards. We validate inlet temperature against the configuration at quote time.
- FPGA and specialty cards. The PCIe Gen3 x16 slots accept FPGA and specialty cards within the 150W per-slot limit. PCIe Gen3 bandwidth is the ceiling - workloads needing PCIe Gen4 GPU bandwidth belong on Gen10 Plus or Gen11.
Management - iLO 4 Generation
The DL380 Gen9 ships with HPE iLO 4: remote console (an iLO Advanced license enables full graphical KVM), virtual media, IPMI, SNMP telemetry, Active Health System logging, and HPE OneView compatibility - the same iLO 4 generation across the Gen9 line, which is part of the platform's operational-standardization value. The key difference from Gen10 is that iLO 4 has no Silicon Root of Trust; the hardware-anchored firmware-verification chain arrived with iLO 5 on Gen10. UEFI Secure Boot is supported and is the right pattern for production Gen9 builds, with compensating controls where a compliance framework requires firmware-integrity attestation. iLO Advanced is typically a separate cost and is rarely optional for production data-center deployments; we quote it explicitly.
Power and Cooling
The same HPE Flex Slot power supplies as the rest of the DL380 Gen9 family - 500W, 800W, or 1400W Platinum/Titanium in 1+1 redundant configurations, plus the optional HPE Flexible Slot Battery Backup module. The eight-bay configuration's lower drive count means lower total power draw than the 16- or 24-bay variants: 500W PSUs are adequate for many eight-bay builds, and 800W in 1+1 covers all common dual-socket configurations including E5-2680/2690 v4 with full memory.
For high-TDP CPUs (E5-2699 v4 at 145W, E5-2667 v4 at 135W) or builds with double-wide GPUs, 1400W PSUs are required. We run the HPE Power Advisor against every DL380 Gen9 quote to validate PSU sizing. Thermal: ASHRAE A3 (40 C) and A4 (45 C) extended-ambient operation is supported with the performance heatsinks (auto-included for 120W+ CPUs).
Physical Specs & Platform Notes
- Form factor: 2U rackmount, standard-depth Gen9 enclosure shared across the DL380 Gen9 bay-count variants; with the cable management arm installed, plan for additional rear clearance.
- PCIe expansion: up to six PCIe Gen3 slots with both CPUs populated (three with one CPU), split full-height and low-profile across the primary and secondary risers; the secondary riser requires the second processor.
- Parts availability: excellent. The DL380 Gen9 shipped in one of the largest install bases of any 2U generation, so drives, PSUs, risers, heatsinks, FlexibleLOM cards, and Smart Array controllers are widely available, and third-party maintenance spares depth is strong in major metros.
- Accessories we recommend: the 2U SFF ball-bearing sliding rail kit (P/N 679365-001 / 737412-001; see the DL380 G8/G9 2U SFF sliding rail kit), the optional Universal Media Bay (PN 724865-B21) for front VGA and USB, the rear-2-SFF kit for boot placement, and the GPU power-cable kit (PN 669777-B21) on accelerator builds.
- Platform notes: CPU hot-plug is not supported, and v3/v4 CPUs cannot be mixed. NVMe via the Express Bay option consumes specific front-bay positions, which is a tighter tradeoff at eight bays than at sixteen or twenty-four. Confirm FlexibleLOM and drive-backplane compatibility against the specific build at quote time.
Our Assessment
Where it excels: The eight-bay DL380 Gen9 is the right answer for compute-driven dual-socket workloads where primary storage is networked. It is a strong fit for VM cluster nodes with SAN-backed datastores, application and middleware servers with modest local data, mid-tier SQL Server and Oracle hosts whose datafiles live on SAN, domain controllers and infrastructure services, branch-office and ROBO deployments, and lab or staging environments mirroring Gen9 production. When the workload does not need a large local SSD pool, paying for sixteen or twenty-four bays is wasted budget, and the eight-bay build delivers the same Gen9 platform for less.
Where to look instead: If the workload needs a large local storage tier, the DL380 Gen9 16-Bay 2.5" is the SFF sweet spot and the DL380 Gen9 24-Bay 2.5" is the maximum-density option; for bulk HDD capacity, the DL380 Gen9 12-Bay 3.5" is purpose-built. New mission-critical deployments that need iLO 5 Silicon Root of Trust, PCIe Gen4, or DDR4-2933+ bandwidth should move to the DL380 Gen10 16-Bay 2.5". Budget-driven 2U Gen9 deployments that can trade SFF for LFF should compare the HPE ProLiant DL180 Gen9 LFF value tier. Dell-standardized shops should compare the Dell PowerEdge R730 8-Bay 2.5", the equivalent 2U Grantley platform at the same bay count.
Bottom line: The eight-bay DL380 Gen9 is the cost-disciplined member of the family - the build you choose when the compute is the point and the storage lives on the network. The typical customer is an IT team adding SAN-backed VM cluster nodes to an existing Gen9 estate, standing up application or infrastructure servers, or deploying branch and ROBO sites where local capacity is modest. Buy it when eight bays genuinely cover the local-storage requirement; step up to the 16- or 24-bay companions the moment a large local SSD pool is in the picture, and step to Gen10 when current-generation security and memory bandwidth matter.
Honest Limitations
- Same Gen9 platform limitations as the rest of the family. HPE active warranty has ended; iLO 4 has no Silicon Root of Trust; DDR4 speed caps at DDR4-2400 (v4) or DDR4-2133 (v3) and drops further under full DIMM population; PCIe Gen3 only; the FBWC battery is a wear item; v3/v4 CPU mixing is not supported; and HPE Smart Memory is required for rated speeds.
- Boot-drive consumption hurts more at eight bays. Two bays for an OS RAID 1 mirror is 25% of the budget, so rear-bay or M.2 boot is strongly preferred on builds where data capacity matters.
- Future expansion to 16 or 24 bays requires backplane and cage kits. The field upgrade is supported but not trivial - if the workload may grow into more bays within the platform's service life, start at the 16- or 24-bay variant to avoid the expansion exercise.
- NVMe consumes bay count. The Express Bay NVMe option occupies physical front-bay positions, so on the eight-bay build the storage budget tightens further when NVMe-tier performance is required.
Workload Fit
| This server is right for | Consider alternatives for |
|---|---|
| ✅ VM cluster nodes with SAN-backed datastores | ❌ VDI hosts requiring SFF-bay-heavy storage (use 16-Bay) |
| ✅ Application servers and middleware tier | ❌ HCI nodes needing a high local drive count (use 16- or 24-Bay) |
| ✅ Database hosts with networked primary storage | ❌ Database hosts with local primary storage (use 16- or 24-Bay) |
| ✅ Domain controllers and infrastructure services | ❌ Workloads requiring more than eight bays at Gen9 |
| ✅ Branch office and ROBO deployments | ❌ New mission-critical deployments needing iLO 5 |
| ✅ Lab/dev/staging mirroring Gen9 production | ❌ Memory-bandwidth-sensitive workloads (Gen10+) |
Where to Look Instead
- Need more SFF bays at Gen9? → DL380 Gen9 16-Bay 2.5" - the SFF sweet spot for VDI, HCI, and database hosts with local SSD.
- Need maximum SFF density at Gen9? → DL380 Gen9 24-Bay 2.5" - 24 SFF bays at Gen9.
- Need LFF (3.5") drives for bulk capacity? → DL380 Gen9 12-Bay 3.5" - high-capacity NL-SAS HDD pool.
- Want a lower-cost 2U Gen9 value tier? → HPE ProLiant DL180 Gen9 LFF - cost-optimized 2U dual-socket Gen9.
- Need Gen10 with iLO 5, DDR4-2933, and Silicon Root of Trust? → DL380 Gen10 16-Bay 2.5" - current-generation 2U.
- Dell shop alternative at the same Gen9 dual-socket 2U tier? → Dell PowerEdge R730 8-Bay 2.5" - 2U 2S Grantley, equivalent positioning.
- Mounting hardware? → DL380 G8/G9 2U SFF sliding rail kit (P/N 679365-001 / 737412-001).
Ready to Configure?
Tell us the workload, CPU generation preference (v3 vs v4), memory target, storage configuration (drive types, RAID layout, controller preference, boot pattern), networking requirement (embedded 1 GbE vs FlexibleLOM), PSU configuration, and quantity. We respond within 24 hours with a validated configuration including HPE Power Advisor sizing and third-party maintenance coordination when requested. 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 or use the quote form below.
Save Your Design
Click the Add to Quote button at the bottom of your screen to save your design as a draft order for future reference and to check for discounts, lead time, and availability. Most servers ship within 1-3 days.
