HPE ATP Storage Solutions cheat sheet

HPE Primera: mission-critical, 100% availability guarantee, NVMe-based, scale-up/scale-out, AI-driven InfoSight.

HPE Nimble Storage: midrange, adaptive flash, predictive analytics, scale-out, all-flash and hybrid options.

HPE MSA: entry-level, affordable, simple SAN/DAS, ideal for SMBs and remote offices.

HPE StoreOnce: deduplication backup appliances, Catalyst and RMC protocols, cloud-integrated.

HPE Apollo: high-performance compute/storage for AI, analytics, and unstructured data.

HPE 3PAR: legacy midrange/high-end, now superseded by Primera; still supported.

HPE SimpliVity: hyperconverged infrastructure with built-in backup and deduplication.

Positioning rule: Primera for tier-0/1, Nimble for tier-2, MSA for tier-3/entry, StoreOnce for backup.

HPE Cloud Volumes: cloud-native block storage for AWS/Azure, managed by HPE, for dev/test and DR.

RAID 5: single parity, tolerates one drive failure, usable capacity = (N-1)*drive size.

RAID 6: double parity, tolerates two drive failures, usable capacity = (N-2)*drive size.

RAID 10: mirrored stripes, tolerates one drive per mirror pair, usable capacity = 50%.

NVMe over Fabrics (NVMe-oF): uses RDMA or FC for low-latency access to remote NVMe devices.

Thin provisioning: allocates capacity on demand; monitor overcommit ratio to avoid depletion.

Deduplication: inline (during ingest) vs. post-process (after write); affects performance and capacity.

Replication: synchronous (zero RPO, latency-sensitive) vs. asynchronous (low RPO, distance).

Erasure coding: distributes data and parity across nodes, higher capacity efficiency than mirroring.

iSCSI: SCSI over TCP/IP; FC: SCSI over Fibre Channel; FCoE: FC over Ethernet.

Capacity planning formula: usable capacity = raw capacity × (1 - parity overhead) × (1 - spare overhead).

IOPS estimation: total IOPS = (number of drives) × (IOPS per drive); adjust for RAID write penalty (RAID 5/6 = 4/6).

Latency target: for OLTP, <1 ms; for analytics, <5 ms; for backup, <20 ms.

Bandwidth: throughput (MB/s) = IOPS × block size; ensure network and HBA can sustain.

Growth factor: design for 3-year projected growth; include 20-30% headroom.

Sizing for replication: RPO/RTO determine sync vs. async and bandwidth required.

HPE Ninja: use HPE NinjaSTACK or Sizer tools for validated configurations.

Power/cooling: calculate per-device wattage and BTU; ensure rack power density.

Multipath: design for redundant fabrics (FC) or NICs (iSCSI) to avoid single point of failure.

Initial setup: use HPE SmartStart or array's setup wizard; configure management IP, DNS, NTP.

Host connectivity: install HPE Smart SAN or multipath software; zone FC switches with single-initiator-single-target zones.

Array initialization: create CPGs (Primera/3PAR) or pools (Nimble) before provisioning volumes.

Volume creation: set size, provisioning type (thin/thick), and presentation to host set.

Host OS configuration: use HPE Host Explorer or guidance for MPIO settings (e.g., Round Robin).

Firmware update: use HPE Service Console or StoreServ Management Console; follow compatibility matrix.

Security: change default passwords, enable LDAP/AD authentication, configure encryption (SEDs or external KMS).

Validation: run HPE InfoSight or array health check; verify host sees LUNs and can read/write.

Documentation: record WWNs, IPs, zoning, and LUN mapping for future reference.

HPE InfoSight: cloud-based AI analytics for predictive support, cross-stack insights, and recommendations.

HPE Primera/3PAR: use SSMC (StoreServ Management Console) for configuration and monitoring.

HPE Nimble: use Nimble GUI or CLI; integrate with InfoSight for predictive alerts.

Alerting: configure SNMP traps, email alerts, and syslog for proactive monitoring.

Snapshots: point-in-time copies; schedule and retain per policy; use for quick recovery.

Replication: set up peer persistence (Primera/3PAR) or Nimble replication for DR.

Backup integration: use HPE StoreOnce with Catalyst for deduplicated backup from backup software.

Ransomware protection: immutable snapshots and air-gapped copies; HPE StoreOnce supports Catalyst over Fibre Channel.

Reporting: generate capacity, performance, and health reports via InfoSight or array tools.

Check InfoSight for known issues and recommendations before deep troubleshooting.

Performance bottleneck isolation: use array performance metrics (IOPS, latency, queue depth) and host metrics.

Common issue: high latency due to saturated ports; check FC port utilization and queue depth.

Common issue: disk failure; identify via array alerts and replace drive; rebuild time depends on RAID and drive size.

Common issue: path failure; verify multipath status and switch zoning.

Optimization: align host I/O size with array block size; enable write cache (if protected).

Optimization: use wide striping (Primera/3PAR) to distribute load across all drives.

Troubleshooting network: check iSCSI jumbo frames, flow control, and network congestion.

Log collection: use HPE support tools to collect logs for analysis; avoid rebooting without guidance.

Virtualization integration: HPE Primera/Nimble support VMware vSphere VAAI, VVols, and SRM.

Database solutions: use HPE Primera for low-latency Oracle/SQL; leverage snapshots for clone refresh.

Cloud integration: HPE Cloud Volumes for AWS/Azure; StoreOnce Cloud Bank for long-term retention.

Disaster recovery: design with HPE Peer Persistence (sync) or Nimble replication (async) and orchestration.

Hyperconverged: HPE SimpliVity for VM-centric, built-in backup and deduplication.

AI/analytics: HPE Apollo for high-throughput, scale-out storage with parallel file systems.

Customer scenario: SMB needing simple, affordable SAN → HPE MSA.

Customer scenario: mission-critical with zero downtime → HPE Primera with Peer Persistence.

Customer scenario: backup deduplication and cloud tier → HPE StoreOnce with Cloud Bank.