Dell Proven Professional · Maintenance · D-PSC-MN-23
Dell PowerScale
Maintenance
Validates knowledge of PowerScale product installation, cabling, and maintenance including hardware components, Gen 6/6.5 node concepts, FRU/CRU procedures, and cluster implementation.
Master Dell PowerScale storage maintenance — from Gen 6/6.5 hardware concepts and cluster management tools through hands-on FRU/CRU procedures, rack installation, and full cluster implementation with OneFS.
›Cluster management tools — WebUI (port 8080, HTTPS, certificate-based); CLI (SSH to cluster IP or node IP, OneFS shell); Serial Console (RS-232, used during initial cluster creation or node recovery); PAPI (REST API for programmatic cluster management, JSON responses)
Module 02
PowerScale Hardware Maintenance
›Preliminary maintenance tasks — verify cluster health using isi status, node status, and drive status commands; confirm no existing faults; escalate to Support if protection is below requested level before proceeding
›Preparing a node for maintenance — smartfail vs halt: smartfail gracefully removes node from cluster and redistributes data to remaining nodes; halt powers off immediately without data redistribution
›Drive replacement (FRU) — hot-swap NVMe, SSD, or HDD; identify failed drive by amber LED blink pattern; pull tray; insert replacement; verify drive reintegration with isi devices output
›PSU hot-swap (FRU) — swap one PSU at a time; verify green LED on replacement; confirm redundancy retained throughout; node remains online
›Back-end NIC/InfiniBand adapter (FRU) — node offline required before removal; full ESD precautions; verify back-end connectivity resumes after replacement
›Motherboard/system board (FRU) — requires full node removal from rack; transfer NVRAM battery to replacement board; re-register node after reinsertion
›Pre-installation planning — site survey: floor load capacity, rack space, PDU circuit amperage, cooling airflow direction, cable run lengths; IP address planning: management IP, back-end IPs per node, front-end IP pool size
›Site preparation tasks — clear aisle for access, verify PDU breaker amperage against node power draw, pre-run and label network cables, mark rack unit positions for each node
›Rack and subcomponent installation — select rack type: Dell PowerEdge rack or customer-provided 4-post/2-post; install node-specific rail kit (telescoping rails vs static rails); attach cable management arm; insert node, secure thumb screws, verify cable routing
›Cabling nodes correctly — back-end InfiniBand: each node connects to dedicated InfiniBand switch; direct-connect topology valid for 3-node clusters only; front-end: connect node front-end ports to ToR switches per IP pool design
›Power cabling — connect both node PSUs to separate PDU circuits for full power redundancy; never share circuits between PSU A and PSU B on same node
›Cable labelling convention — source-port to destination-port notation; include rack-unit position in label; verify cable seating and bend radius after routing; document final cable map for service records
Module 04
PowerScale Implementation
›Creating a new cluster — Cluster Join wizard via serial console or WebUI: set cluster name, apply OneFS license, configure SmartConnect zone, assign node IPs, define back-end subnet, join first node as cluster primary
›Joining additional nodes to an existing cluster — SmartFail existing node first if replacing; add new node via WebUI Node Join page; use isi_reformat for full disk reformatting when redeploying a previously used node
›Node reimaging — boot from USB or network image when OneFS OS is corrupted; required when node fails to join cluster due to software state; restores factory-default OneFS partition
›SRS (Secure Remote Services) — ESRS gateway registration for proactive Dell support; enables phone-home for alerts and automated support ticket creation; SRS VE (virtual edition) deployment in customer VMware environment
›Compliance and WORM settings — WORM compliance mode sets cluster permanently to compliance mode; action is irreversible without Dell Support unlock; used for regulatory data retention (SEC, FINRA, HIPAA)
›Hadoop / HDFS support — activate HDFS protocol in OneFS; integrate with Apache Ambari and Hadoop ecosystem for big data analytics workloads; configure HDFS access zones and permissions
›OneFS rolling upgrade — isi upgrade commands: isi upgrade start, isi upgrade plan, node-by-node online upgrade; commit upgrade after all nodes complete; verify version with isi version; patch installation via isi pkg
Technologies You Will Master
Gen 6/6.5 Nodes
SolVe Tool
WebUI / CLI / PAPI
Serial Console
SmartFail
FRU Procedures
CRU Procedures
NVMe / SSD / HDD
InfiniBand
ESD Handling
OneFS Cluster Join
Node Reimaging
SRS Registration
Rolling Upgrade
Compliance Mode
HDFS Support
Gen 6/6.5 Nodes
SolVe Tool
WebUI / CLI / PAPI
Serial Console
SmartFail
FRU Procedures
CRU Procedures
NVMe / SSD / HDD
InfiniBand
ESD Handling
OneFS Cluster Join
Node Reimaging
SRS Registration
Rolling Upgrade
Compliance Mode
HDFS Support
Full Curriculum
4-Module D-PSC-MN-23 Programme
Complete PowerScale maintenance curriculum — from Gen 6/6.5 hardware concepts and cluster management tools through hands-on FRU/CRU procedures, rack installation, and full cluster implementation with OneFS.
›Installation methods and tools — SolVe Online tool for guided installation procedures; installation phases: pre-install assessment, site preparation, rack and cable, power-on, initial configuration
›PowerScale Gen 6 node design — higher-density chassis architecture; 60-drive chassis nodes: H600 (hybrid NVMe+HDD) and H5600 (maximum-capacity HDD) for large-scale unstructured storage
›PowerScale Gen 6.5 node types — A300 (archive, cost-optimised), A3000 (high-capacity archive), F810 (all-flash NVMe performance), H700 (hybrid performance), H7000 (high-capacity hybrid); each node independently addresses the OneFS cluster
›Node form factors — 1U chassis (F810 all-flash nodes) vs multi-U chassis (H-series and A-series capacity nodes); chassis-based nodes share a backplane vs independent nodes with self-contained components
›Node internal components — NVMe drives for tier 0, SSD for cache/tier 1, HDD for capacity tier; dual-port 40/100 GbE back-end InfiniBand or Ethernet; front-end 10/25 GbE Ethernet NICs; redundant hot-swap PSU; hot-swap fan modules
›Cluster management tools — WebUI: HTTPS on port 8080, certificate-based authentication, full GUI administration; CLI: SSH to cluster virtual IP or individual node IP, OneFS restricted shell; Serial Console: RS-232 direct connection, essential during initial cluster creation and node recovery without network access; PAPI: RESTful API, JSON over HTTPS, used for scripting and third-party integration
›Preliminary maintenance tasks — run isi status to view overall cluster health; check per-node status and drive status to confirm no existing faults; verify data protection level meets requested threshold before beginning work; escalate to Dell Support if protection is degraded
›Preparing a node for maintenance — smartfail: gracefully removes the node from the cluster, triggers OneFS to redistribute data across remaining nodes before powering off (preferred method); halt: immediate power-off without data redistribution (emergency only)
›Drive replacement (FRU) — hot-swap procedure for NVMe, SSD, and HDD drives; identify failed drive by amber LED blink pattern on drive tray; pull tray while node is online; insert replacement drive; verify drive reintegration using isi devices output; OneFS automatically begins rebuild
›PSU hot-swap (FRU) — swap one PSU at a time to maintain power redundancy; verify green LED on replacement PSU indicates full power delivery; confirm power redundancy status in WebUI or isi status before swap
›Back-end NIC/InfiniBand adapter (FRU) — node must be taken offline (smartfailed) before NIC removal; ESD precautions required (wrist strap, mat); reseat replacement adapter, bring node back online, verify back-end InfiniBand connectivity in isi network list
›Motherboard/system board (FRU) — requires full physical removal of node from rack; transfer NVRAM battery to replacement board (NVRAM holds in-flight writes); re-insert node; node rejoins cluster using original node number if serial number matches
›CRU procedures — front bezel removal and replacement; rail kit adjustment for rack depth; cable management arm attachment and routing adjustment; rack ear replacement for panel mounting; ESD wrist strap, mat, and grounding required for all electronic components; bag removed components in anti-static packaging
›Pre-installation planning — site survey items: floor load capacity per rack position, rack U-space availability, PDU circuit amperage vs node power draw, cooling airflow direction (front-to-back), cable run lengths between rack and ToR switch
›IP address planning — management IP for cluster virtual SmartConnect zone; individual node back-end IPs per node (used for internal InfiniBand/Ethernet); front-end IP pool sizing based on client access zone requirements and SmartConnect load-balancing policy
›Site preparation tasks — clear hot-aisle and cold-aisle access corridors; verify each PDU breaker amperage rating against node nameplate; pre-run and label all network cables from rack to ToR switches; mark rack unit positions and verify square nut placement for rail mounting
›Rack and subcomponent installation — select rack type: Dell PowerEdge 42U/48U rack or customer-provided 4-post 19" rack; install node-specific rail kit (telescoping rails for tool-less insertion vs static rails for permanent mounting); attach cable management arm to maintain cable bend radius; slide node into rack, tighten thumb screws, verify cable routing does not obstruct airflow
›Back-end cabling — InfiniBand topology: each node connects its back-end IB port to a dedicated InfiniBand switch (1:1 connectivity required for 4-node and larger clusters); direct-connect back-end topology supported for 3-node clusters without a switch; verify IB cable type and length matches the SFP/QSFP standard for the node generation
›Front-end and power cabling — connect each node’s front-end Ethernet ports to ToR switches per IP pool design document; connect PSU A to PDU circuit A and PSU B to PDU circuit B on separate breakers; cable label format: <source-node>-<port> to <dest-switch>-<port> including rack-unit position; document final cable map for service records
›Creating a new cluster — access Cluster Join wizard via serial console (RS-232) or WebUI; required inputs: cluster name, OneFS license key, SmartConnect zone name, management IP, back-end subnet and mask, node-specific IPs; first node designated as cluster primary and generates cluster configuration database
›Joining additional nodes — when adding a node to an existing cluster: connect cabling, power on node; access WebUI Node Join page or CLI; node discovery via back-end InfiniBand; if reusing a node previously in another cluster, run isi_reformat to clear existing cluster membership data and reformat all drives
›Node reimaging — required when OneFS OS partition is corrupted or unbootable; boot node from USB installer or network PXE image; reimaging restores factory-default OneFS and clears all node configuration; node joins cluster fresh after reimaging
›SRS (Secure Remote Services) — deploy ESRS (EMC Secure Remote Services) gateway VM on customer network; register each PowerScale cluster with ESRS gateway; enables proactive phone-home for alerts, automated service ticket creation, and remote diagnostic access by Dell Support
›WORM Compliance Mode — enables SmartLock compliance mode cluster-wide; once set to compliance mode the cluster cannot be reverted without Dell Support intervention; required for regulatory compliance environments (SEC Rule 17a-4, FINRA, HIPAA); objects stored under WORM become immutable for the configured retention period
›Hadoop / HDFS support — activate HDFS protocol in OneFS via WebUI Protocols section; configure HDFS access zones and authentication (Kerberos or simple); integrate with Apache Ambari or Hadoop management tools; PowerScale presents as an HDFS-compatible namenode for big data analytics workloads
›OneFS rolling upgrade procedure — run isi upgrade start to initiate; isi upgrade plan shows node-by-node upgrade sequence; each node upgrades and reboots independently while the cluster remains online; run isi upgrade commit after all nodes complete to finalise; install patches with isi pkg; verify version with isi version
Earn the D-PSC-MN-23 Dell Proven Professional certification and demonstrate your expertise performing hardware installation, FRU/CRU replacements, and OneFS cluster implementation on PowerScale Gen 6 and Gen 6.5 systems.