Skip to content

HPE ProLiant Servers

Scope

HPE ProLiant rack and tower servers: model selection (DL, ML, DL Gen11), iLO (Integrated Lights-Out) management, firmware lifecycle, HPE OneView integration, server profiles, and hardware configuration for datacenter deployments. Covers initial provisioning, ongoing management, and capacity expansion.

Checklist

  • [Critical] Is the correct ProLiant platform selected for the workload? (DL series for rack-dense compute, ML series for tower/edge, DL Gen11 for latest Intel/AMD with PCIe Gen5 and DDR5)
  • [Critical] Is iLO firmware updated to the latest revision and configured with a dedicated management network (not shared with production traffic)?
  • [Critical] Are iLO accounts configured with role-based access control, Active Directory/LDAP integration, and default credentials changed before deployment?
  • [Critical] Is HPE OneView deployed for fleet management when managing more than a handful of servers, with server profiles defining BIOS, boot, firmware, and network configurations as code?
  • [Critical] Is the firmware baseline established and managed via HPE OneView or iLO Amplifier Pack, with Service Pack for ProLiant (SPP) tested in a staging environment before production rollout?
  • [Critical] Is the RAID controller configured correctly? (HPE Smart Array or MR controller with appropriate RAID level, write cache policy, and battery/capacitor-backed cache for write-intensive workloads)
  • [Recommended] Are server profiles in OneView used to template BIOS settings, boot order, firmware levels, and logical network connections for repeatable provisioning?
  • [Recommended] Is Intelligent Provisioning used for initial OS deployment, or is a PXE/image-based deployment pipeline (e.g., Ansible with Redfish) established for fleet-scale provisioning?
  • [Recommended] Is HPE Agentless Management (AMS) or iLO-based agentless monitoring configured to report hardware health to management tools without in-OS agents?
  • [Recommended] Is the power supply configuration N+1 redundant with the correct input power (AC vs DC, single-phase vs three-phase), and are power capping policies configured for dense deployments?
  • [Recommended] Are thermal and fan policies reviewed for the installation environment? (ProLiant Gen11 supports extended ambient temperature up to 40C/45C in specific configurations)
  • [Recommended] Is the Redfish API used for automation instead of legacy RIBCL/SMASH, with scripts or Ansible modules targeting the iLO Redfish endpoint for firmware updates, power control, and configuration?
  • [Recommended] Is HPE InfoSight for Servers enabled to provide predictive analytics, firmware recommendations, and proactive support case generation?
  • [Optional] Is HPE Trusted Platform Module (TPM 2.0) enabled for measured boot, BitLocker/LUKS key sealing, and attestation in security-sensitive environments?
  • [Optional] Is Silicon Root of Trust verified as active? (Gen10+ ProLiant servers embed a hardware root of trust in the iLO ASIC that validates firmware integrity on every boot)
  • [Optional] Is Persistent Memory (Intel Optane PMem or CXL-attached memory) evaluated for latency-sensitive database or caching workloads on supported Gen10/Gen11 models?

Why This Matters

ProLiant servers are one of the most widely deployed x86 server platforms in enterprise datacenters. iLO management is central to the operational model -- misconfigured iLO (default passwords, shared networks, outdated firmware) is a common attack vector and operational liability. HPE's Service Pack for ProLiant bundles firmware, drivers, and agents into a tested baseline, but applying SPP to production without staging has caused outages from BIOS setting resets and NIC firmware regressions. OneView server profiles are the key to treating bare-metal configuration as infrastructure-as-code, but they require careful planning around logical enclosures, network sets, and firmware baselines. Gen11 introduces PCIe Gen5 and DDR5, which changes NIC and GPU compatibility assumptions from Gen10 designs.

Common Decisions (ADR Triggers)

  • Server tier -- DL360 (1U dense) vs DL380 (2U expandable) vs DL560 (4-socket) based on workload density and expansion requirements
  • Management platform -- iLO standalone vs HPE OneView vs HPE GreenLake for Compute Ops Management (cloud-managed)
  • Firmware lifecycle -- SPP-based updates via OneView (scheduled, staged) vs manual iLO updates vs GreenLake cloud-managed firmware
  • RAID controller -- HPE Smart Array (hardware RAID) vs software RAID vs pass-through for hyperconverged/SDS (VMware vSAN, Nutanix)
  • Deployment method -- Intelligent Provisioning (interactive) vs PXE + Kickstart/Preseed vs Ansible Redfish automation vs OneView server profiles
  • Out-of-band network -- Dedicated iLO NIC on isolated management VLAN vs shared NIC (not recommended for production)
  • Security posture -- Silicon Root of Trust + TPM + Secure Boot full chain vs selective enablement based on compliance requirements

See Also

  • general/compute.md -- general compute architecture patterns
  • general/hardware-sizing.md -- hardware sizing methodology
  • general/hardware-vendor-partnerships.md -- vendor partnership considerations
  • providers/hpe/synergy.md -- HPE Synergy composable infrastructure
  • providers/hpe/greenlake.md -- HPE GreenLake cloud services