{"title":"Proliant DL360 Gen 9","description":"\u003cp data-end=\"490\" data-start=\"149\"\u003eThe \u003cstrong data-end=\"180\" data-start=\"153\"\u003eHPE ProLiant DL360 Gen9\u003c\/strong\u003e is a powerful and reliable \u003cstrong data-end=\"226\" data-start=\"208\"\u003e1U rack server\u003c\/strong\u003e designed for businesses, IT professionals, and home lab environments. Powered by \u003cstrong data-end=\"351\" data-start=\"308\"\u003eIntel Xeon E5-2600 v3 and v4 processors\u003c\/strong\u003e, it delivers strong multi-core performance for workloads such as \u003cstrong data-end=\"489\" data-start=\"417\"\u003evirtualization, web hosting, file storage, and database applications\u003c\/strong\u003e.\u003c\/p\u003e\n\u003cp data-end=\"836\" data-start=\"492\"\u003eSupporting up to \u003cstrong data-end=\"538\" data-start=\"509\"\u003e1.5TB of DDR4 SmartMemory\u003c\/strong\u003e and flexible storage with \u003cstrong data-end=\"594\" data-start=\"565\"\u003eSATA, SAS, or NVMe drives\u003c\/strong\u003e, the DL360 Gen9 offers excellent scalability in a compact form factor. With built-in \u003cstrong data-end=\"711\" data-start=\"680\"\u003eHPE iLO 4 remote management\u003c\/strong\u003e, PCIe expansion, and hardware RAID options, it provides the performance and reliability needed for modern IT infrastructure.\u003c\/p\u003e\n\u003cp data-end=\"1031\" data-start=\"838\"\u003eAt \u003cstrong data-end=\"862\" data-start=\"841\"\u003eWholesale Servers\u003c\/strong\u003e, we provide professionally tested \u003cstrong data-end=\"933\" data-start=\"897\"\u003erefurbished HPE ProLiant servers\u003c\/strong\u003e ready to power your \u003cstrong data-end=\"1000\" data-start=\"954\"\u003edata center, business network, or home lab\u003c\/strong\u003e with fast nationwide shipping.\u003c\/p\u003e","products":[{"product_id":"dl360-g9-3-5-4-bay-chassis","title":"HPE ProLiant DL360 Gen9 4-Bay 3.5\" Drives [Gen9]","description":"\u003cp\u003eThe refurbished HPE ProLiant DL360 Gen9 4-Bay 3.5\" is the large-form-factor (LFF) member of HPE's Gen9 1U dual-socket line: four 3.5\" SAS\/SATA hot-swap bays in a 1U chassis. It carries the same Intel Grantley platform as the rest of the family, Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the C610 chipset, 24 DDR4 DIMM slots with a 3 TB ceiling, HPE modular Smart Array controllers, the FlexibleLOM mezzanine, and iLO 4 management. What is specific to this variant is the form factor: four LFF bays in 1U is a deliberately narrow combination that pairs high rack density with bulk-capacity HDD storage for edge, branch, ROBO, and backup roles where both rack space and capacity matter.\u003c\/p\u003e\u003cp\u003eThis page covers what changes on the LFF chassis: the workloads where 1U LFF is the right tool, the RAID math at four large-capacity drives, and the storage decisions that differ from the SFF configurations. The shared compute, memory, networking, and management platform is the same across every DL360 Gen9 chassis; for the full platform reference, the \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the primary page for the family.\u003c\/p\u003e\u003cp\u003eTo configure a build, call 1-800-778-1545 or use the quote form below. Every refurbished unit ships under our 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units.\u003c\/p\u003e\u003chr\u003e\u003ch2\u003eWhen 1U LFF Is the Right Combination\u003c\/h2\u003e\u003cp\u003e1U LFF is a narrow configuration on purpose. Most LFF workloads run on 2U or 4U platforms with twelve or more bays for capacity scaling, and most 1U workloads run SFF for performance per bay. The DL360 Gen9 4-Bay 3.5\" earns its place only when both 1U rack density and large-format HDD capacity are genuine requirements at the same time:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eEdge compute with bulk local storage.\u003c\/strong\u003e Edge sites running analytics preprocessing, IoT collection, or video work that also need local capacity for staging before central upload. Four 8 TB to 12 TB NL-SAS drives deliver roughly 24 TB to 36 TB raw in 1U of rack space.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eBranch office multi-role servers.\u003c\/strong\u003e A single 1U box running AD, DNS, DHCP, file shares, print, and modest local virtualization. Four LFF bays in RAID 6 or RAID 10 cover branch-scale file capacity without a separate storage chassis.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eROBO file servers at 1U density.\u003c\/strong\u003e Regional and satellite offices with hard rack-space limits, where a 1U compute-plus-storage box covers most branch infrastructure on one server.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRemote backup targets.\u003c\/strong\u003e Backup repositories at sites where a 2U-or-larger storage box does not fit the rack envelope. Four high-capacity NL-SAS drives in RAID 6 hold a useful branch retention window.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmall to mid-scale surveillance NVR.\u003c\/strong\u003e Camera counts that fit within four LFF drives of retention, where the recorder shares the rack with other 1U gear.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eLab, dev, and test bulk storage.\u003c\/strong\u003e Lower-priority capacity where 1U cost and density matter more than scale.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eIf four LFF bays are enough and 1U density is the constraint, this configuration does real work. If bulk capacity is the primary driver and 1U is not required, the \u003ca href=\"\/products\/hp-proliant-dl380-g9-12-bay-3-5-chassis\"\u003eDL380 Gen9 12-Bay 3.5\"\u003c\/a\u003e at 2U is the better fit, with three times the bay count for twice the rack space.\u003c\/p\u003e\u003ch2\u003eStorage - 4 LFF Bays\u003c\/h2\u003e\u003cp\u003eFour 3.5\" SAS\/SATA hot-swap bays sit across the front of the 1U chassis. NVMe options are limited on the LFF chassis; the Express Bay NVMe support lives on the SFF DL360 Gen9 variants. When a 1U LFF build needs some flash alongside HDD capacity, the practical patterns are LFF SSDs in 3.5\" carriers or an M.2 SATA SSD on a PCIe enablement card for the boot and high-performance tier. At full population, four LFF bays reach roughly 56 TB raw with 14 TB drives, before RAID overhead.\u003c\/p\u003e\u003cp\u003eThe drive portfolio spans the full Gen9 LFF range:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eNL-SAS HDDs.\u003c\/strong\u003e The bulk-capacity workhorse: 4 TB through 14 TB MDL drives at 7,200 RPM, optimized for sequential throughput and capacity per dollar.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e10K and 15K SAS HDDs.\u003c\/strong\u003e Higher per-drive IOPS than NL-SAS at a lower capacity ceiling: 10K LFF tops near 2.4 TB, 15K LFF near 900 GB.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eLFF SSDs.\u003c\/strong\u003e SAS or SATA SSDs in 3.5\" carriers. Rarely the efficient choice on form factor alone, but valid when LFF is locked in for chassis reasons.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSelf-encrypting drives (SED).\u003c\/strong\u003e Available for compliance-regulated bulk storage.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eRAID at 4 LFF\u003c\/h3\u003e\u003cp\u003eRAID layout at four LFF drives works differently than at twelve, because the small bay count changes the overhead math:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eRAID 10 (two mirrored pairs, striped).\u003c\/strong\u003e 50 percent overhead, roughly 24 TB usable with 12 TB drives. Our usual default at exactly four drives: the overhead is identical to RAID 6 but write performance is better and the rebuild scope is smaller.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRAID 6 (dual parity).\u003c\/strong\u003e Same 50 percent overhead at four drives, roughly 22 TB usable with 12 TB drives. Preferred when read capacity matters more than write speed; dual parity protects against a second drive failing during the multi-day rebuild that large LFF drives require.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRAID 5 (single parity).\u003c\/strong\u003e 25 percent overhead, roughly 36 TB usable with 12 TB drives, but not recommended at LFF capacity: single parity across multi-day rebuild windows is high risk. We will quote it on explicit request and flag the risk in writing.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eJBOD \/ HBA pass-through.\u003c\/strong\u003e When the storage abstraction lives in software (ZFS, software-defined NAS), the H240ar HBA is the right controller.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eBoot Drives\u003c\/h3\u003e\u003cp\u003eThe 4-Bay LFF chassis has a boot problem: two of four bays for a mirrored OS volume is 50 percent of the storage budget, which is rarely acceptable. M.2 SATA on the HPE M.2 SSD enablement card is the answer on essentially every 4-Bay LFF production build, since it puts boot in a PCIe slot and preserves all four LFF bays for data. Front-bay boot is a fallback only when two bays of data are genuinely sufficient, and a single unmirrored boot drive is not a production pattern. We default to M.2 boot on every 4-Bay LFF quote and confirm the PCIe slot allocation accordingly.\u003c\/p\u003e\u003ch2\u003eStorage Controllers\u003c\/h2\u003e\u003cp\u003eThe same HPE modular Smart Array family runs here, with selection skewed toward the cache sizing a four-drive bulk-storage workload actually needs:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P440ar (2 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e The right production controller at four LFF; the 2 GB cache is comfortably sized for a four-drive array. Mounts in the modular slot without consuming a PCIe slot.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array H240ar (HBA \/ pass-through).\u003c\/strong\u003e For software-defined NAS or ZFS file servers that want raw disks rather than hardware RAID.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P840 (4 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e Available for write-intensive needs, but rarely required at four LFF; the P440ar is usually sufficient.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eB140i (embedded software RAID via the chipset).\u003c\/strong\u003e Acceptable for boot mirroring only; not a production data-RAID controller on a dual-socket platform.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eAll P-series controllers require the HPE Smart Storage Battery for write-back caching, and the flash-backed cache module is a wear item we verify and replace as part of build prep on any refurbished unit.\u003c\/p\u003e\u003ch2\u003eProcessors\u003c\/h2\u003e\u003cp\u003eOne or two sockets of Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the Grantley C610 platform, drop-in compatible within the socket but not mixable in one server. Core counts reach 18 per CPU on v3 and 22 per CPU on v4, with TDPs from roughly 55 W to 145 W. Bulk-storage roles rarely need the top bins, so CPU selection on this variant usually skews lower than on a dense SFF compute node:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2620 v4 (8 cores, 85 W).\u003c\/strong\u003e The common branch and edge pick; 16 cores across two sockets is plenty for file and infrastructure roles, and the low TDP sits easily inside the 1U envelope.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2640 v4 (10 cores, 90 W) or E5-2650 v4 (12 cores, 105 W).\u003c\/strong\u003e Mid-tier choices when the box runs compute alongside the storage role.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2680 v4 or higher.\u003c\/strong\u003e Reserved for consolidated branch or edge-analytics builds that pair four-LFF storage with meaningful compute.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eTop-bin CPUs still require the performance heatsink and high-performance fan kit, and a single-socket build exposes only 12 of the 24 DIMM slots and half the PCIe lanes, so the second socket matters whenever memory or expansion does.\u003c\/p\u003e\u003ch2\u003eMemory\u003c\/h2\u003e\u003cp\u003e24 DDR4 DIMM slots, 12 per CPU. The platform ceiling is 3 TB with 128 GB LRDIMMs; RDIMM is the mainstream choice, with 64 GB to 128 GB total covering most branch and edge bulk-storage workloads. ZFS-based file servers are the exception and benefit from more memory (128 GB to 256 GB) for ARC cache. Speed is population-dependent: DDR4-2400 on v4 and DDR4-2133 on v3 at one DIMM per channel, stepping down a tier at full two-DIMM-per-channel population. HPE Smart Memory is required for rated speeds; third-party DIMMs may train slower or fail to post. This platform predates Intel Optane persistent memory, so PMem is not part of the conversation here.\u003c\/p\u003e\u003ch2\u003eNetworking and PCIe Expansion\u003c\/h2\u003e\u003cp\u003eNetworking runs through the FlexibleLOM mezzanine, so the network personality is a build choice that does not cost a PCIe slot. For bulk-storage roles, 10 GbE FlexibleLOM (the 530FLR-SFP+ or 534FLR-SFP+ dual-port adapters) is the sensible default on backup-target and edge-analytics builds, while the 331FLR quad-port 1 GbE is acceptable for a pure branch file server. PCIe expansion is the 1U constraint: up to three PCIe Gen3 slots with both CPUs populated, and on this variant one of them is usually committed to the M.2 boot card. When more expansion than that is needed, the 2U DL380 Gen9 is the right move.\u003c\/p\u003e\u003ch2\u003eGPU Support\u003c\/h2\u003e\u003cp\u003eGPU support in the 1U envelope is limited to single-width, low-profile cards within the chassis power and thermal budget; there is no room for double-wide accelerators. On a bulk-storage LFF build a GPU is rarely part of the spec, but for the edge-analytics case a single low-profile accelerator such as an NVIDIA T-series card fits. Anything heavier (double-wide, multi-GPU, or high per-card power) belongs in the 2U DL380 Gen9 rather than this chassis.\u003c\/p\u003e\u003ch2\u003eManagement - iLO 4 Generation\u003c\/h2\u003e\u003cp\u003eThe DL360 Gen9 uses iLO 4. iLO Standard is included for health monitoring, power control, and basic remote access; iLO Advanced is the licensed tier for full graphical remote console, virtual media, and remote KVM, and it is usually quoted separately, which matters for lights-out edge and branch sites where remote hands are scarce. Intelligent Provisioning handles firmware and driver deployment, and the Active Health System log is the first place to check a refurbished unit's history. On security, iLO 4 predates the Silicon Root of Trust introduced on Gen10; UEFI Secure Boot is the firmware integrity baseline, and a TPM module is available where a hardware root is required.\u003c\/p\u003e\u003ch2\u003ePower and Cooling\u003c\/h2\u003e\u003cp\u003ePower comes from HPE Flex Slot hot-plug supplies in 500 W, 800 W, and 1400 W ratings at Platinum and Titanium efficiency, configured 1+1 for redundancy. Four LFF HDDs draw less storage-tier power than eight to ten SAS SSDs, so an 800 W pair is the standard production choice and 500 W is adequate for an entry-tier branch build on E5-2620-class CPUs. For edge sites without a rack UPS, the HPE Flex Slot battery backup module is worth adding for short power-event ride-through. Thermally the 1U chassis supports ASHRAE A3 (40 C) and A4 (45 C) inlet ranges; LFF HDDs run cooler than dense SSD arrays, and the wider ambient range helps at edge sites without dedicated cooling. We target a 25 C to 30 C inlet on production deployments.\u003c\/p\u003e\u003ch2\u003ePhysical Specs \u0026amp; Platform Notes\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eForm factor:\u003c\/strong\u003e 1U rack chassis, standard depth for four-post racks, with the usual rear allowance for cable management.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePCIe expansion:\u003c\/strong\u003e up to three PCIe Gen3 slots with both CPUs populated, low-profile only; one slot is typically committed to the M.2 boot card on this variant.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eParts availability:\u003c\/strong\u003e excellent. Gen9 LFF drives, controllers, FlexibleLOM adapters, PSUs, and rails are deep on the secondary market. HPE active warranty support has ended, and third-party maintenance is the standard production support path in 2026.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eAccessories we recommend:\u003c\/strong\u003e the \u003ca href=\"\/products\/hp-dl360-g8-g9-sff-1u-sliding-rails-679368-001-728437-001\"\u003eHPE 1U ball-bearing sliding rail kit (P\/N 679368-001 \/ 728437-001)\u003c\/a\u003e for the DL360 Gen9 chassis, a cable management arm, and the HPE M.2 SSD enablement card so boot stays off the four LFF bays.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePlatform notes:\u003c\/strong\u003e M.2 boot is effectively mandatory here to avoid spending half the bays on the OS; CPU hot-plug is not supported; LFF rebuild windows run into days at production load, so plan RAID 10 or RAID 6 rather than RAID 5; and NVMe is not a practical option on the LFF chassis.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eOur Assessment\u003c\/h2\u003e\u003cp\u003e\u003cstrong\u003eWhere it excels:\u003c\/strong\u003e This is the configuration for the genuine 1U-plus-LFF case: edge compute nodes that need local bulk storage, branch and ROBO servers consolidating file and infrastructure roles, small backup targets and surveillance recorders where a 2U storage box will not fit the rack. Four high-capacity NL-SAS drives in RAID 10 or RAID 6, an M.2 boot device, and a low-TDP dual-socket CPU pairing make a capable, rack-efficient branch server.\u003c\/p\u003e\u003cp\u003e\u003cstrong\u003eWhere to look instead:\u003c\/strong\u003e If bulk capacity is the real driver and 1U is not a hard constraint, the \u003ca href=\"\/products\/hp-proliant-dl380-g9-12-bay-3-5-chassis\"\u003eDL380 Gen9 12-Bay 3.5\"\u003c\/a\u003e gives three times the bays at 2U. If SFF SSD performance matters more than LFF capacity, the \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the standard 1U compute build, and the \u003ca href=\"\/products\/dl360-g9-2-5-10-bay-hot-swap-psu\"\u003eDL360 Gen9 10-Bay 2.5\"\u003c\/a\u003e is the densest SFF option. For current-generation 1U LFF with iLO 5, the \u003ca href=\"\/products\/hpe-proliant-dl360-g10-4-bay-3-5-build-your-own-server\"\u003eDL360 Gen10 4-Bay 3.5\"\u003c\/a\u003e is the step up.\u003c\/p\u003e\u003cp\u003e\u003cstrong\u003eBottom line:\u003c\/strong\u003e A niche but useful build for the IT team that has decided, for good rack-space reasons, that bulk storage has to live in 1U. It is the right tool for edge, branch, ROBO, and small backup roles, and the wrong tool the moment capacity scaling or SFF performance becomes the priority. Size the four-bay capacity ceiling against the deployment's growth before committing, because field expansion past four LFF bays means a different chassis.\u003c\/p\u003e\u003ch2\u003eHonest Limitations\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003eOnly four LFF bays: the raw capacity ceiling is roughly 56 TB with 14 TB drives, less after RAID. Capacity-led builds belong on the DL380 Gen9 12-Bay LFF.\u003c\/li\u003e\n\u003cli\u003eBoot consumes 50 percent of the bays in the front, so M.2 boot is effectively mandatory.\u003c\/li\u003e\n\u003cli\u003eLFF rebuild times run into days at production load; RAID 10 or RAID 6 strongly preferred, RAID 5 is high risk at this capacity.\u003c\/li\u003e\n\u003cli\u003eNVMe is not a practical option on the LFF chassis; for NVMe at 1U Gen9, use the SFF 8-Bay or 10-Bay variants.\u003c\/li\u003e\n\u003cli\u003eiLO 4 management, without the Silicon Root of Trust hardware attestation introduced on Gen10.\u003c\/li\u003e\n\u003cli\u003eDDR4 speed caps at DDR4-2400 (v4) or DDR4-2133 (v3) and steps down under full DIMM population; HPE Smart Memory required for rated speeds.\u003c\/li\u003e\n\u003cli\u003ePCIe Gen3 only, three slots in 1U, with one usually committed to the M.2 boot card.\u003c\/li\u003e\n\u003cli\u003eNo double-wide GPU support in 1U; single-width low-profile accelerators only.\u003c\/li\u003e\n\u003cli\u003eHPE active warranty has ended; production support is via third-party maintenance.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eWorkload Fit\u003c\/h2\u003e\u003ctable\u003e\n\u003ctr\u003e\n\u003cth\u003eThis server is right for\u003c\/th\u003e\n\u003cth\u003eConsider alternatives for\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eEdge compute with bulk local storage at 1U\u003c\/td\u003e\n\u003ctd\u003eMore than four LFF bays needed (use DL380 Gen9 12-Bay 3.5\")\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eBranch and ROBO multi-role file servers\u003c\/td\u003e\n\u003ctd\u003eSFF SSD performance as the primary driver (use the 8-Bay or 10-Bay 2.5\")\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eRemote backup targets where 1U is required\u003c\/td\u003e\n\u003ctd\u003eNew deployments needing iLO 5 and Silicon Root of Trust\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eSmall to mid-scale surveillance NVR at 1U\u003c\/td\u003e\n\u003ctd\u003eWorkloads needing PCIe Gen4 NVMe bandwidth\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eLab, dev, and test bulk storage at 1U density\u003c\/td\u003e\n\u003ctd\u003eMore than roughly 56 TB raw at 1U\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/table\u003e\u003ch2\u003eWhere to Look Instead\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed more LFF bays?\u003c\/strong\u003e The \u003ca href=\"\/products\/hp-proliant-dl380-g9-12-bay-3-5-chassis\"\u003eDL380 Gen9 12-Bay 3.5\"\u003c\/a\u003e is the 2U LFF platform for backup, file serving, and archive at scale.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed SFF performance in 1U Gen9?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the standard 1U compute build, and the \u003ca href=\"\/products\/dl360-g9-2-5-10-bay-hot-swap-psu\"\u003eDL360 Gen9 10-Bay 2.5\"\u003c\/a\u003e is the densest SFF option.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed current-generation 1U LFF with iLO 5?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl360-g10-4-bay-3-5-build-your-own-server\"\u003eDL360 Gen10 4-Bay 3.5\"\u003c\/a\u003e is the direct generational step up.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWorking to the tightest budget on a 1U Gen9 LFF build?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl160-gen9-4-bay-lff-build-your-own\"\u003eDL160 Gen9 4-Bay 3.5\"\u003c\/a\u003e is the value-tier 1U Gen9 step down.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eStandardized on Dell?\u003c\/strong\u003e The \u003ca href=\"\/products\/dell-poweredge-r630-10-bay-chassis\"\u003eDell PowerEdge R630 10-Bay 2.5\"\u003c\/a\u003e is the equivalent 1U dual-socket Grantley platform from the same generation.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eReady to Configure?\u003c\/h2\u003e\u003cp\u003eTell us the workload (edge compute, branch file server, ROBO, remote backup, or surveillance NVR), the drive capacity target (8, 10, 12, or 14 TB), the RAID layout (RAID 10 or RAID 6 are the usual picks at four drives), the controller preference, the M.2 boot configuration, the FlexibleLOM choice (1 or 10 GbE is typical here), the PSU configuration, and the quantity. We respond within 24 hours with a validated configuration including drive-capacity verification, RAID sizing math, and HPE Power Advisor sizing. Every refurbished unit ships with the Wholesale Servers 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units. Call 1-800-778-1545 or use the quote form below.\u003c\/p\u003e","brand":"HPE","offers":[{"title":"Default Title","offer_id":45951242764487,"sku":"BP-013615","price":412.24,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0748\/4493\/0247\/files\/server-design-lab-hpe-proliant-dl360-gen9-4-bay-35-drives-188542.png?v=1765539623"},{"product_id":"hpe-proliant-dl360-g9-8-bay-2-5-chassis","title":"HPE ProLiant DL360 Gen9 8-Bay 2.5\" Drives [Gen9]","description":"\u003cp\u003eThe refurbished HPE ProLiant DL360 Gen9 8-Bay 2.5\" is the standard SFF configuration of HPE's Gen9 1U dual-socket platform, and the build we treat as the default DL360 Gen9 for general production. Eight 2.5\" SAS\/SATA hot-swap front bays sit ahead of the Intel Grantley platform: Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) processors on the C610 chipset, 24 DDR4 DIMM slots with a 3 TB memory ceiling, HPE modular Smart Array storage controllers, the FlexibleLOM network mezzanine, and iLO 4 out-of-band management. Eight bays is the right-sized storage footprint for the large majority of 1U dual-socket Gen9 workloads, where primary storage is networked and local disk handles boot, cache, and logs.\u003c\/p\u003e\u003cp\u003eThis is the full platform reference for the DL360 Gen9: processors, memory, storage controllers, networking, management, power, and an honest read on where a 2014-era platform belongs in a 2026 procurement decision. Where a build needs more than eight bays in the same 1U chassis, the 10-Bay configuration is the maximum-SFF option; where the workload wants large-format capacity, the 4-Bay 3.5\" covers it; and where eight bays of local storage are not enough, the 1U form factor is the wrong tool and the 2U DL380 Gen9 is the better answer. Each alternative is linked in context below.\u003c\/p\u003e\u003cp\u003eTo configure a build, call 1-800-778-1545 or use the quote form below. Every refurbished unit ships under our 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units.\u003c\/p\u003e\u003chr\u003e\u003ch2\u003eWhere the DL360 Gen9 Fits in the Family\u003c\/h2\u003e\u003cp\u003eThe DL360 Gen9 is HPE's 1U dual-socket Gen9 workhorse, the 1U pair-partner to the 2U DL380 Gen9. Within the 1U DL360 Gen9 line there are three storage chassis: the 8-Bay 2.5\" SFF on this page, the higher-density \u003ca href=\"\/products\/dl360-g9-2-5-10-bay-hot-swap-psu\"\u003eDL360 Gen9 10-Bay 2.5\"\u003c\/a\u003e for the maximum SFF count in 1U, and the \u003ca href=\"\/products\/dl360-g9-3-5-4-bay-chassis\"\u003eDL360 Gen9 4-Bay 3.5\"\u003c\/a\u003e for large-format bulk capacity at the edge. The compute, memory, networking, and management platform is identical across all three; the only thing that changes is the front-bay layout.\u003c\/p\u003e\u003cp\u003eStep outside the 1U envelope and the decision is about expansion, not generation. When a workload needs more than three PCIe slots or more than ten drives, the \u003ca href=\"\/products\/hp-proliant-dl380-g9-2-5-8-bay-server\"\u003eDL380 Gen9 8-Bay 2.5\"\u003c\/a\u003e 2U companion is the right move: same Grantley platform, six PCIe slots, double-wide GPU support, and far more storage headroom. For a Dell-shop equivalent at the same 1U Gen9 tier, the \u003ca href=\"\/products\/dell-poweredge-r630-10-bay-chassis\"\u003eDell PowerEdge R630 10-Bay 2.5\"\u003c\/a\u003e is the direct counterpart: 1U, dual-socket, the same E5-2600 v3\/v4 Grantley generation.\u003c\/p\u003e\u003ch2\u003eStorage - 8 SFF Bays\u003c\/h2\u003e\u003cp\u003eEight 2.5\" SAS\/SATA hot-swap bays line the front of the chassis. This is the standard DL360 Gen9 SFF layout; the 10-Bay variant adds two more bays in the position the optical drive would otherwise occupy. NVMe is supported through the Express Bay option in specific front-bay positions, trading SAS\/SATA bay count for PCIe-attached NVMe lanes when low-latency local flash is the requirement.\u003c\/p\u003e\u003cp\u003eThe drive portfolio spans the full Gen9 SFF range: SAS SSDs in mixed-use and read-intensive endurance classes, cost-optimized SATA SSDs, 10K and 15K SAS HDDs for spinning capacity, self-encrypting (SED) variants for compliance-driven deployments, and NVMe via Express Bay. At full population, eight SFF bays deliver tens of terabytes of raw capacity depending on drive selection, with high-capacity SAS and SATA SSDs reaching the largest per-bay figures.\u003c\/p\u003e\u003cp\u003eCommon 8-Bay storage profiles we quote:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003evSphere cluster node, SAN datastore primary.\u003c\/strong\u003e 2x SSDs in RAID 1 for ESXi boot (or M.2 boot to preserve all eight bays), six SSDs for a vSAN cache device or local datastore. Primary VM storage lives on shared FC or iSCSI.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eHyper-V cluster node with CSV cache.\u003c\/strong\u003e 2x SSDs for Windows Server boot, six SSDs for CSV cache. Primary VM storage on Storage Spaces Direct or SAN.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eKubernetes worker host.\u003c\/strong\u003e 2x SSDs in RAID 1 for the OS, six SSDs for ephemeral local storage and container image layers. Persistent volumes ride a network CSI provider.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWeb and application tier.\u003c\/strong\u003e 2x SSDs in RAID 1 for the OS, four to six SSDs in RAID 5 or RAID 10 for application data, logs, and staging.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eMid-tier SQL Server or Oracle, networked primary storage.\u003c\/strong\u003e 2x SSDs for the OS, 2x SSDs mirrored for tempdb or redo, four SSDs in RAID 10 for logs. Primary datafiles on SAN.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eInfrastructure and branch services.\u003c\/strong\u003e 2x SSDs in RAID 1 for the OS and roles, the remaining bays for supplementary capacity or left open for field growth.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eBoot Drives\u003c\/h3\u003e\u003cp\u003eM.2 SATA via the HPE M.2 SSD enablement card is the cleanest boot pattern on this chassis: it consumes one PCIe slot but keeps all eight front bays available for data. The alternative is 2x SFF SSDs in RAID 1 in the front bays, which costs two of the eight bays. On an 8-bay build that is 25 percent of the storage budget spent on boot, so we default to M.2 boot on DL360 Gen9 8-Bay quotes unless the PCIe slot budget is already committed, in which case front-bay RAID 1 is the fallback.\u003c\/p\u003e\u003ch2\u003eStorage Controllers\u003c\/h2\u003e\u003cp\u003eThe DL360 Gen9 runs HPE's modular Smart Array controller family. We quote the controller to the workload rather than defaulting to the top SKU:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P440ar (2 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e The mainstream production default for mixed and read-heavy workloads. Mounts in the dedicated modular slot without consuming a standard PCIe slot.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P840 (4 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e The write-intensive choice for transactional databases and heavy logging, where the larger cache earns its place. Occupies a PCIe slot.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array H240ar (HBA \/ pass-through mode).\u003c\/strong\u003e The right controller for software-defined and hyperconverged storage stacks (vSAN, Storage Spaces Direct, Ceph, ZFS) that want raw disks rather than hardware RAID.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eB140i (embedded software RAID via the chipset).\u003c\/strong\u003e OS-boot and light-duty only. We do not quote B140i as a production data-RAID controller; it is a boot-volume convenience, not a storage controller.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eAll P-series controllers require the HPE Smart Storage Battery for write-back caching, and the flash-backed write cache module is a documented wear item that should be checked on any refurbished unit. For any production array with a meaningful write path, the P440ar or P840 with battery-backed cache is the recommendation; pass-through HBA only when the storage layer is handled in software.\u003c\/p\u003e\u003ch2\u003eProcessors\u003c\/h2\u003e\u003cp\u003eOne or two sockets of Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the Grantley C610 platform. The two generations are drop-in compatible within the same socket but cannot be mixed in a single server. Core counts run up to 18 per CPU on v3 (E5-2699 v3) and up to 22 per CPU on v4 (E5-2699 v4), so a dual-socket v4 build reaches 44 cores. TDPs span roughly 55 W on low-power SKUs to 145 W on the top bins.\u003c\/p\u003e\u003cp\u003eThe 1U thermal envelope is the constraint that drives CPU selection here. Top-bin parts such as the E5-2699 v4 at 145 W and the high-frequency E5-2667 v4 at 135 W require the performance heatsink and the high-performance fan kit; we validate thermal headroom on every top-bin quote rather than assuming the standard heatsink will carry it. A common field error is ordering a high-TDP CPU into a chassis that shipped with standard cooling, so we confirm the heatsink and fan configuration matches the CPU bin before a build leaves.\u003c\/p\u003e\u003cp\u003eOne more trap worth naming: a single-socket DL360 Gen9 populates only half the platform. Single-CPU builds expose just 12 of the 24 DIMM slots and roughly half the PCIe lanes, so if memory capacity or expansion matters, the second socket is not optional.\u003c\/p\u003e\u003ch2\u003eMemory\u003c\/h2\u003e\u003cp\u003e24 DDR4 DIMM slots, 12 per CPU across the platform's memory channels. The ceiling is 3 TB using 128 GB LRDIMMs across all 24 slots; a more typical production build runs 32 GB or 64 GB RDIMMs, landing at 768 GB or 1.5 TB respectively. RDIMM is the mainstream choice; LRDIMM unlocks the highest capacities at a latency cost; NVDIMM-N is available as a niche persistence option on supported configurations. This platform predates Intel Optane persistent memory, which arrived with the later Cascade Lake generation, so PMem is not part of the DL360 Gen9 conversation.\u003c\/p\u003e\u003cp\u003eMemory speed is population-dependent. Broadwell-EP v4 CPUs run DDR4-2400 at one DIMM per channel; Haswell-EP v3 tops out at DDR4-2133. Fully populating to two DIMMs per channel steps the speed down a tier, which is the expected behavior, not a fault, and worth planning around when a build needs both maximum capacity and maximum bandwidth. HPE Smart Memory is required to hit rated speeds; third-party DIMMs may train at a lower speed or refuse to post.\u003c\/p\u003e\u003ch2\u003eNetworking and PCIe Expansion\u003c\/h2\u003e\u003cp\u003eNetworking runs through the FlexibleLOM mezzanine rather than fixed onboard ports, so the network personality is a build-spec choice that does not consume a PCIe slot. Options span the 331FLR quad-port 1 GbE for management-tier and branch builds, the 530FLR-SFP+ and 534FLR-SFP+ dual-port 10 GbE SFP+ adapters for mainstream production, 10GBASE-T for copper 10 GbE plants, and 25 GbE SFP28 for the densest east-west fabrics. For dense 1U production, dual-port 10 GbE FlexibleLOM is the typical default.\u003c\/p\u003e\u003cp\u003ePCIe expansion is where the 1U chassis shows its limits. With both CPUs populated the DL360 Gen9 offers up to three PCIe Gen3 slots through its riser options, and the slot budget gets allocated carefully: a discrete Smart Array controller, an M.2 boot card, and an add-in NIC or accelerator can fill the chassis quickly. When a workload needs more than three slots, that is the signal to move to the 2U DL380 Gen9 rather than fight the 1U slot count.\u003c\/p\u003e\u003ch2\u003eGPU Support\u003c\/h2\u003e\u003cp\u003eGPU and accelerator support in the DL360 Gen9 1U envelope is limited to single-width, low-profile cards within the chassis power and thermal budget; the NVIDIA T-series single-slot accelerators are the representative fit. There is no room for double-wide compute GPUs in 1U. Workloads that need full-height double-wide accelerators, multiple GPUs, or higher per-card power belong in the 2U DL380 Gen9 or a 4-socket DL560-class platform, where the chassis can deliver the airflow and slot width a large accelerator requires.\u003c\/p\u003e\u003ch2\u003eManagement - iLO 4 Generation\u003c\/h2\u003e\u003cp\u003eThe DL360 Gen9 uses iLO 4, not the iLO 5 that arrived with Gen10. iLO Standard is included and covers health monitoring, power control, and basic remote access; iLO Advanced is the licensed tier that unlocks full graphical remote console, virtual media, and integrated remote KVM, and it is usually quoted separately. Intelligent Provisioning handles firmware and driver deployment, and the Active Health System log is the first place to look when diagnosing a refurbished unit's history.\u003c\/p\u003e\u003cp\u003eOn the security baseline, iLO 4 predates HPE's Silicon Root of Trust, which is a Gen10 hardware feature. UEFI Secure Boot is the firmware integrity baseline on Gen9, and a TPM module is available for platforms that need a hardware root for BitLocker or measured boot. Deployments with a hard requirement for Silicon Root of Trust hardware attestation should step to the Gen10 platform.\u003c\/p\u003e\u003ch2\u003ePower and Cooling\u003c\/h2\u003e\u003cp\u003ePower comes from HPE Flex Slot hot-plug supplies in 500 W, 800 W, and 1400 W ratings at Platinum and Titanium efficiency, configured 1+1 for redundancy. The 800 W Flex Slot pair is the standard production configuration and carries a dual-socket build with a full complement of drives; 500 W is adequate for entry-tier single-CPU builds; 1400 W is reserved for top-bin CPUs paired with high-draw expansion. HPE Power Advisor sizing is part of every quote so the PSU matches the as-configured draw rather than a nameplate guess.\u003c\/p\u003e\u003cp\u003eThermally the chassis supports ASHRAE A3 (40 C) and A4 (45 C) inlet ranges with the performance heatsinks fitted, though we target a 25 C to 30 C inlet for service-life optimization on production deployments. Top-bin CPUs require the high-performance fan kit and a thermal confirmation at quote time; the 1U envelope leaves less headroom than the 2U platform, so cooling is validated rather than assumed.\u003c\/p\u003e\u003ch2\u003ePhysical Specs \u0026amp; Platform Notes\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eForm factor:\u003c\/strong\u003e 1U rack chassis, standard-depth, fitting standard four-post racks; cable management adds the usual rear depth allowance behind the chassis.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePCIe expansion:\u003c\/strong\u003e up to three PCIe Gen3 slots with both CPUs populated, allocated through the riser options; low-profile cards only in the 1U envelope.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eParts availability:\u003c\/strong\u003e excellent. Gen9 is one of the most widely deployed enterprise generations ever shipped, so drives, controllers, FlexibleLOM adapters, PSUs, and rails are deep on the secondary market. HPE active warranty support has ended, and third-party maintenance is the standard production support path in 2026.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eAccessories we recommend:\u003c\/strong\u003e the \u003ca href=\"\/products\/hp-dl360-g8-g9-sff-1u-sliding-rails-679368-001-728437-001\"\u003eHPE 1U SFF ball-bearing sliding rail kit (P\/N 679368-001 \/ 728437-001)\u003c\/a\u003e for tool-less rack mounting, a cable management arm for serviceability, and the high-performance fan and heatsink kit on any top-bin CPU build.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePlatform notes:\u003c\/strong\u003e the 8-bay and 10-bay backplanes differ, so a field upgrade from eight to ten bays is a cage-and-backplane job rather than a drop-in; CPU hot-plug is not supported; M.2 boot is strongly preferred over front-bay boot on this chassis to preserve data bays; and top-bin CPU builds need their cooling validated before deployment.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eOur Assessment\u003c\/h2\u003e\u003cp\u003e\u003cstrong\u003eWhere it excels:\u003c\/strong\u003e The 8-Bay DL360 Gen9 is the right answer for 1U dual-socket nodes where storage is networked and local disk carries boot, cache, and logs. It is a strong fit for vSphere and Hyper-V cluster nodes backed by SAN or hyperconverged datastores, Kubernetes worker pools, stateless web and application tiers in dense racks, mid-tier databases with networked primary storage, and the broad class of infrastructure services (domain controllers, DNS, monitoring) that want a reliable 1U box at low acquisition cost.\u003c\/p\u003e\u003cp\u003e\u003cstrong\u003eWhere to look instead:\u003c\/strong\u003e If a build needs ten drives in 1U, start at the \u003ca href=\"\/products\/dl360-g9-2-5-10-bay-hot-swap-psu\"\u003eDL360 Gen9 10-Bay 2.5\"\u003c\/a\u003e rather than expanding bays later. If it needs more than three PCIe slots, double-wide GPUs, or substantially more storage, the \u003ca href=\"\/products\/hp-proliant-dl380-g9-2-5-8-bay-server\"\u003e2U DL380 Gen9\u003c\/a\u003e is the correct form factor. If the deployment requires iLO 5 and Silicon Root of Trust, step up to the \u003ca href=\"\/products\/dl360-g10-chassis\"\u003eDL360 Gen10 8-Bay 2.5\"\u003c\/a\u003e. Where budget is the hard constraint and a value-tier 1U Gen9 will do, the \u003ca href=\"\/products\/hpe-proliant-dl160-gen9-4-bay-lff-build-your-own\"\u003eDL160 Gen9 4-Bay 3.5\"\u003c\/a\u003e is the cost-floor option.\u003c\/p\u003e\u003cp\u003e\u003cstrong\u003eBottom line:\u003c\/strong\u003e This is the default 1U Gen9 build for an organization that runs networked primary storage and wants dependable dual-socket compute at refurbished pricing. It suits the IT team standardizing a virtualization or container cluster on proven hardware, or filling out an infrastructure tier where the per-node cost matters more than the latest platform features. Buyers who need maximum local storage, heavy PCIe expansion, or current-generation security hardware should read the alternatives above before committing.\u003c\/p\u003e\u003ch2\u003eWhere the DL360 Gen9 Fits in 2026\u003c\/h2\u003e\u003cp\u003eThe DL360 Gen9 launched in 2014 on the Grantley platform, which makes it roughly eleven to twelve years into its service life as of 2026. That is mature, not obsolete. The platform is two HPE generations behind the current Gen11 line and one behind Gen10 and Gen10 Plus, and HPE's own active warranty coverage has lapsed, so the honest framing is that this is a cost-driven acquisition where the workload fits comfortably inside the E5-2600 v3\/v4 envelope and third-party maintenance covers production support.\u003c\/p\u003e\u003cp\u003eThat describes a lot of real workloads. Dev, test, and staging environments, branch and edge compute, infrastructure services, lab build-outs, and budget-constrained virtualization clusters all run well on Gen9 hardware at a fraction of current-generation cost. Where the workload needs the newer platform's memory bandwidth, PCIe Gen4, larger core counts, or Silicon Root of Trust security, the \u003ca href=\"\/products\/dl360-g10-chassis\"\u003eDL360 Gen10\u003c\/a\u003e is the step up. The decision is a straightforward cost-versus-capability tradeoff, and for a large share of 1U dual-socket workloads the Gen9 still lands on the right side of it.\u003c\/p\u003e\u003ch2\u003eHonest Limitations\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003eiLO 4 management, without the Silicon Root of Trust hardware attestation introduced on Gen10.\u003c\/li\u003e\n\u003cli\u003eDDR4 speed caps at DDR4-2400 (v4) or DDR4-2133 (v3), and steps down a tier under full two-DIMM-per-channel population.\u003c\/li\u003e\n\u003cli\u003ePCIe Gen3 only, and just three slots in the 1U chassis, which is a tight expansion budget.\u003c\/li\u003e\n\u003cli\u003eThe flash-backed write cache module on Smart Array P-series controllers is a wear item and should be verified on any refurbished unit.\u003c\/li\u003e\n\u003cli\u003ev3 and v4 CPUs cannot be mixed in the same server, and single-socket builds expose only half the DIMM slots and PCIe lanes.\u003c\/li\u003e\n\u003cli\u003eHPE Smart Memory is required to reach rated DIMM speeds; third-party memory may train slower or fail to post.\u003c\/li\u003e\n\u003cli\u003eNetworking is FlexibleLOM-only with no fixed onboard ports, so a FlexibleLOM adapter is mandatory, not optional.\u003c\/li\u003e\n\u003cli\u003eThe 1U thermal envelope limits cooling headroom; top-bin CPUs require the performance fan and heatsink kit and a thermal check.\u003c\/li\u003e\n\u003cli\u003eNo double-wide GPU support in 1U; single-width low-profile accelerators only.\u003c\/li\u003e\n\u003cli\u003eMoving from eight to ten bays in the field is a cage-and-backplane change, not a drive add, so size the bay count up front.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eWorkload Fit\u003c\/h2\u003e\u003ctable\u003e\n\u003ctr\u003e\n\u003cth\u003eThis server is right for\u003c\/th\u003e\n\u003cth\u003eConsider alternatives for\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eVM and container cluster nodes with networked datastores at 1U density\u003c\/td\u003e\n\u003ctd\u003eBuilds needing ten drives in 1U (use the 10-Bay 2.5\")\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eStateless web and application tiers in dense compute racks\u003c\/td\u003e\n\u003ctd\u003eWorkloads needing more than three PCIe slots (use the 2U DL380 Gen9)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eKubernetes worker hosts and infrastructure services\u003c\/td\u003e\n\u003ctd\u003eDouble-wide GPU or multi-GPU compute (use DL380 Gen9 or DL560)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMid-tier databases with primary storage on SAN\u003c\/td\u003e\n\u003ctd\u003eDatabase hosts needing large local primary storage\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eBranch, edge, lab, dev, and staging deployments\u003c\/td\u003e\n\u003ctd\u003eNew deployments requiring iLO 5 and Silicon Root of Trust\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCost-driven 1U Gen9 standardization on proven hardware\u003c\/td\u003e\n\u003ctd\u003eWorkloads that need current-generation memory bandwidth or PCIe Gen4\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/table\u003e\u003ch2\u003eWhere to Look Instead\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed ten SFF bays in the same 1U chassis?\u003c\/strong\u003e The \u003ca href=\"\/products\/dl360-g9-2-5-10-bay-hot-swap-psu\"\u003eDL360 Gen9 10-Bay 2.5\"\u003c\/a\u003e is the maximum-SFF density option on the platform.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed large-format capacity at the edge?\u003c\/strong\u003e The \u003ca href=\"\/products\/dl360-g9-3-5-4-bay-chassis\"\u003eDL360 Gen9 4-Bay 3.5\"\u003c\/a\u003e takes LFF drives in the same 1U body for bulk-capacity branch and backup roles.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed more PCIe slots or more storage than 1U allows?\u003c\/strong\u003e The \u003ca href=\"\/products\/hp-proliant-dl380-g9-2-5-8-bay-server\"\u003eDL380 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the 2U companion with six PCIe slots and double-wide GPU support.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed current-generation 1U with iLO 5 and Silicon Root of Trust?\u003c\/strong\u003e The \u003ca href=\"\/products\/dl360-g10-chassis\"\u003eDL360 Gen10 8-Bay 2.5\"\u003c\/a\u003e is the direct generational step up.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWorking to the tightest budget on a 1U Gen9 build?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl160-gen9-4-bay-lff-build-your-own\"\u003eDL160 Gen9 4-Bay 3.5\"\u003c\/a\u003e is the value-tier 1U Gen9 step down.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eStandardized on Dell?\u003c\/strong\u003e The \u003ca href=\"\/products\/dell-poweredge-r630-10-bay-chassis\"\u003eDell PowerEdge R630 10-Bay 2.5\"\u003c\/a\u003e is the equivalent 1U dual-socket Grantley platform from the same generation.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eReady to Configure?\u003c\/h2\u003e\u003cp\u003eTell us the workload, the CPU generation preference (v3 versus v4), the CPU TDP context (the 1U thermal envelope matters for top-bin choices), the memory target, the storage configuration (drive types, RAID layout, controller preference, and M.2 versus front-bay boot), the FlexibleLOM choice (1, 10, or 25 GbE), the PSU configuration, and the quantity. We respond within 24 hours with a validated configuration that includes HPE Power Advisor sizing, thermal validation on high-TDP builds, and third-party maintenance coordination when you want it. Every refurbished unit ships with the Wholesale Servers 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units. Call 1-800-778-1545 or use the quote form below.\u003c\/p\u003e","brand":"HPE","offers":[{"title":"Default Title","offer_id":45951242928327,"sku":"BP-013616","price":307.83,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0748\/4493\/0247\/files\/server-design-lab-hpe-proliant-dl360-gen9-8-bay-25-drives-104158.png?v=1765539623"},{"product_id":"dl360-g9-2-5-10-bay-hot-swap-psu","title":"HPE ProLiant DL360 Gen9 10-Bay 2.5\" Drives [Gen9]","description":"\u003cp\u003eThe refurbished HPE ProLiant DL360 Gen9 10-Bay 2.5\" is the maximum-SFF-density configuration of HPE's Gen9 1U dual-socket line: ten 2.5\" SAS\/SATA hot-swap bays, the most front storage the 1U chassis holds. It runs the same Intel Grantley platform as the rest of the family, Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the C610 chipset, up to 22 cores per CPU on v4 (44 cores across two sockets), 24 DDR4 DIMM slots with a 3 TB ceiling, HPE modular Smart Array controllers, the FlexibleLOM mezzanine, and iLO 4 management. The two extra bays over the standard 8-Bay build are the point: this is the variant to pick when SFF capacity inside a single rack unit is the driver.\u003c\/p\u003e\n\n\u003cp\u003eThis page covers what is specific to the 10-bay layout, the high-density storage profiles it enables and where the extra two bays earn their place. The compute, memory, networking, and management platform is shared across every DL360 Gen9 chassis; the \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the primary page for the family and the standard build for compute-driven 1U workloads that do not need the extra bays.\u003c\/p\u003e\n\n\u003cp\u003eTo configure a build, call 1-800-778-1545 or use the quote form below. Every refurbished unit ships under our 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units.\u003c\/p\u003e\n\n\u003chr\u003e\n\n\u003ch2\u003eWhen 10 SFF Bays Is the Right Density\u003c\/h2\u003e\n\u003cp\u003eTen 2.5\" bays is the maximum SFF count the DL360 Gen9 1U chassis supports, and the two bays beyond the standard eight come from the area a slim optical drive would otherwise occupy. The 10-Bay is the right pick when local SFF capacity matters within the 1U footprint, rather than the compute-first case the 8-Bay covers:\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eDense vSphere cluster nodes with vSAN.\u003c\/strong\u003e Boot off M.2 or a front-bay pair, two cache SSDs, and six to eight capacity SSDs in the vSAN disk group, all in 1U. The extra bays let a hyperconverged node carry a useful local capacity tier without going to 2U.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eHyper-V and Storage Spaces Direct nodes at rack density.\u003c\/strong\u003e The same HCI pattern on the Microsoft side, where S2D ReadyNode-style designs benefit from the extra capacity disks per 1U node.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWeb and application tier in dense racks.\u003c\/strong\u003e A boot pair plus four to eight SSDs for application data and logs, at 1U per server, for deployments running many instances behind load balancers.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eHyperscale-style compute pools.\u003c\/strong\u003e Kubernetes worker pools and container hosts where local SSD carries ephemeral data and image layers while persistent volumes ride the network.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eMid-tier databases with local working storage.\u003c\/strong\u003e Boot, a mirrored tempdb or redo pair, and additional SSDs for local working sets, with primary datafiles on SAN.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eEdge and ROBO where one 1U box does everything.\u003c\/strong\u003e Ten bays cover branch file services, infrastructure roles, and local virtualization on a single rack-efficient server.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003eIf the workload is compute-driven and does not need ten bays, the \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the standard build. If bulk large-format capacity is the requirement, the \u003ca href=\"\/products\/dl360-g9-3-5-4-bay-chassis\"\u003eDL360 Gen9 4-Bay 3.5\"\u003c\/a\u003e takes LFF drives in the same 1U body. And if storage or expansion needs exceed what 1U holds, the 2U \u003ca href=\"\/products\/dl380-g9-2-5-16-bay-chassis\"\u003eDL380 Gen9 16-Bay 2.5\"\u003c\/a\u003e is the move.\u003c\/p\u003e\n\n\u003ch2\u003eStorage - 10 SFF Bays\u003c\/h2\u003e\n\u003cp\u003eTen 2.5\" SAS\/SATA hot-swap bays across the front of the chassis, the maximum SFF configuration for the 1U DL360 Gen9. NVMe is supported through the Express Bay option in specific positions (up to four SFF NVMe), trading SAS\/SATA bay count for PCIe-attached NVMe lanes. At full population, ten SFF bays deliver tens of terabytes of raw capacity depending on drive choice, with high-capacity SAS and SATA SSDs reaching the largest per-bay figures.\u003c\/p\u003e\n\u003cp\u003eThe drive portfolio spans the full Gen9 SFF range: SAS SSDs in mixed-use and read-intensive endurance classes, cost-optimized SATA SSDs, 10K and 15K SAS HDDs for moderate-IOPS spinning storage, self-encrypting (SED) variants for compliance, and NVMe via Express Bay. Common 10-Bay storage profiles we quote:\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003evSAN node.\u003c\/strong\u003e M.2 or front-bay boot, two cache SSDs, six to eight capacity SSDs in the disk group, primary VM storage distributed across the cluster.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eS2D node.\u003c\/strong\u003e A boot pair, cache and capacity SSDs split per the Storage Spaces Direct design, with the extra bays improving per-node capacity.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWeb and app tier.\u003c\/strong\u003e 2x SSDs in RAID 1 for the OS, four to eight SSDs in RAID 5 or RAID 10 for application data and logs.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eContainer host.\u003c\/strong\u003e A boot pair plus local SSDs for ephemeral data and image layers; persistent volumes on a network CSI provider.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eMid-tier database.\u003c\/strong\u003e An OS pair, mirrored tempdb or redo, additional SSDs for local working sets, primary datafiles on SAN.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003ch3\u003eBoot Drives\u003c\/h3\u003e\n\u003cp\u003eM.2 SATA via the HPE M.2 SSD enablement card is the cleanest boot pattern: it sits in a PCIe slot and preserves all ten front bays for data. The alternative is 2x SFF SSDs in RAID 1 in the front bays, which costs two of the ten bays, a 20 percent bite versus 25 percent on the 8-Bay. The DL360 Gen9 chassis has no rear drive bays, so M.2 and front-bay are the two practical boot options, and we default to M.2 boot on 10-Bay quotes where the PCIe budget allows so the full ten bays stay available for the storage tier.\u003c\/p\u003e\n\n\u003ch2\u003eStorage Controllers\u003c\/h2\u003e\n\u003cp\u003eThe DL360 Gen9 takes HPE's modular Smart Array controllers, which mount in a chassis-specific slot rather than a PCIe expansion position. That matters more here than on the 2U platform because the 1U chassis has only three PCIe slots total, so keeping the storage controller off PCIe leaves all three slots for networking, an HBA, or an accelerator.\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P440ar (2 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e The mainstream production controller; full hardware RAID 0\/1\/5\/6\/10\/50\/60 in the modular slot. The right pick for most 10-Bay production builds.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array H240ar (HBA \/ pass-through, modular).\u003c\/strong\u003e The HCI and software-defined choice (vSAN, Storage Spaces Direct, Ceph), clean SAS pass-through without consuming a PCIe slot.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eSmart Array P840 (PCIe, 4 GB flash-backed write cache, battery-backed).\u003c\/strong\u003e For write-intensive workloads where the 2 GB cache is not enough; it occupies a PCIe slot, so it is a deliberate tradeoff against the tight 1U slot budget.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eB140i (embedded software RAID via the chipset).\u003c\/strong\u003e Boot mirroring only; not a production data-RAID controller on a dual-socket platform.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003eAll P-series controllers require the HPE Smart Storage Battery for write-back caching. The flash-backed cache module is a wear item with a finite service life, and many refurbished units arrive with batteries past spec, so we check battery state and replace aged modules as part of build prep.\u003c\/p\u003e\n\n\u003ch2\u003eProcessors\u003c\/h2\u003e\n\u003cp\u003eOne or two sockets of Intel Xeon E5-2600 v3 (Haswell-EP) or v4 (Broadwell-EP) on the Grantley C610 platform, drop-in compatible within the socket but not mixable in one server. Single-socket builds are supported but expose only 12 of the 24 DIMM slots and a reduced PCIe count, so two sockets is the production standard. The 1U thermal envelope is the real constraint: the chassis carries CPUs up to 145 W, but the top bins need the performance heatsink and high-performance fan kit and an inlet-temperature check, because the 1U margin is tighter than the 2U platform. Common picks:\u003c\/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2620 v4 (8 cores, 85 W).\u003c\/strong\u003e Entry-tier production; 16 cores at two sockets, the easiest thermal envelope, the budget and branch choice.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2650 v4 (12 cores, 105 W).\u003c\/strong\u003e Balanced mid-tier; 24 cores at modest TDP, the general-purpose 1U dual-socket default.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2680 v4 (14 cores, 120 W).\u003c\/strong\u003e Production mainstream; 28 cores at two sockets, comfortable in the 1U envelope on the standard heatsink.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2667 v4 (8 cores, 135 W, 3.2 GHz base).\u003c\/strong\u003e High-frequency, low-core-count for per-core-licensed software; requires the performance heatsink.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eE5-2699 v4 (22 cores, 145 W).\u003c\/strong\u003e Top bin; 44 cores at two sockets is achievable but at the edge of the 1U thermal envelope, so it requires the performance heatsink, performance fans, and an inlet-temperature confirmation. We validate thermal headroom on every top-bin quote.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003eA common field error is dropping a high-TDP CPU into a chassis that shipped with standard cooling, so we confirm the heatsink and fan kit match the CPU bin before a build leaves.\u003c\/p\u003e\n\n\u003ch2\u003eMemory\u003c\/h2\u003e\n\u003cp\u003e24 DDR4 DIMM slots, 12 per CPU, the same memory architecture as the 2U DL380 Gen9; HPE did not cut DIMM count for the 1U form factor on Gen9. The ceiling is 3 TB with 128 GB LRDIMMs across all 24 slots, while a typical production build runs 32 GB or 64 GB RDIMMs for 768 GB or 1.5 TB. RDIMM is the mainstream choice and LRDIMM unlocks the top capacities at a latency cost; NVDIMM-N is available on v4 configurations as a niche persistence option. Speed is population-dependent: DDR4-2400 on v4 and DDR4-2133 on v3 at one DIMM per channel, stepping down a tier at full two-DIMM-per-channel population. HPE Smart Memory is required for rated speeds, and this platform predates Intel Optane persistent memory, so PMem is not part of the conversation here.\u003c\/p\u003e\n\n\u003ch2\u003eNetworking and PCIe Expansion\u003c\/h2\u003e\n\u003cp\u003eThe DL360 Gen9 uses FlexibleLOM as the primary networking mezzanine, with no embedded fixed ports alongside it, so the FlexibleLOM choice is part of the build spec and does not consume a PCIe slot. Options span the 331FLR quad-port 1 GbE, the 530FLR-SFP+ and 534FLR-SFP+ dual-port 10 GbE SFP+ adapters, 10GBASE-T, and 25 GbE SFP28 FlexFabric. Dense 1U racks usually pair with top-of-rack 10 GbE, so 10 GbE FlexibleLOM is the typical default; 1 GbE is acceptable for branch and edge.\u003c\/p\u003e\n\u003cp\u003ePCIe expansion is the 1U constraint: up to three PCIe Gen3 slots with both CPUs populated, versus six on the 2U DL380 Gen9. A single-socket build populates only two of the three. The slot budget is tight, so it gets allocated deliberately between an add-in HBA, a discrete controller, an accelerator, or specialty cards. When a workload needs more than three slots, the DL380 Gen9 is the right form factor.\u003c\/p\u003e\n\n\u003ch2\u003eGPU Support\u003c\/h2\u003e\n\u003cp\u003eGPU support in the 1U envelope is single-width, low-profile cards only, within the chassis power and thermal budget, with an NVIDIA T-series single-slot accelerator as the representative fit. There is no room for double-wide compute GPUs in 1U. Workloads that need double-wide accelerators, multiple GPUs, or higher per-card power belong in the 2U DL380 Gen9 or a 4-socket DL560-class platform, where the chassis can deliver the slot width and airflow a larger accelerator requires.\u003c\/p\u003e\n\n\u003ch2\u003eManagement - iLO 4 Generation\u003c\/h2\u003e\n\u003cp\u003eThe DL360 Gen9 ships with iLO 4, the same management generation as the rest of the Gen9 line, which is a real value point for a fleet standardizing on Gen9. iLO Standard covers health monitoring, power control, IPMI, SNMP telemetry, and the Active Health System log; iLO Advanced is the licensed tier for full graphical remote console and virtual media, and for a 1U lights-out deployment it is rarely optional given the operational benefit. We quote iLO Advanced explicitly when the deployment needs it.\u003c\/p\u003e\n\u003cp\u003eThe architectural difference from Gen10 is the security baseline: iLO 4 predates the Silicon Root of Trust hardware verification introduced with iLO 5. UEFI Secure Boot is the firmware integrity baseline on Gen9, and a TPM module is available where a hardware root is required. Deployments under compliance frameworks that require platform attestation should treat this as a documented gap versus Gen10 and either add compensating controls or step up.\u003c\/p\u003e\n\n\u003ch2\u003ePower and Cooling\u003c\/h2\u003e\n\u003cp\u003eHPE Flex Slot hot-plug supplies in 1+1 redundancy. The 500 W Platinum supply suits low-TDP single-CPU or modest dual-CPU builds; the 800 W Platinum or Titanium pair is the standard production choice and carries a dual-socket build with full memory and ten SSDs, including E5-2680 or E5-2690 v4; the 1400 W Platinum is the pick for top-bin E5-2699 v4 or NVMe-heavy builds. HPE Power Advisor sizing is part of every quote so the supply matches the as-configured draw.\u003c\/p\u003e\n\u003cp\u003eThermally the 1U chassis runs hotter than 2U at equivalent CPU TDP because the envelope is tighter. ASHRAE A3 (40 C) and A4 (45 C) extended ambient operation are supported on most configurations with performance heatsinks, though the top of A4 eats operational margin. We target a 25 C to 30 C inlet for service-life optimization on production deployments and validate thermals on every top-bin CPU quote.\u003c\/p\u003e\n\n\u003ch2\u003ePhysical Specs \u0026amp; Platform Notes\u003c\/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eForm factor:\u003c\/strong\u003e 1U rack chassis, standard depth for four-post racks, with the usual rear allowance for cable management.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePCIe expansion:\u003c\/strong\u003e up to three PCIe Gen3 slots with both CPUs populated (two with a single CPU), low-profile in the 1U envelope; the modular controller slot keeps the Smart Array off PCIe.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eParts availability:\u003c\/strong\u003e excellent. Gen9 SFF drives, modular controllers, FlexibleLOM adapters, PSUs, and rails are deep on the secondary market. HPE active warranty support has ended, and third-party maintenance is the standard production support path in 2026.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eAccessories we recommend:\u003c\/strong\u003e the \u003ca href=\"\/products\/hp-dl360-g8-g9-sff-1u-sliding-rails-679368-001-728437-001\"\u003eHPE 1U SFF ball-bearing sliding rail kit (P\/N 679368-001 \/ 728437-001)\u003c\/a\u003e for tool-less mounting, a cable management arm, and the HPE M.2 SSD enablement card so boot stays off the ten front bays.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePlatform notes:\u003c\/strong\u003e the two extra bays over the 8-Bay come from the optical-drive area and share the same backplane family; there are no rear drive bays on this chassis; CPU hot-plug is not supported; and top-bin CPU builds need their cooling validated before deployment.\u003c\/li\u003e\n\u003c\/ul\u003e\n\n\u003ch2\u003eOur Assessment\u003c\/h2\u003e\n\u003cp\u003e\u003cstrong\u003eWhere it excels:\u003c\/strong\u003e The 10-Bay is the configuration for hyperconverged and storage-leaning 1U nodes: vSAN and Storage Spaces Direct clusters that want a real local capacity tier per node, dense web and application tiers, and edge or ROBO boxes that consolidate compute and storage into a single rack unit. When the design calls for the most SFF capacity HPE puts in a 1U Gen9 chassis, this is the build.\u003c\/p\u003e\n\u003cp\u003e\u003cstrong\u003eWhere to look instead:\u003c\/strong\u003e If the workload is compute-driven and does not need ten bays, the \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the standard and more economical build. For large-format bulk capacity, the \u003ca href=\"\/products\/dl360-g9-3-5-4-bay-chassis\"\u003eDL360 Gen9 4-Bay 3.5\"\u003c\/a\u003e is the LFF option. For more PCIe slots, double-wide GPUs, or more than ten drives, the \u003ca href=\"\/products\/dl380-g9-2-5-16-bay-chassis\"\u003eDL380 Gen9 16-Bay 2.5\"\u003c\/a\u003e at 2U is the form factor, and for current-generation 1U with iLO 5 and Silicon Root of Trust, the \u003ca href=\"\/products\/hpe-proliant-dl360-g10-10-bay-2-5-chassis\"\u003eDL360 Gen10 10-Bay 2.5\"\u003c\/a\u003e is the step up.\u003c\/p\u003e\n\u003cp\u003e\u003cstrong\u003eBottom line:\u003c\/strong\u003e This is the maximum-density 1U Gen9 SFF build, for the team that wants local storage capacity without giving up rack density. It suits HCI cluster nodes and storage-leaning compute at refurbished pricing, and it is more chassis than a compute-only workload needs, where the 8-Bay is the better-value pick. Buyers who need current-generation security hardware, PCIe Gen4, or active HPE support should weigh the Gen10 step before committing.\u003c\/p\u003e\n\n\u003ch2\u003eWhere the DL360 Gen9 Fits in 2026\u003c\/h2\u003e\n\u003cp\u003eThe DL360 Gen9 launched in 2014 on Haswell-EP with a Broadwell-EP refresh in 2016, which puts it roughly eleven to twelve years into service as of 2026. HPE active warranty and ProSupport on Gen9 hardware has ended for both v3 and v4 builds, and third-party maintenance from established providers is the standard production support path. Gen10 (Skylake-SP and Cascade Lake-SP) introduced iLO 5 with Silicon Root of Trust, faster DDR4, and per-core performance gains, while Gen10 Plus and Gen11 brought PCIe Gen4 and DDR5.\u003c\/p\u003e\n\u003cp\u003eThe honest framing is a cost-versus-capability call. For new mission-critical deployments that need current security baselines, faster memory, PCIe Gen4, or active HPE support, the \u003ca href=\"\/products\/hpe-proliant-dl360-g10-10-bay-2-5-chassis\"\u003eDL360 Gen10\u003c\/a\u003e is the right answer. The Gen9 10-Bay earns its place extending an existing Gen9 fleet, in rack-density-driven workloads, in lab and staging that mirrors Gen9 production, and in budget-driven deployments where the workload fits the platform and the broad Gen9 parts supply is operationally valuable.\u003c\/p\u003e\n\n\u003ch2\u003eHonest Limitations\u003c\/h2\u003e\n\u003cul\u003e\n\u003cli\u003eHPE active warranty and ProSupport on Gen9 has ended for both v3 and v4 builds; production support is via third-party maintenance.\u003c\/li\u003e\n\u003cli\u003eiLO 4, not iLO 5, so no Silicon Root of Trust; firmware integrity rests on UEFI Secure Boot, a documented gap for platform-attestation compliance.\u003c\/li\u003e\n\u003cli\u003eDDR4 speed caps at DDR4-2400 (v4) or DDR4-2133 (v3) and steps down under full DIMM population; HPE Smart Memory required for rated speeds.\u003c\/li\u003e\n\u003cli\u003ePCIe expansion is three slots maximum; workloads needing more belong on the 2U DL380 Gen9 with six slots.\u003c\/li\u003e\n\u003cli\u003eNo double-wide GPU support in 1U; single-width low-profile accelerators only.\u003c\/li\u003e\n\u003cli\u003ePCIe Gen3 only; PCIe Gen4 needs Gen10 Plus or Gen11.\u003c\/li\u003e\n\u003cli\u003eThe 1U thermal envelope is tighter than 2U; top-bin CPUs require the performance heatsink and fan kit and an inlet check.\u003c\/li\u003e\n\u003cli\u003eThe flash-backed write cache module on P-series controllers is a wear item and is verified on every refurbished unit.\u003c\/li\u003e\n\u003cli\u003eNetworking is FlexibleLOM-only with no embedded fixed ports, so a FlexibleLOM adapter is mandatory.\u003c\/li\u003e\n\u003c\/ul\u003e\n\n\u003ch2\u003eWorkload Fit\u003c\/h2\u003e\n\u003ctable\u003e\n\u003ctr\u003e\n\u003cth\u003eThis server is right for\u003c\/th\u003e\n\u003cth\u003eConsider alternatives for\u003c\/th\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eDense 1U vSphere and Hyper-V nodes with local capacity\u003c\/td\u003e\n\u003ctd\u003eCompute-only 1U workloads that do not need ten bays (use the 8-Bay 2.5\")\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003evSAN and Storage Spaces Direct nodes at rack density\u003c\/td\u003e\n\u003ctd\u003eWorkloads needing more than three PCIe slots (use the 2U DL380 Gen9)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eWeb and application tier in dense racks\u003c\/td\u003e\n\u003ctd\u003eDouble-wide GPU or multi-GPU compute (use DL380 Gen9 or DL560)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eKubernetes worker and container hosts at 1U\u003c\/td\u003e\n\u003ctd\u003eNew deployments requiring iLO 5 and Silicon Root of Trust\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eEdge and ROBO consolidating compute and storage in 1U\u003c\/td\u003e\n\u003ctd\u003eLarge-format bulk capacity (use the 4-Bay 3.5\" or DL380 Gen9 LFF)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCapacity-add to an existing DL360 Gen9 fleet\u003c\/td\u003e\n\u003ctd\u003eWorkloads needing current memory bandwidth or PCIe Gen4\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/table\u003e\n\n\u003ch2\u003eWhere to Look Instead\u003c\/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eCompute-driven and do not need ten bays?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl360-g9-8-bay-2-5-chassis\"\u003eDL360 Gen9 8-Bay 2.5\"\u003c\/a\u003e is the standard 1U build for the family.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed large-format capacity in 1U?\u003c\/strong\u003e The \u003ca href=\"\/products\/dl360-g9-3-5-4-bay-chassis\"\u003eDL360 Gen9 4-Bay 3.5\"\u003c\/a\u003e takes LFF drives for edge and backup roles.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed more PCIe slots or more than ten drives?\u003c\/strong\u003e The \u003ca href=\"\/products\/dl380-g9-2-5-16-bay-chassis\"\u003eDL380 Gen9 16-Bay 2.5\"\u003c\/a\u003e is the 2U companion with six PCIe slots.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNeed current-generation 1U with iLO 5 and Silicon Root of Trust?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl360-g10-10-bay-2-5-chassis\"\u003eDL360 Gen10 10-Bay 2.5\"\u003c\/a\u003e is the direct generational step up.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWorking to the tightest budget on a 1U Gen9 build?\u003c\/strong\u003e The \u003ca href=\"\/products\/hpe-proliant-dl160-gen9-4-bay-lff-build-your-own\"\u003eDL160 Gen9 4-Bay 3.5\"\u003c\/a\u003e is the value-tier 1U Gen9 step down.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eStandardized on Dell?\u003c\/strong\u003e The \u003ca href=\"\/products\/dell-poweredge-r630-10-bay-chassis\"\u003eDell PowerEdge R630 10-Bay 2.5\"\u003c\/a\u003e is the equivalent 1U dual-socket Grantley platform from the same generation.\u003c\/li\u003e\n\u003c\/ul\u003e\n\n\u003ch2\u003eReady to Configure?\u003c\/h2\u003e\n\u003cp\u003eTell us the workload, the CPU generation preference (v3 versus v4), the CPU TDP context (the 1U thermal envelope matters for top-bin choices), the memory target, the storage configuration (drive types, RAID layout, controller preference, and M.2 versus front-bay boot), the FlexibleLOM choice (1, 10, or 25 GbE), the PSU configuration, and the quantity. We respond within 24 hours with a validated configuration including HPE Power Advisor sizing, thermal validation on high-TDP builds, and third-party maintenance coordination when you want it. Every refurbished unit ships with the Wholesale Servers 180-day warranty after a 12+ hour burn-in test, and volume pricing starts at 5 units. Call 1-800-778-1545 or use the quote form below.\u003c\/p\u003e","brand":"HPE","offers":[{"title":"Default Title","offer_id":45951242830023,"sku":"BP-003825","price":144.01,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0748\/4493\/0247\/files\/server-design-lab-hpe-proliant-dl360-gen9-10-bay-25-drives-742204.png?v=1765539623"}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0748\/4493\/0247\/collections\/proliant-dl360-gen-9-635843.jpg?v=1765540188","url":"https:\/\/wholesaleservers.com\/collections\/proliant-dl360-gen-9.oembed","provider":"Wholesale Servers","version":"1.0","type":"link"}