FBAA_004 FlashBlade Architect Associate PDF Dumps

Prévia do material em texto

1 / 8
Exam : FBAA_004
Title :
https://www.passcert.com/FBAA_004.html
Pure Storage FlashBlade
Architect Associate
2 / 8
1.A batch application reading data from FlashBlade is changed to issue fewer, larger reads. During its
next run, the engineer observes lower read IOPS but higher read throughput. The application owner asks
whether storage performance has deteriorated.
Which two additional measurements are most useful for assessing these observations? (Choose two.)
A. The array's data-reduction ratio during each run
B. The average size of completed read requests during each run
C. The physical storage utilization at the end of each run
D. The read I/O response times during each run
E. The number of files retained after each run
Answer: B, D
Explanation:
● B: Request size helps explain why fewer operations can transfer more data per second. A lower IOPS
value by itself does not establish a regression when the workload changes.
● D: Response times show whether individual reads are taking longer, adding a responsiveness
measure to the observed transfer rate. Interpret this alongside the application's requirements; the
measurements alone do not establish a root cause.
● A: Data reduction describes storage efficiency, not the size and completion time of the observed reads.
● C: Physical utilization may matter in a broader investigation, but it does not directly resolve the IOPS-
versus-throughput comparison.
● E: File count does not measure the data transferred by each read or its response time.
Authoritative reference: Arista and Pure Storage — Deploying NVMe-oF for Enterprise Leaf-Spine
Architecture, page 10, “Test Results”; verified 2026-09-21. Used for the general relationship among
request size, IOPS, throughput, and latency—not as evidence of FlashBlade NVMe-oF support.
2.A team is evaluating FlashBlade//S for two applications. One requires NFS access to its working files.
The other writes a separate data set through an S3 API. The team wants to keep both application
interfaces and consolidate storage without introducing a protocol gateway.
Which proposed design best fits this requirement?
A. Place both data sets in NFS file systems and translate the S3 application's requests through a
gateway.
B. Place both data sets in S3 buckets and modify the file application to use object requests.
C. Use native NFS file services for one application and native S3 buckets for the other on FlashBlade//S.
D. Deploy separate FlashBlade//S systems because file and object services require different storage
platforms.
Answer: C
Explanation:
● C: FlashBlade//S supports native file and object access, allowing these separate workloads to share a
platform.
● A: This introduces the gateway the customer wants to avoid.
● B: This requires the application-interface change the customer has ruled out.
● D: Separate arrays are not required merely because one workload uses NFS and the other S3.
This is a protocol-suitability decision, not a completed sizing recommendation. It does not imply that an
NFS file is also exposed as an S3 object.
Authoritative reference: Everpure — FlashBlade//S Data Sheet, ID19249-02-en-04/26, page 1,
3 / 8
“Advanced hardware and software working together”; verified 2026-09-21. The linked product address
returned the five-page official datasheet when checked.
3.During a FlashBlade procurement discussion, the finance team requires ownership of the storage
hardware. It also wants capacity consumption to determine subscription charges, subject to an agreed
reserve commitment, rather than purchasing the full capacity entitlement upfront.
Which Evergreen model best matches these requirements?
A. Evergreen//One
B. Evergreen//Flex
C. Evergreen//Forever
Answer: B
Explanation:
● B: Evergreen//Flex combines customer hardware ownership with a consumption-based subscription.
Its reserve commitment means this is not an unrestricted zero-minimum payment model.
● A: Evergreen//One provides storage as a service rather than the customer-owned hardware model
requested here.
● C: Evergreen//Forever supports traditional ownership with ongoing modernization, but does not provide
the requested Flex consumption structure.
Specific configuration eligibility, minimum commitments, and prices still require a commercial quote; none
is inferred from this model-selection question.
Authoritative references:
● Pure Storage — Data Storage That Gives You Flexibility with Control, PS2374-02, September 2023,
pages 3–4 and 8, portfolio comparison; verified 2026-09-21.
● Evergreen//Flex FAQs, “What is a flexible storage ownership subscription?” and “What is the reserve
commit and on-demand usage?”; live document checked 2026-09-21.
● Pure Storage — Own and Pay as You Go with Evergreen//Flex, official transcript, 03:20–04:13,
distinguishing provider-owned service from customer ownership; verified 2026-09-21.
4.What distinguishes FlashBlade//S DirectFlash media management from an architecture that relies on
the internal controllers of individual commodity SSDs?
A. Purity//FB coordinates flash-media management across the system.
B. Each DirectFlash Module independently sets the system's media-management policy.
C. Pure1 performs flash-media management remotely for every client write.
D. Client file-system drivers coordinate media management across the modules.
Answer: A
Explanation:
● A: Purity//FB manages resources globally, providing system-level media coordination.
● B: This incorrectly assigns the global policy to isolated module-level management.
● C: Pure1 monitoring is not the local flash-media-management function.
● D: Client drivers do not take over Purity//FB's media-management role.
Authoritative reference: Everpure — FlashBlade//S Data Sheet, ID19249-02-en-04/26, page 2,
“Unmatched performance efficiency” and “Take control of your unstructured data”; verified 2026-09-21.
5.A FlashBlade//S customer expects retained data to grow much faster than processing demand. An
4 / 8
initial budget assumes that every capacity increase must include a proportional increase in compute
resources. The customer asks whether this is an architectural requirement.
Which response is most accurate?
A. That assumption is required because blade count determines capacity and compute in a fixed ratio.
B. That assumption can be avoided by using data compression as a substitute for capacity planning.
C. That assumption is required for NFS workloads but can be avoided by using object storage.
D. That assumption can be revised because the modular design separates capacity and performance
growth.
Answer: D
Explanation:
● D: FlashBlade//S supports independent capacity and performance scaling; validate the actual
expansion against configuration rules.
● A: A universal fixed ratio contradicts that modular design.
● B: Compression does not establish the resources future workloads need.
● C: The architectural distinction is not conditional on choosing S3 instead of NFS.
Authoritative reference: Everpure — FlashBlade//S Data Sheet, ID19249-02-en-04/26, page 2, “Scale
compute and storage independently,” and page 4, “Storage that gets better over time”; verified 2026-09-
21.
6.In an illustrative FlashBlade planning worksheet, 240 TB of logical live data occupies 120 TB
physically. A new 60 TB logical data set is expected to achieve the same 2:1 logical-to-physical ratio.
Snapshot usage and other overhead remain unchanged and are excluded from these figures.
What physical live-data usage should the worksheet predict after ingestion?
A. 150 TB
B. 180 TB
C. 240 TB
D. 300 TB
Answer: A
Explanation:
● A: The new data needs 60/2 = 30 TB physically; 120 + 30 = 150 TB.
● B: This adds logical growth directly to physical usage, ignoring the stated reduction.
● C: This is the original logical size, not the new physical total.
● D: This is the new logical total before reduction.
Authoritative references:
● The Pure Storage Platform for AI — page 7, Data Reduction; compression principle only. All values
and exclusions are exercise assumptions; PS2571-02-en 05/24.Verified 2026-09-21.
7.A synthetic sizing exercise predicts 360 TB of physical user data after growth. At least 20% of usable
capacity must remain free, and the tested workload must sustain 8 GB/s. The configurations below have
already passed compatibility checks; capacity is after system overhead, and throughput is measured
with this workload.
Which is the smallest usable-capacity configuration that meets both requirements?
A. 400 TB usable; 10 GB/s
B. 450 TB usable; 7 GB/s
5 / 8
C. 480 TB usable; 9 GB/s
D. 600 TB usable; 12 GB/s
Answer: C
Explanation:
● A: 360/400 = 90% used; the free-space requirement fails.
● B: 360/450 = 80% used, but 7 GB/s misses the performance target.
● C: 360/0.80 = 450 TB is the capacity minimum. This option also exceeds 8 GB/s and is smaller than D.
● D: It passes both checks but is not the smallest qualifying configuration.
Authoritative references:
● FlashBlade Architect Associate Exam Guide — Section 2.2, workload configuration sizing. Numeric
configuration capabilities are synthetic inputs, not vendor specifications; FBAA_004; checked 2026-09-
21. Verified 2026-09-21.
8.An architect is comparing FlashBlade//S and FlashBlade//E for a growing repository whose access
requirements are modest relative to its retained capacity.
Which distinction should guide the initial platform discussion?
A. FlashBlade//E uses rotating disks, whereas FlashBlade//S uses flash.
B. FlashBlade//E emphasizes capacity economics; FlashBlade//S emphasizes demanding performance
workloads.
C. FlashBlade//E supports objects but requires FlashBlade//S for file access.
D. FlashBlade//E is an offline archive, whereas FlashBlade//S provides online access.
Answer: B
Explanation:
● A: Both are all-flash platforms; the distinction is not disk versus flash.
● B: The repository-oriented positioning of //E fits this initial discussion. Actual requirements still need
validation.
● C: //E also provides native file services.
● D: //E is an online storage platform, not an offline archive.
Authoritative references:
● FlashBlade//E Data Sheet — page 1, platform introduction and native protocols; ID19250-01-en-04/26.
Verified 2026-09-21.
● The Pure Storage Platform for AI — page 7, Spectrum of Performance and Cost Options; PS2571-02-
en 05/24. Verified 2026-09-21.
9.An illustrative capacity forecast shows 70 TB used out of 100 TB usable, with growth of 5 TB per
month. Policy requires expansion before usage exceeds 80 TB. Procurement and installation take three
months, and no deletion project is planned.
What should the architect recommend?
A. Wait until usage reaches 80 TB before ordering.
B. Order after one month because 25 TB will still be free.
C. Wait until physical capacity is exhausted to avoid unused investment.
D. Start expansion planning now and address the timing gap.
Answer: D
Explanation:
6 / 8
● A: The threshold is reached in two months; starting then adds a further three-month delay.
● B: A one-month wait would make installation four months away.
● C: This ignores the explicit operating threshold and risks exhausting capacity.
● D: (80 - 70)/5 = 2 months to the threshold, less than the three-month lead time. An interim measure or
accelerated delivery is needed.
Authoritative references:
● FlashBlade with Commvault Cloud Design Guide — page 14, Predicting Future Needs; forecast and
lead-time figures are original exercise assumptions; 15-page edition checked 2026-09-21; respect
explicitly stated deployment releases. Verified 2026-09-21.
10.Two proposals meet a repository's capacity and performance requirements. One has a lower
purchase price; the other includes an Evergreen modernization approach.
Which two additional cost categories should be included before concluding which proposal has the lower
five-year total cost of ownership? (Choose two.)
A. Power, cooling, and rack-space costs
B. Only the first month's subscription invoice
C. Planned refresh, migration, and operational disruption costs
D. The number of marketing awards received
E. The amount of capacity before protection overhead, without its usable equivalent
Answer: A, C
Explanation:
● A: Operating infrastructure costs continue throughout the comparison period.
● B: A single invoice cannot represent the full five-year cost stream.
● C: Refresh and transition costs can change lifecycle economics even when purchase prices differ.
● D: Awards are not a cost category.
● E: An incomparable raw-capacity number does not quantify lifecycle cost.
Authoritative references:
● FlashBlade//E Data Sheet — pages 1–2, long-term ownership cost and Evergreen modernization;
ID19250-01-en-04/26. Verified 2026-09-21.
11.In a multi-chassis FlashBlade//S design, what role do external fabric modules (XFMs) serve?
A. They execute backup policies as replacement MediaAgents.
B. They provide the external fabric needed to interconnect the expanded system.
C. They store the customer data in place of DirectFlash Modules.
D. They convert file-system snapshots into application-consistent database backups.
Answer: B
Explanation:
● A: MediaAgents are backup servers, not the fabric role of XFMs.
● B: The reference design requires XFMs for multi-chassis FlashBlade//S configurations.
● C: Fabric connectivity is distinct from flash-media storage.
● D: Application consistency is not created by a fabric module.
Authoritative references:
● FlashBlade with Commvault Cloud Design Guide — page 9, Pure Storage FlashBlade//S; multi-chassis
XFM requirement; 15-page edition checked 2026-09-21; respect explicitly stated deployment releases.
7 / 8
Verified 2026-09-21.
12.A Linux training workflow spends most of its preparation time walking a wide directory tree on
FlashBlade. Streaming the selected files is already fast.
Which change most directly targets the identified preparation bottleneck?
A. Increase the retention period of the file-system snapshots.
B. Replace the training job's output format with a compressed archive.
C. Evaluate RapidFile's parallel file-discovery tools for the directory walk.
D. Add physical capacity without changing the serial traversal process.
Answer: C
Explanation:
● A: Retention controls recovery history, not directory enumeration speed.
● B: Changing output formatting does not address the input directory walk.
● C: RapidFile parallelizes metadata traversal, matching the measured preparation issue.
● D: Additional capacity does not itself remove serial client behavior.
Authoritative references:
● Remove Bottlenecks in AI Workflows — Why Listing Large Datasets is a Challenge; RapidFile Toolkit =
Faster Unix Commands; official article checked 2026-09-21. Verified 2026-09-21.
013.A prospect's retained image archive is growing rapidly. Its existing disk environment meets current
access targets, but rack space and power are becoming the main expansion constraints.
Which discovery direction is most relevant to a FlashBlade opportunity?
A. Compare repository economics and footprint at the required capacity and access level.
B. Assume the archive needs the highest-performance configuration because images are used in AI.
C. Limit the comparison to purchase price per raw terabyte.
D. Defer storage planning until the existing system misses its latency target.
Answer: A
Explanation:
● A: The stated constraints make capacity efficiency and lifecycle operating costs relevant, even before a
latency failure.
● B: An AI association does not establish a high-performance requirement.
● C: Raw purchase price excludes usable capacity and the operating constraints described.
● D: Power and space can force expansion decisions independently of latency.
Authoritative references:
● FlashBlade//E Data Sheet — page 1, repository positioning and capacity efficiency; ID19250-01-en-
04/26. Verified 2026-09-21.
14.A design review claims that FlashBlade's internal hardware redundancy makes separate recovery
planning unnecessary.
Which two observations identify gaps in that claim? (Choose two.)
A. Hardware redundancy makes file protocols unnecessary.
B. An accepted accidentaloverwrite still requires an appropriate recovery point.
C. Adding network uplinks creates a historical copy of the data.
D. Loss of the entire site requires a copy outside that failure domain.
8 / 8
E. Capacity expansion automatically creates an independently administered backup.
Answer: B, D
Explanation:
● A: Access protocols and hardware redundancy solve different problems.
● B: Redundancy does not supply a previous logical state after an unwanted write.
● C: Network paths are not point-in-time copies.
● D: Copies confined to the lost site cannot satisfy site-loss recovery.
● E: More capacity is not itself another protected copy.
Authoritative references:
● The Pure Storage Platform for AI — page 6, Reliability and Availability; PS2571-02-en 05/24. Verified
2026-09-21.
● Ransomware Mitigation from Pure Storage — pages 1–2, recovery copies and protection against
accidental deletion; PS2100-01 07/2021; protection principles, not current enablement procedures.
Verified 2026-09-21.
● FlashBlade with Commvault Cloud Design Guide — pages 12–13, Multiple Sites; 15-page edition
checked 2026-09-21; respect explicitly stated deployment releases. Verified 2026-09-21.
15.A synthetic FlashBlade estimate contains 120 TB of text expected to reduce at 2:1 and 180 TB of
already-compressed images expected to reduce at 1:1. Ratios are logical bytes divided by physical
bytes. Ignore snapshots and other overhead.
How much physical space should be budgeted for these two data sets?
A. 150 TB
B. 200 TB
C. 240 TB
D. 300 TB
Answer: C
Explanation:
● A: This incorrectly applies 2:1 reduction to all 300 TB.
● B: This uses the unweighted mean ratio of 1.5:1, which is not valid for these unequal data sets.
● C: 120/2 + 180/1 = 60 + 180 = 240 TB. Calculate each component before adding.
● D: This ignores the stated compression benefit for the text.
Authoritative references:
● The Pure Storage Platform for AI — page 14, data-reduction footnote; original worksheet supplies the
actual ratios; PS2571-02-en 05/24. Verified 2026-09-21.