I85PREP
⌘ K
Huawei HCIA-Storage study materials
Huawei HCIA-Storage study resources

Huawei HCIA-Storage practice questions

One question at a time. Build your Huawei HCIA-Storage understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1ICT Infrastructure

In a storage array, which RAID level provides striping with dedicated parity and can survive the failure of exactly one member disk?

  1. RAID 6
  2. RAID 0
  3. RAID 5 Correct answer.
  4. RAID 1

Explanation

RAID 5 stripes both data and parity across all member disks, so parity is distributed rather than dedicated to one drive. It tolerates a single disk failure; a second failure before rebuild destroys the data. RAID 6 extends this with dual parity and survives two failures.

Question 2ICT Infrastructure

Which protocol is designed specifically to carry block-level SCSI commands over a TCP/IP network?

  1. CIFS
  2. FTP
  3. iSCSI Correct answer.
  4. NFS

Explanation

iSCSI encapsulates SCSI command descriptor blocks in TCP/IP packets, presenting remote block storage as a local SCSI disk to the initiator. NFS and CIFS are file-sharing protocols, and FTP transfers files rather than exposing block devices.

Question 3ICT Infrastructure

A host needs a redundant path to a LUN so that the loss of one HBA or one fabric switch does not interrupt I/O. Which feature should be configured?

  1. Quality of service
  2. Multipathing Correct answer.
  3. Thin provisioning
  4. Snapshot

Explanation

Multipathing software such as UltraPath presents multiple physical paths to the same LUN and keeps them under one logical device, failing over automatically when a path breaks. Thin provisioning, snapshots, and QoS change capacity or performance behavior but do not provide path redundancy.

Question 4ICT Infrastructure

Which storage media type is non-volatile and has no moving mechanical parts, making it suitable for high-IOPS workloads in an all-flash array?

  1. Solid-state drive Correct answer.
  2. SAS HDD
  3. Tape cartridge
  4. Nearline SATA HDD

Explanation

A solid-state drive stores data in NAND flash with no rotating platters or actuator arms, so it delivers low latency and high random IOPS. Hard disks and tape are electromechanical media with much higher access latency.

Question 5ICT Infrastructure

What is the primary purpose of a LUN masking configuration on a storage array?

  1. Balancing I/O across controller enclosures
  2. Compressing data before it is written to disk
  3. Restricting which host initiators can access a given LUN Correct answer.
  4. Encrypting data at rest with a self-encrypting drive

Explanation

LUN masking maps a LUN to specific host initiator WWNs or IQNs, so only authorized hosts can see and use that logical unit. It is an access-control mechanism and does not compress, encrypt, or load-balance traffic.

Question 6ICT Infrastructure

Which backup type copies only the data that changed since the last backup of any kind, whether full or incremental?

  1. Incremental backup Correct answer.
  2. Synthetic full backup
  3. Differential backup
  4. Full backup

Explanation

An incremental backup captures everything modified since the most recent backup of any type, so each run is small but a restore needs the last full plus every incremental in sequence. A differential backup instead captures changes since the last full backup only.

Question 7ICT Infrastructure

In Fibre Channel SAN terminology, what does a WWN identify?

  1. A physical port or node on the fabric Correct answer.
  2. A volume group inside a file system
  3. A RAID stripe unit size
  4. A TCP port on an iSCSI target

Explanation

A World Wide Name is a unique 64-bit identifier assigned to a Fibre Channel node or port, analogous to a MAC address, and is used in zoning and LUN masking. It has no relationship to TCP ports, file system volume groups, or stripe sizes.

Question 8ICT Infrastructure

Which capability distinguishes a storage snapshot from a full clone?

  1. A snapshot converts thick LUNs into thin LUNs permanently.
  2. A snapshot can only be taken when all hosts are offline.
  3. A snapshot captures a point-in-time image that typically shares data blocks with the source. Correct answer.
  4. A snapshot always consumes a full copy of the source capacity at creation.

Explanation

Snapshots record a point-in-time state while initially sharing unchanged blocks with the source through copy-on-write or redirect-on-write, so they are fast and space-efficient. A full clone is an independent physical copy that consumes its own capacity.

Question 9ICT Infrastructure

A customer wants block storage for a database server over an existing Ethernet network without adding Fibre Channel switches. Which Huawei storage access method fits this requirement?

  1. NFS exported from a NAS appliance
  2. iSCSI over the existing IP network Correct answer.
  3. Object storage through S3 buckets
  4. FC SAN through a dual-fabric director

Explanation

iSCSI carries block-level SCSI commands over TCP/IP, so a database server can use existing Ethernet infrastructure without FC switches. NAS and object storage present file or object interfaces rather than raw block devices.

Question 10ICT Infrastructure

An array reports that a RAID 5 group has lost one member disk and is rebuilding. What is the most serious risk during the rebuild window?

  1. Read performance improves because parity reads are skipped
  2. A second disk failure in the same group causes data loss Correct answer.
  3. Snapshots of the LUN are deleted automatically
  4. The surviving disks automatically convert to RAID 0

Explanation

While a RAID 5 group is degraded, the array reconstructs lost blocks from parity on every read, and a further member failure leaves no parity to reconstruct from, so the group fails. Rebuild also stresses the surviving disks, raising the chance of that second failure.

Question 11ICT Infrastructure

A thin LUN is created with a 1 TB capacity on a storage pool, but only 200 GB is written. How is physical capacity consumed in a thin-provisioned pool?

  1. Physical consumption equals the pool size regardless of writes.
  2. The full 1 TB is reserved immediately in the pool.
  3. Capacity is consumed only when a snapshot is taken.
  4. Only the blocks actually written are allocated from the pool. Correct answer.

Explanation

Thin provisioning allocates physical extents on demand as host writes arrive, so a 1 TB thin LUN with 200 GB of data consumes roughly 200 GB of pool space. This raises pool utilization efficiency but requires monitoring so the pool does not run out of space.

Question 12ICT Infrastructure

In Fibre Channel zoning, what is the effect of placing two host ports and one array port in the same zone?

  1. Both hosts can access the array port's LUNs permitted by masking Correct answer.
  2. The array port is isolated from the fabric
  3. The zone automatically creates a RAID group
  4. The two hosts can see each other's LUNs directly

Explanation

Zoning controls which fabric ports may communicate, so ports in the same zone can log in to each other. Combined with LUN masking, each host then sees only the LUNs mapped to its initiator; zoning alone does not grant or deny access to specific LUNs.

Question 13ICT Infrastructure

Which characteristic makes erasure coding attractive compared with triple replication in a distributed storage system?

  1. It provides lower write latency in all workloads
  2. It requires fewer disks than replication for the same protection level
  3. It eliminates the need for any redundancy
  4. It achieves comparable durability with lower capacity overhead Correct answer.

Explanation

Erasure coding such as 4+2 stores parity fragments so usable capacity is roughly two-thirds of raw, versus one-third for three-way replication. The trade-off is higher CPU and network cost for encoding and reconstruction, so it is not automatically lower latency.

Question 14ICT Infrastructure

A host using iSCSI cannot log in to a target after the network team changed the target portal IP address. Which configuration must be updated on the initiator?

  1. The host's RAID controller cache policy
  2. The iSCSI target portal address Correct answer.
  3. The LUN's snapshot schedule
  4. The file system block size

Explanation

An iSCSI initiator discovers and connects to targets through configured portal addresses, so changing the target IP requires updating the portal entry on the initiator. Cache policy, block size, and snapshot schedules do not affect target discovery.

Question 15ICT Infrastructure

Which statement correctly describes the difference between SAN and NAS?

  1. SAN provides block-level access, while NAS provides file-level access. Correct answer.
  2. NAS requires Fibre Channel, while SAN requires Ethernet.
  3. SAN and NAS both provide only object-level access.
  4. SAN provides file-level access, while NAS provides block-level access.

Explanation

A SAN delivers block devices such as LUNs to hosts, which then apply their own file systems, whereas NAS delivers shared file systems through protocols such as NFS and CIFS. Either can run over different transports, so the transport is not the defining difference.

Question 16ICT Infrastructure

Why does a storage pool typically reserve free capacity beyond the sum of its LUN sizes?

  1. To satisfy a Fibre Channel zoning requirement
  2. To enable the array to convert RAID 5 to RAID 6 automatically
  3. To reduce the number of host HBAs needed
  4. To allow for metadata, snapshots, and unexpected growth without pool exhaustion Correct answer.

Explanation

Pools hold metadata and may host snapshots, clones, and thin LUN growth, so administrators keep headroom to avoid a full pool that would stall writes. Zoning, RAID-level conversion, and HBA counts are unrelated to pool headroom.

Question 17ICT Infrastructure

A backup policy runs a full backup on Sunday and differential backups on Monday through Saturday. How many backup sets are needed to restore Thursday's data?

  1. Sunday's full set plus Monday, Tuesday, Wednesday, and Thursday differentials
  2. Sunday's full set plus Thursday's differential set Correct answer.
  3. Only Sunday's full set
  4. Only Thursday's differential set

Explanation

A differential backup contains all changes since the last full backup, so Thursday's set already includes Monday through Wednesday changes. Restore therefore requires the Sunday full plus the Thursday differential only.

Question 18ICT Infrastructure

Which action best reduces the risk of silent data corruption being returned to an application reading from a LUN?

  1. Enabling end-to-end data integrity checking such as T10 PI Correct answer.
  2. Disabling multipathing on the host
  3. Increasing the LUN's snapshot retention period
  4. Using thick provisioning for all LUNs

Explanation

T10 protection information appends checksums to each block so the array and host can detect corruption introduced along the path or in media. Snapshots, multipathing, and provisioning type address availability and capacity, not silent corruption detection.

Question 19ICT Infrastructure

An administrator needs a point-in-time copy of a production LUN that can be mounted on a test server and modified without affecting production. Which mechanism is most appropriate?

  1. A backup to tape restored to the test server
  2. A clone of the LUN presented to the test server Correct answer.
  3. A zone change adding the test server to the production zone
  4. A read-only snapshot mounted directly by the test server

Explanation

A clone is an independent writable copy, so the test server can modify it without touching production data. A read-only snapshot cannot be written, a tape restore is slower and unnecessary here, and zoning changes would expose production LUNs.

Question 20ICT Infrastructure

Which statement about the relationship between RAID and hot spare disks is correct?

  1. A hot spare doubles the write performance of the RAID group
  2. A hot spare is an idle disk that is automatically substituted when a member fails Correct answer.
  3. A hot spare is a member of the RAID group and stores parity
  4. A hot spare is required for RAID 0 to function

Explanation

A hot spare is not part of the RAID group's addressable capacity; it waits idle and is automatically used to replace a failed member so the group can rebuild. It improves resilience rather than performance and is irrelevant to RAID 0, which has no redundancy.

Question 21ICT Infrastructure

An array's cache is configured in write-back mode. What happens to host writes acknowledged to the host but not yet written to disk if the array loses power?

  1. The writes are lost unless cache is protected by a battery or capacitor Correct answer.
  2. The writes are reconstructed from RAID parity
  3. The writes are automatically replayed from the host's page file
  4. The writes are recovered from the host's snapshot

Explanation

Write-back caching acknowledges writes once they reach cache, so unprotected cache contents would be lost on power failure. Arrays protect this window with battery-backed or supercapacitor-backed cache that flushes to persistent media.

Question 22ICT Infrastructure

A storage pool shows 95 percent utilization while thin LUNs are still growing. Which consequence is most likely if no action is taken?

  1. The array automatically shrinks thin LUN capacities
  2. New writes to thin LUNs may fail once the pool is exhausted Correct answer.
  3. All snapshots are converted to clones
  4. The pool converts itself to a RAID 6 group

Explanation

Thin LUNs draw from pool free space as writes arrive, so an exhausted pool cannot allocate new extents and writes fail or the LUNs go offline. Administrators should expand the pool, delete unneeded snapshots, or reclaim space before exhaustion.

Question 23ICT Infrastructure

A customer requires that a LUN remain accessible if an entire storage controller fails. Which design element is essential?

  1. Increasing the LUN's snapshot retention count
  2. Enabling thin provisioning on the LUN
  3. Using a dual-controller array with LUN ownership failover and multipathing on the host Correct answer.
  4. Configuring the LUN on a single controller with no failover path

Explanation

Dual controllers with transparent failover let the surviving controller take over LUN ownership, and host multipathing redirects I/O to the remaining path. Provisioning type and snapshot retention affect capacity and recovery points, not controller-level availability.

Question 24ICT Infrastructure

When comparing copy-on-write and redirect-on-write snapshots, which statement accurately describes a key difference?

  1. Both mechanisms require a full physical copy of the source at creation
  2. Redirect-on-write cannot be used with thin provisioning
  3. Copy-on-write preserves the original block by copying it before overwrite, while redirect-on-write sends new writes elsewhere and leaves the original block intact Correct answer.
  4. Copy-on-write writes new data to a new location, while redirect-on-write overwrites the original location

Explanation

With copy-on-write, the first write to a protected block copies the old data to a snapshot area before the new data overwrites it. Redirect-on-write instead redirects new writes to new blocks and leaves the original block untouched for the snapshot, which reduces write amplification but fragments the source layout.

Question 25ICT Infrastructure

A four-node distributed storage cluster uses a 4+2 erasure coding scheme. Two nodes go offline simultaneously. What is the resulting state of the affected data?

  1. The data is permanently lost because more than one node failed
  2. The data is converted to triple replication automatically
  3. The data remains readable because two parity fragments can reconstruct the missing data Correct answer.
  4. The data is readable only after the cluster is reduced to three nodes

Explanation

In a 4+2 scheme, any four of the six fragments can rebuild the original data, so losing two fragments still leaves enough information to read and reconstruct. Data would be lost only if more than two fragments in the same protection group became unavailable.

Question 26ICT Infrastructure

Which three planes are typically distinguished in the architecture of an enterprise storage system such as Huawei OceanStor?

  1. Front-end plane, back-end plane, and power plane
  2. Control plane, data plane, and management plane Correct answer.
  3. Read plane, write plane, and cache plane
  4. Host plane, fabric plane, and disk plane

Explanation

Storage array architecture is commonly described in terms of a control plane that handles metadata and configuration, a data plane that moves user data through front-end ports, cache and back-end loops, and a management plane for operations tools. Separating these planes clarifies where performance bottlenecks and management traffic occur.

Question 27ICT Infrastructure

Which disk interface is designed primarily for direct-attached enterprise HDDs and SSDs inside a storage enclosure, using dual ports for high availability?

  1. Serial Attached SCSI (SAS) Correct answer.
  2. Fibre Channel (FC)
  3. Ethernet
  4. Serial ATA (SATA)

Explanation

SAS is the mainstream drive interface inside enterprise storage enclosures and supports dual-ported drives so that a path remains if one expander or controller link fails. SATA drives are single-ported and aimed at cost-sensitive use.

Question 28ICT Infrastructure

What is the primary purpose of the back-end loop in a dual-controller storage array?

  1. To provide out-of-band management access for administrators
  2. To replicate data to a remote site over IP
  3. To carry data between the controllers and the disk enclosures Correct answer.
  4. To connect hosts to the array through the SAN fabric

Explanation

The back-end loop connects controller enclosures to disk enclosures and carries I/O to and from the physical drives. Front-end ports serve host connectivity, and management ports serve administrators.

Question 29ICT Infrastructure

Which statement best describes the role of a storage array's cache in read operations?

  1. It compresses all read data before returning it to the host
  2. It holds recently read data so that repeated reads can be served without accessing the disks Correct answer.
  3. It permanently stores the most frequently read blocks instead of the disks
  4. It encrypts read data before it leaves the array

Explanation

Cache is volatile memory that keeps recently accessed data to reduce disk access latency for repeated reads. It is not persistent storage, and compression or encryption are separate features.

Question 30ICT Infrastructure

In a RAID 1 configuration, how is data protected?

  1. Data is written identically to two disks so either copy can serve reads Correct answer.
  2. Data is striped across all member disks with parity
  3. Data is written to a single disk and mirrored only during rebuild
  4. Data is split into blocks with two parity blocks per stripe

Explanation

RAID 1 mirrors data by writing identical copies to paired disks, so a read can be served from either member and the array survives the loss of one disk. Parity-based schemes belong to RAID 5 and RAID 6.

Question 31ICT Infrastructure

Which protocol is used by a host to access file shares over an Ethernet network with built-in authentication and locking mechanisms?

  1. Non-Volatile Memory Express over Fabrics (NVMe-oF)
  2. Fibre Channel Protocol (FCP)
  3. Server Message Block (SMB) Correct answer.
  4. Internet Small Computer Systems Interface (iSCSI)

Explanation

SMB is a file-sharing protocol that provides shared access to files with authentication, locking and permission semantics. FCP and iSCSI carry block-level SCSI commands, so they do not provide file sharing.

Question 32ICT Infrastructure

Which metric most directly describes the number of input/output operations per second a storage system can sustain?

  1. Bit error rate
  2. Capacity in terabytes
  3. IOPS Correct answer.
  4. Mean time between failures

Explanation

IOPS counts completed I/O operations per second and is the standard measure of storage transaction throughput. Capacity, error rate and reliability figures describe other characteristics.

Question 33ICT Infrastructure

What does the acronym LUN refer to in block storage?

  1. A logical unit presented to a host as a block device Correct answer.
  2. A logical network segment used for iSCSI traffic
  3. A license unit that enables array features
  4. A physical disk enclosure in the back-end loop

Explanation

A LUN is a logical unit of block storage that a host sees as a disk device. It may be backed by part of a RAID group or a storage pool rather than by a single physical disk.

Question 34ICT Infrastructure

A host must access a LUN over an existing IP network using standard SCSI commands and CHAP authentication. Which host-side component is required?

  1. An iSCSI initiator Correct answer.
  2. A SAS expander
  3. An NFS client
  4. A Fibre Channel host bus adapter

Explanation

An iSCSI initiator encapsulates SCSI commands in TCP/IP and supports CHAP authentication with the target. A Fibre Channel HBA requires an FC fabric, and NFS is a file protocol.

Question 35ICT Infrastructure

An administrator wants to improve sequential read performance for a LUN whose data is spread across many disks. Which RAID characteristic provides this benefit?

  1. Dedicating one disk as a hot spare
  2. Striping data across multiple member disks Correct answer.
  3. Mirroring the data on two disks
  4. Reducing the stripe size to a single block

Explanation

Striping spreads consecutive data blocks across member disks so that multiple disks can serve a sequential request in parallel, raising throughput. Mirroring improves availability and read options but does not spread a single stream across many disks.

Question 36ICT Infrastructure

A storage pool with 10 TB of raw usable capacity contains thin LUNs whose subscribed capacity totals 30 TB. What does this configuration illustrate?

  1. Over-subscription enabled by thin provisioning Correct answer.
  2. A RAID 10 configuration
  3. Thick provisioning with guaranteed allocation
  4. Erasure coding failure

Explanation

Thin provisioning lets administrators present more logical capacity than is physically available because space is allocated only as data is written. This over-subscription requires monitoring so that the pool does not run out of space.

Question 37ICT Infrastructure

Which feature allows a storage array to keep serving I/O after a controller failure without administrator intervention?

  1. Manual LUN migration to a spare controller
  2. Controller active-active redundancy with path failover on the host Correct answer.
  3. Snapshot rollback of the affected LUN
  4. Write-through cache mode

Explanation

In an active-active dual-controller array, the surviving controller takes over the failed controller's LUNs, and multipath software on the host switches to the remaining path. This happens automatically and preserves access.

Question 38ICT Infrastructure

Which statement accurately describes the difference between a backup and a disaster recovery plan?

  1. Backup protects data copies, while disaster recovery defines how services are restored at an alternate site Correct answer.
  2. Backup restores only files, while disaster recovery restores only databases
  3. Disaster recovery replaces the need for backups at the primary site
  4. Backup and disaster recovery are the same activity performed by different teams

Explanation

Backup creates recoverable copies of data, whereas disaster recovery is a broader plan covering alternate sites, replication, runbooks and recovery objectives. A complete strategy uses both together.

Question 39ICT Infrastructure

A company must retain monthly backups for seven years at low cost while keeping the last 30 days quickly restorable. Which approach fits best?

  1. Store all backups on the primary array's SSD tier
  2. Store all backups on host local disks
  3. Keep only the most recent backup and delete older copies
  4. Tier recent backups on fast storage and archive older backups to lower-cost media Correct answer.

Explanation

A tiered backup strategy keeps recent recovery points on fast storage for quick restore and moves older retention copies to cost-efficient archive media. This balances recovery time against long-term retention cost.

Question 40ICT Infrastructure

Which capability allows a storage system to serve the same LUN to multiple hosts simultaneously while coordinating access?

  1. Disk scrubbing
  2. Cluster-aware shared LUN access with SCSI reservations or persistent reservations Correct answer.
  3. Single-initiator zoning only
  4. Write-through caching

Explanation

Shared access to a LUN by multiple hosts requires coordination through SCSI reservations or persistent reservations so that cluster software can arbitrate ownership. Without this, concurrent writes could corrupt data.

Question 41ICT Infrastructure

What is the main benefit of using multipath I/O software on a host connected to a dual-controller array?

  1. It converts block LUNs into file shares
  2. It provides redundant paths and can balance I/O across them Correct answer.
  3. It encrypts data in transit between host and array
  4. It increases LUN capacity automatically

Explanation

Multipath software presents multiple physical paths as one logical device, so I/O continues if a path fails and load can be distributed across paths. It does not change capacity or protocol type.

Question 42ICT Infrastructure

A customer needs to protect a LUN against accidental deletion and restore it quickly to a state from two hours ago. Which feature is most appropriate?

  1. Disk spin-down
  2. Link aggregation
  3. Snapshot with rollback capability Correct answer.
  4. Thick LUN provisioning

Explanation

A snapshot captures the state of the LUN at a point in time and can be rolled back to restore that state quickly after accidental changes. Provisioning and network features do not provide point-in-time recovery.

Question 43ICT Infrastructure

Which protocol provides block-level access over a Fibre Channel fabric using SCSI command semantics?

  1. NFS
  2. HTTP
  3. FCP Correct answer.
  4. SMB

Explanation

FCP carries SCSI commands over Fibre Channel, giving hosts block-level access to LUNs. NFS and SMB are file-sharing protocols, and HTTP is an application protocol.

Question 44ICT Infrastructure

An array reports that a disk in a RAID 6 group has failed and the group continues to serve I/O. What is the immediate operational priority?

  1. Convert the group to RAID 0 to improve performance
  2. Replace the failed disk so the group returns to full redundancy Correct answer.
  3. Rebuild the group onto the failed disk
  4. Disable the hot spare to save power

Explanation

RAID 6 tolerates two simultaneous failures, but with one member already lost the group has reduced protection. Replacing the failed disk promptly restores full redundancy before a second failure occurs.

Question 45ICT Infrastructure

Which statement best explains why storage arrays use battery-backed or flash-backed cache protection?

  1. To keep acknowledged writes that are still in cache from being lost during a power failure Correct answer.
  2. To increase the capacity of the cache beyond physical memory
  3. To replace the need for RAID protection
  4. To compress data before it reaches the disks

Explanation

In write-back mode the array acknowledges writes once they are in cache, so cache contents must survive a power loss. Battery or flash modules preserve cached data until it can be written to disk.

Question 46ICT Infrastructure

A four-node distributed storage cluster with a 2-replica policy loses two nodes that held copies of the same data. What is the resulting state?

  1. All data remains available because replicas are independent
  2. The cluster converts the data to erasure coding automatically
  3. The affected data becomes unavailable until a node returns Correct answer.
  4. The cluster automatically creates a third replica on the failed nodes

Explanation

With a two-replica policy each data block has exactly two copies, so losing both nodes that hold those copies leaves no available replica and the data is temporarily inaccessible. Additional replicas or erasure coding would be needed to survive that loss.

Question 47ICT Infrastructure

An administrator observes high write latency on a thin LUN while pool utilization is low. Which cause is most consistent with these symptoms?

  1. The pool has insufficient free capacity
  2. The cache is in write-through mode or the LUN is receiving many small random writes Correct answer.
  3. The host is using a single-path connection with no failover
  4. The LUN is over-provisioned with too much free space

Explanation

Low pool utilization rules out space exhaustion, so latency likely comes from write behavior or cache policy. Small random writes require more disk operations per unit of data, and write-through mode forces each write to disk before acknowledgment.

Question 48ICT Infrastructure

A customer requires that no single disk failure cause data loss and that the array tolerate the simultaneous loss of two disks in the same group. Which RAID level meets these requirements?

  1. RAID 5
  2. RAID 0
  3. RAID 6 Correct answer.
  4. RAID 1

Explanation

RAID 6 uses two independent parity blocks per stripe, allowing the group to survive two simultaneous member failures. RAID 5 tolerates only one failure per group.

Question 49ICT Infrastructure

Which design choice best protects against a site-level outage while keeping recovery time low for a tier-1 database?

  1. Synchronous or near-synchronous replication to a remote array with automated failover Correct answer.
  2. Nightly tape backups stored offsite
  3. RAID 6 protection within the local array
  4. Snapshots retained on the same array

Explanation

Site-level protection requires a copy at a separate location, and replication with automated failover keeps recovery time low. Local RAID and on-array snapshots do not survive loss of the entire site.

Question 50ICT Infrastructure

An administrator must ensure that a LUN presented to a clustered application is not written concurrently by two hosts that are unaware of each other. Which control is most appropriate?

  1. LUN masking combined with SCSI persistent reservations managed by the cluster Correct answer.
  2. Adding more front-end ports to the array
  3. Increasing the LUN's thin provisioning capacity
  4. Enabling write-back cache on the array

Explanation

LUN masking limits which initiators can see the LUN, and persistent reservations let cluster software arbitrate ownership so that only the rightful node writes. Capacity, cache and port changes do not prevent conflicting writes.