6.4 Managing HPE Alletra Storage MP: Onboarding, Provisioning, Monitoring & Local Tools
Key Takeaways
A new B10000 is racked, cabled and powered, then claimed in the GreenLake workspace and configured from Data Services Cloud Console in a few clicks.
Day-to-day B10000 work in DSCC covers hosts and host groups, volumes and volume sets, intent-based provisioning and protection policies.
Data Ops Manager reports headroom trends, latency hotspots, workload drift, resource contention, capacity savings and sustainability metrics.
Data-plane software upgrades are delivered from the cloud, matched to each system and applied non-disruptively.
The B10000 also has a local management UI, a CLI and a Web Services API, and HPE offers disconnected management for sites without cloud connectivity.
Managing HPE Alletra Storage MP
Quick Answer: Alletra Storage MP systems are cloud managed. You claim a new system in your HPE GreenLake workspace, then use Data Services Cloud Console (DSCC) and Data Ops Manager to configure it, create hosts and volumes, provision by intent, apply protection policies, monitor performance, capacity and sustainability, and run non-disruptive upgrades. The B10000 also keeps a local UI, CLI and Web Services API, and HPE offers disconnected management for isolated sites.
Day 0: deploy and onboard
HPE describes B10000 deployment as: rack the infrastructure, plug in the power cords and connect the network cables; then, with a few clicks, the system is configured and appears in the fleet, ready to serve data. The typical steps are:
- Install the chassis, nodes, JBOFs and, for switched systems, the back-end switches and cabling.
- Connect the management network so the system can reach the HPE GreenLake cloud outbound.
- Claim the device in the GreenLake workspace (Chapter 1) and assign it to DSCC with its subscription.
- Complete setup in Data Ops Manager: system name, network settings, time and DNS, and support connectivity.
Day 1: provision storage
Hosts and host groups
Define each server as a host with its initiators (FC WWPNs, iSCSI IQNs or NVMe host NQNs) and group clustered servers into a host group so every node of a cluster sees the same volumes.
Volumes and volume sets
Create volumes and collect related volumes into volume sets, for example all volumes of one database. Volume sets let snapshots, replication and protection policies treat those volumes as one consistent unit.
Intent-based provisioning
Instead of choosing a system, layout and ports by hand, the administrator states the storage tier, workload type, capacity and protection policy. AI-driven logic recommends the best-suited system in the fleet and applies the settings. This is how DSCC turns provisioning from infrastructure-centric to application-centric.
Protection policies
A protection policy defines snapshot schedules and retention, replication and backup targets. For the B10000, policies can send application-consistent backups to HPE StoreOnce systems onboarded to Data Ops Manager, and HPE GreenLake for Backup and Recovery can protect array volumes through volume protection groups, with copies on premises and in the HPE Cloud Protection Store.
Day 2: operate and optimize
| Task | Where and how |
|---|---|
| Health and alerts | Fleet dashboard in Data Ops Manager; InfoSight predictive analytics |
| Performance | Headroom utilization trends by volume set, top hotspots by latency, workload drift and resource contention detection |
| Capacity | Capacity reporting, savings and efficiency metrics, and forecasting |
| Sustainability | Energy consumption and carbon emissions metrics |
| Upgrades | Cloud-delivered data-plane upgrades, matched to each system and applied non-disruptively |
| Expansion | Add drives in pairs, JBOFs one at a time, or nodes on switched systems |
| Support | Cases and diagnostics through the GreenLake console and HPE Support Center |
| Automation | Published DSCC REST APIs |
Because management is SaaS, new data services become available without installing software, which HPE calls "invisible upgrades".
Access control
GreenLake identity and role-based access control decide who can view or change storage. Combine this with the B10000's Virtual Domains when several tenants or departments administer their own volumes on one system.
Managing the X10000
The X10000 appears in DSCC through its object service. Administrators create buckets, access keys and policies, configure bucket replication between X10000 systems, and monitor capacity and performance in Data Ops Manager. Applications then use standard S3 or NFSv4.1 access; HPE notes that the AWS and MinIO S3 SDKs are tested. DSCC APIs for the X10000 are published and S3-compatible.
Local and disconnected management
| Option | When to use it |
|---|---|
| Local management UI (B10000) | Direct administration on the array, for example during an internet outage |
| CLI | Scripting and advanced tasks |
| Web Services API (WSAPI) | Local automation |
| HPE Alletra Storage MP B10000 Disconnected and Alletra Storage MP Disconnected for X10000 | Sites that must run without cloud connectivity, such as air-gapped or classified environments |
Remember that losing the cloud connection never stops host I/O; it pauses cloud administration and telemetry only.
A worked onboarding example
An administrator receives a B10130 with 16 SSDs:
- The team racks and cables it and connects the management ports.
- The administrator adds the device in the GreenLake workspace and assigns it to Data Ops Manager.
- In DSCC, they create host group "ESX-Prod" with the WWPNs of four ESXi hosts.
- They provision 20 TB for "VMware datastores, tier: production", with a protection policy of hourly immutable snapshots kept for seven days and nightly backup to StoreOnce.
- Data Ops Manager later flags rising latency on one volume set; InfoSight shows the cause is a host CPU shortage rather than the array.
Common exam traps
- Cloud management does not move data; it manages the control plane only.
- Host groups and volume sets exist to keep clusters and multi-volume applications consistent.
- Disconnected sites have supported options; they are not excluded from Alletra Storage MP.
- Upgrades are non-disruptive and cloud-delivered, not manual downloads from a portal.
An administrator is provisioning storage for a three-node clustered database and wants every cluster node to see the same volumes and every database volume to be snapshotted at the same instant. Which B10000 objects support this?
A separate host for each node, with no grouping of the volumes
Pool A and pool B, one for each cluster node
A host group for the cluster nodes and a volume set for the database volumes
Virtual Domains created for each cluster node
How are data-plane software upgrades delivered to a B10000 managed from DSCC?
From the cloud, non-disruptively
Only by an HPE engineer on site during a scheduled outage window
By reinstalling ArcusOS from a USB drive on each controller node
Through SPOCK, which pushes new firmware down to the array
A defense agency needs a B10000 at a site with no internet connectivity. Which statement is accurate?
The B10000 cannot operate without a permanent connection to the HPE cloud
The agency must use an MSA Gen7 instead, because only the MSA works offline
Host I/O works, but the array can never be administered at that site
Use HPE's disconnected management option, plus the B10000's local UI, CLI and WSAPI
Which analytics does Data Ops Manager provide to help troubleshoot B10000 performance?
Tape cartridge load counts and drive cleaning history for the attached MSL library
Headroom trends, latency hotspots, workload drift and contention
SAF Collector upload status for the latest assessment
Pearson VUE exam scheduling for the storage team
Sections you finish are checked off in the contents.