H13-622 Huawei Certified ICT Professional-Constructing Cloud Storage Solution Questions and Answers

The cluster name can be set to9000cluster.

If the InfoLink license has not been purchased, you may skip the InfoLink DNS configuration.

The front-end network gateway will serve as the default gateway of all storage nodes after being configured.

The data redundancy ratio is thedefault ratio of all directories on the OceanStor 9000.

TRUE

FALSE

a switch

b switch

c switch

The relationship between redundancy ratio and read/write performance from high to low is: N+1> N+2=N+2: 1> N+3: 1> N+4

Only non-empty directories can modify the redundancy ratio

The relationship between stripe size and read/write performance is: 512KB> 256KB> 128KB> 16KB

The modified stripe size is only valid for the file to be written. The file size before the modification remains unchanged.

True

False

Log in to the ISM to check alarms.

Check the indicators on the storage system.

Reinstall the ISM software.

Restart the sharing service.

Oceanstor 9000 can distribute requests to relatively idle nodes according to the current business load smart address of each node, thus enhancing customers.

Oceanstor 9000 g automatically senses the fault and adjusts the service to other normal nodes, thus improving the reliability of client service access.

This feature is a basic feature. You do not need to purchase a license to save costs.

Allow the client to access the OceanStor 9000 using the domain name. The client operator does not need to know the Ip address for each specific access.

Management operations, which reduces management costs

Check file system configurations, and then check storage device configurations.

Check network connection, and then check NAS device configurations.

Check device indicators, and then check system logs.

Check thestatus of storage unit disk enclosures, and then check the status of storage unit controller enclosures.

In traditional storage forms, disks are partitioned, file systems record file structures based on a tree relationship, and metadata is used to identify the use of blocks.

Objects are the basic units of OBS. Each object is acomposite of data and data property sets.

In the OBS mechanism, data and metadata are stored and managed in a distributed manner.

OBS is typically oriented to structured data applications.

Huawei big data storage products support InfiniBand/lOGE/GE networking modes.

Huawei big data storage products employ scale-out architecture.

Huawei big data storage products are based on the MPP DB.

Huawei big data storage products support dynamic storage tiering (DST).

ifconfig

ethtool

Ispci Igrep Ether

Automatic load balancing between all nodes

Automatically balance capacity and performance

No manual intervention required

Transparent to the application

abcba

bcbab

acbab

babcb

FAT32

NTFS

NFS

CIFS

Big data technology is more powerful for data processing, but management is more complex than traditional technology

Compared to data processing technology, different value data processing will be handled differently.

The big data era adopts a new way (distributed parallel processing) to process data, its component expansion is flexible, data carrying capacity is strong, and management system

Directory

Documentation

User

User group

If the NFS transport data block is small, it may increase the number of service interactions.

If the data block transmitted by NFS is large, the number of business interactions may increase.

Oceanstor 9000 NFS transport block size defaults to 1MB

Oceanstor 9000 NFS transport block size defaults to 32KB

TRUE

FALSE

3

4

5

6

Adependency relationship between service groups can be defined.

A dependency relationship between resources can be defined.

Each resource corresponds to one agent.

Resource dependency relationships can be defined in a cyclic manner.

Ports AO-1 and AO-1 are connected to the 10GE switch. A back-end network IP address must be planned for each of the twoports.

Ports B3-0 and B3-1 are connected to the front-end service switch. A front-end network IP address must be planned for each of the two ports.

NIC 0 is connected to the management and maintenance switch. You do not need to assign an IP addressfor this port.

Port IPMI is connected to the management and maintenance switch. You do not need to assign an IP address for this port.

The maximum capacity of a LUN is 16 TB.

LUNs in one RAID group belong to the same controller.

The 64 KB stripe depth is recommended for random services.

It is recommended that LUNs with heavy loads be in different RAID groups.

FCSAN

Object Storage

Cluster

10GE or infiniband

A VM disk is controlled by the VxVM andtypically located in a disk group. Each VM disk corresponds to at least one physical disk or disk partition.

A disk group is a collection of VM disks.

A subdisk is the basic unit for the VxVM to allocate disk space. To improve the efficiency ofunderlying resource use, a subdisk can be a group of inconsecutive disk blocks.

A plex consists of one or multiple subdisks on one or multiple physical disks. Each plex contains a duplicate of the data selected in a volume.

default policy has the lowest priority

default policy can be modified

Relationship between policy parameter combinations is OR

Support up to 128 file pool policies

The namespace is automatically created when the Oceanstor 9000 starts, and its name is consistent with the system name.

The namespace is the root directory of the entire system and cannot be shared and accessed, as long as the subdirectories below it can be shared and accessed.

A unified namespace provides users with a unified storage resource pool to achieve centralized management and allocation of storage space.

Each file share directory is a subdirectory of the namespace

It is recommended to use two switches to connect all nodes for redundancy, whether it is a front-end business network or a back-end storage network.

Only need to connect the storage node to the management network, other devices do not need to connect to the management network

The switches of the front-end service network can be stacked, and the back-end storage network switches are not recommended for stacking.

To meet management requirements, only the first three nodes of the cluster need to be connected to the management network.

TRUE

FALSE

WORM files can be modified and deleted

The WORM clock can only be set once. After the setting is successful, the change is no longer allowed.

When a directory is configured as a WORM root directory, all subdirectories, new directories, and new files of the directory are automatically obtained from the parent directory.

WORM attribute

WORM protection submission method only supports automatic submission

NTFS

HDFS

WushanFS

EXT3

103

104

105

106

If there is wrong goods or out of stock when unpacking and inspecting the goods, please fill in the “Cargo Error Feedback Form” and give feedback to the local office.

When the goods are inspected and inspected, the project supervisor can open the goods separately.

If the goods appear to be inconsistent with the packing list, the equipment of the unpacked box should be transferred to the room for proper storage.

Disk utilization

CPU utilization

Memory utilization

Node throughput

True

False

True

False

True

False

Front-end business network

Backend storage network

Management network

Heartbeat network

NFSv3 file sharing

NFSv4 file sharing

CIFS file sharing

FTP file sharing

The heartbeat network ports of the clustered NAS engine are correctly connected to the heartbeat network.

The service network ports of the clustered NAS engine are not connected to the customer's service network before they are configured.

The clustered NAS engine is correctly powered on and the statuses of the indicators are normal.

Hardware installation and cable connection of the clustered NAS engine are complete.

8

12

6

10

GE

10GE

Fibre Channel

SAS

Afile system has at most 10 million files.

A directory has at most 10,000 files.

A file system has at most 10 levels of directories.

A file system has at most 10,000 directories.

Data to be written is divided into multiple data blocks of the same size.

The Erasure Code algorithm calculates N parity blocks for every M consecutive data blocks.(The value of N is smaller than that of M.)

M+N data blocks are concurrently stored to different nodes.

In Erasure Code storage mode, aconcurrent failure of N-l disks is tolerated.

Storage space

User

Number of documents

User group

Front-end business network

Backend storage network

Wushan file system business management network

IPMI Management Network

If the DeviceManager process is normal, restart the DeviceManager process.

Check that the DeviceManager process is normal, then check the CM process.

Check that the CM process is normal, and then contact Huawei technical support engineers.

Front-end service network

Back-end storage network

Internal communication network

IPMI management network

The OceanStor Toolkit can be installed on any storage node.

The PC used for installation isconnected to a storage node through port eth0.

The IP address of the network port used to upload the OceanStor Toolkit must be configured on the same network segment as the management network.

After the OceanStor Toolkit is successfully installed, enter https://xx.xx.xx.xx:8088/tools in the Internet Explorer address box to go to the login page.

The N8500 supports a maximum of two mirrors for a single file system.

Resources of different storage pools can be selected to implement file system mirroring.

File system mirroring allows users to select three mirroring layout types.

In the same service scenario, mirrored-stripe and striped-mirror provide the same data reliability and performance.

The file system provides a unified file system namespace.

The name of the namespace provided by the file system is the same as the system name.

The root directory of the namespace does not provide sharing or access.

Added capacity can be directly obtained from the storage pool.

True

False

File system layer

Application layer

User management

Operating system layer

Global Cache enables the memory space on all storage servers to form a unified memory resources pool logically.

Data cached on a node cannot be hit by the read/write service of another node.

Global Cache helps improve thesharing of memory resources among nodes.

The WushanFS uses the distributed lock mechanism to implement global management of cached data. The same service data is cached only on one node. Another node that needs to access the service data can apply for alock to obtain the cached data.

A

B

C

You do not need to connect C nodes to the CE6850 switch.

Super block

Log area

Allocation unit

Folder

Create only one LUN for each RAID group.

Set LUNs to forcible write back without mirroring.

Ensure that the owning controller and the working controller of a LUN are the same.

Disks that compose a file system are evenly distributed on controllers.

True

False