Pass the Pure Storage FAIS FlashArray-Implementation-Specialist Questions and answers with Dumpstech

On theFlashArray//XL,PCIe Slot 5is explicitly reserved forNon-Disruptive Upgrade (NDU)procedures and specific control plane functions.

The FlashArray//XL chassis (5U) features a redesigned controller layout with 9 PCIe slots per controller (numbered 0-8). While slots 0-3 and 6-8 are generally available for Host I/O (FC/iSCSI) or Backend Expansion (SAS/NVMe-oF),Slot 5is kept open or populated with specific interconnect cards used during controller upgrades.

During a "Data-in-Place" upgrade (e.g., moving from an //X to an //XL or upgrading //XL generations), this slot facilitates the high-speed link required to cluster the old and new controllers or to transfer state information. Implementation Engineers must ensure this slot is not populated with standard customer I/O cards (like FC HBAs) during the initial configuration, as doing so would block the ability to perform future non-disruptive hardware upgrades.

When installing rails for a FlashArray in a rack requiring cage nuts (square hole racks that are not using the tool-less mechanism, or specific threaded adaptations), the correct placement for the cage nut is thetop square of the lowest Rack Unit (RU)that the rail will occupy.

Pure Storage rail kits typically secure at the bottom of the targeted rack unit.

The "3-Hole" Rule:A standard Rack Unit (1U) consists of three vertical holes (often referred to as the bottom, middle, and top hole).

Alignment:The rail flange is designed to align such that the retaining screw or positioning stud engages specific holes to support the weight. For the standard securement point, the documentation specifies installing the cage nut in thetop holeof the specific RU where the rail base sits (often the "lowest" RU if the device spans multiple U's, though the rail itself is 1U high).

Precision:Placing it in the "Lowest square of the lowest RU" (Option A) would interfere with the rail's shelf lip or the device below it. Correct alignment ensures the rail is level and the chassis slides in without binding.

For aDark Sitecustomer (where Remote Assist and PhoneHome are disabled), the appropriate next step before performing a shelf evacuation is touse the evac_checks script.

Risk Mitigation:Evacuating a shelf is a resource-intensive operation that permanently removes physical capacity from the system. In a connected environment, Pure Support runs remote diagnostics to verify the array has enough free space and system resources to handle the evacuation safely.

Dark Site Procedure:Since Support cannot log in remotely, the Implementation Engineer must manually run the evac_checks python script (typically loaded via USB or available in the Purity toolkit). This script performs the equivalent validation: checking effectively used capacity, dedup ratios, and system load to predict if the remaining drives can absorb the data from Shelf 01 without filling the array to 100% or causing performance degradation. Relying solely on a generic percentage (Option B or C) is dangerous because it doesn't account for data reduction overhead or specific RAID overheads.

TheFlashArray//XLsupports0 (zero)SAS expansion shelves for permanent capacity.

Architecture:The FlashArray//XL is architected exclusively forNVMestorage media. Its backend connectivity utilizes 100GbE RoCE (RDMA over Converged Ethernet) to connect toDirectFlash Shelves (DFS). It does not possess the legacy SAS controllers or ports required to drive SAS expansion shelves.

Migration Exception:While a "SAS Swing Shelf" (containing legacy SSDs from an older array) might be temporarily attached during a specific Data-in-Place upgrade migration scenario, this is a transient state. In terms ofsupportedoperational configuration for storage capacity, the limit is zero. The platform relies entirely on DirectFlash Modules in the chassis and external DirectFlash Shelves.

Contrast:Older generations (FlashArray//M, FlashArray//X R1/R2) supported SAS shelves, but the //XL platform eliminates this protocol entirely in favor of end-to-end NVMe.

While modern FlashArrays (like the XR4 and XL series) primarily use NVMe-oF/RoCE for backend connectivity, specific configurations or legacy support scenarios may still require aSAS HBA(Serial Attached SCSI Host Bus Adapter), for example, when connecting to older SAS-based expansion shelves during a migration or hybrid configuration.

ForR4B revision controllers(a specific hardware revision of the Gen 4 controller), the strict hardware addressing map requires that any SAS HBA be installed inPCIe Slot 1.

Slot 1provides the specific PCIe lane routing and BIOS interrupts required for the legacy SAS controller to be recognized and function without conflicting with the native NVMe fabric.

Installing the card in Slot 0 (typically reserved for RoCE/Ethernet backend) or Slot 4 would likely result in the card being ignored by Purity or causing boot errors due to resource conflict.

TheFlashArray//XLrequires4 x C19/C20power cables to connect to the rack Power Distribution Units (PDUs).

The FlashArray//XL chassis (5U) is significantly larger and draws more power than the standard FlashArray//X series (3U). It is equipped with four high-capacity Power Supply Units (PSUs) to support the increased thermal and electrical load of the dense DirectFlash Modules and high-performance controllers.

Connector Type:To handle the higher amperage safely, the //XL utilizesIEC C20inlets on the chassis side (rectangular with horizontal pins) rather than the standard C14 inlets found on most servers. Consequently, the required cables must have aC19connector (female) to plug into the array and aC20connector (male) to plug into the PDU (assuming a high-power PDU).

Quantity:Since the chassis has four PSUs for N+2 redundancy, four distinct cables are required. The standard C13/C14 cables (Option A/B) used for lower-wattage IT equipment are physically incompatible and electrically insufficient for the //XL platform.

After configuring the management network and proxy settings, it is essential to validate that the FlashArray can successfully transmit logs and telemetry to the Pure1 cloud. The specific command designed to trigger a manual test of this "Phone Home" functionality ispurearray phonehome test.

purearray phonehome test: Initiates a test cycle where the array attempts to resolve the Pure1 endpoint, establish a secure connection, and send a sample payload. Success here confirms that DNS, firewall rules, and proxy settings are correct for outbound telemetry.

puresupport phonehome: This is often used to manuallysendlogs (e.g., puresupport phonehome --send), but thetestsyntax is specifically under purearray.

purearray test pinghome: This is not a valid standard command for this specific validation.

Using the correct command ensures the engineer receives a clear "Success" or "Failure" message to sign off on the installation checklist.

=========

TheFlashArray//Econnects to its DirectFlash Shelves (DFS) using25Gb Ethernet (RoCE).

The FlashArray//E is the "Capacity Optimized" member of the FlashArray family, designed to replace spinning disk repositories with all-flash efficiency. While the high-performance FlashArray//XL utilizes 100GbE for its backend NVMe-oF connectivity to maximize throughput for mission-critical databases, the //E prioritizes cost-efficiency, density, and power consumption.

Architecture:To maintain this efficiency balance, the backend connectivity for expansion shelves on the //E platform is standardized on25GbE. This provides ample bandwidth for the QLC-based workloads (typically unstructured data, repository, or Tier 2 block) without the higher cost and power draw of 100GbE optics and controllers.

Comparison:Option A (40G) is a legacy QSFP standard not used in modern Pure backend designs. Option C (100G) is reserved for the performance-tier //XL and high-end //X models.

TheFibre Channel (FC)I/O card type is the only option listed that is universally compatible and supported acrossallFlashArray models, from the entry-level //X10 and //X20 up to the high-end //X90 and //XL.

Fibre Channel:Pure Storage architectures rely heavily on Fibre Channel as the primary enterprise storage protocol. All chassis generations and controller sizes feature PCIe risers compatible with Pure's standard 16Gb or 32Gb Fibre Channel HBAs.

Ethernet/iSCSI:While all arrays support iSCSI, the physicalcard typevaries.

10GBase-T:This refers to "copper" Ethernet (RJ45). This card type isnotsupported on all models. High-end FlashArrays (like the //X70, //X90, and //XL) typically utilize SFP+ or QSFP28 cages for optical connectivity (10/25GbE or 40/100GbE) and do not standardly support 10GBase-T copper cards due to power and latency characteristics.

Therefore, while theprotocol(iSCSI) is supported everywhere, the specific physicalcard(10GBase-T) is not. Fibre Channel cards remain the consistent hardware constant across the entire product line.

When upgrading from a FlashArray//X to a FlashArray//XL, the hardware migration often involves handling DirectMemory Modules (DMMs) differently than standard data drives. DMMs are high-performance cache modules that reside in the controller chassis.

In this specific scenario—migrating a fully populated //X90R3 to an //XL that already has its own set of DirectFlash Modules (DFM-Ds)—the DMMs from the old chassis need to be moved to the new chassis to preserve their cache function or be properly decommissioned.

The correct procedure is toEvacuate DMMs as a datapack then install into any slot in //XL.

DMMs cannot be simply "hot-swapped" randomly while the system is live in a way that disrupts the cache map.

They must be logically grouped and evacuated (conceptually "vacated" of active data) so they can be safely removed from the old chassis.

Once removed, they can be installed into the new //XL chassis. The //XL architecture supports DMMs in specific slots, and Purity will detect and re-incorporate them into the cache pool.

This ensures that the expensive SCM (Storage Class Memory) hardware is reused in the new system without causing data integrity issues during the transition.

=========