Download Free HCIA-Big Data V3.0 H13-711_V3.0-ENU Sample Questions

test2022-04-30T06:06:11+00:00

By test H13-711_V3.0-ENU free dumps, H13-711_V3.0-ENU Sample Questions, HCIA-Big Data V3.0 0 Comments

Want to practice online Free HCIA-Big Data V3.0 H13-711_V3.0-ENU Sample Questions? Candidates are required to answer H13-711_V3.0-ENU exam to achieve the success. Now prepare for your HCIA-Big Data V3.0 H13-711_V3.0-ENU exam with new exam dumps, which contents are correct and valid. You will find H13-711_V3.0-ENU exam dumps with actual questions and verified answers are great for preparation.

Take a free H13-711_V3.0-ENU practice exam right now to see how prepared you are!

Page 1 of 12

1. As the following figu r e, regard i ng the flume data transmission architecture, what are the components at the "?"?

Intercept o r

Channel p rocessor

Chan n el selector

None of the above is correct

2. In the Fusioninsight HD cluster t h ree-tier network, the ma n agement n od e, control node, and data node a re re c ommended t o be instal l ed in di f ferent network segments to i mprove reliabilit y .

T rue

False

3. What d a ta does Fusion Insight Mana g er back up re g ularly? (Multiple c h oice)

NameNo d e

LDAPSe r v er

OMS

DBServi c e

4. Regarding DataSet, which of the following statements is incorrect?

DataSet is a strongly typed c o llection composed of objects of a specific domain

DataSet can perform most operations without deserialization

DataSet needs to be deseriali z ed to perform sort, fi l te r , shu f fle and other ope r ations

DataSet is highly similar to RDD, and its performance is better than RDD

5. Which o n e of the foll o wing factors have contributed to the v igor o us development of the big data e r a?

Reduction of hardwa r e cost and inc r ease of network bandw i dth

The rise o f cloud com p uting

The popu l arity of smart terminals and the increase in social d e mand

All above are correct

6. Assuming that HDFS only sav e s 2 copies when writing data. D uring the w riting process, the HDFS

Client first w r ites the data to DataNodel1 and then writes the data to DataNo d e2.

T rue

False

7. What is t h e module used to mana g e the active and standby status of Loader Server p r ocesses in

Loa d er?

Job Scheduler

HA Manager

Job Manager

Resource Manag e r

8. Hive is a data ware ho use software based on h a doop, which c an query a n d manage P B-level distributed data.

Which of the following de s criptions of H ive features is wrong?

Only supports MapRe d uce computing engi n e

E a sy to u se and easy to program

Can directly acce s s HDFS fil e s and HBase

Flexible and conveni e nt ETL (extract/transform/load)

9. What d a ta does Fusioninsight Manag e r back up regul a rly? (Multiple choice)

NameNo d e

LDAP

OMS

DBServi c e

10. Wh e n the size of a Region g r ad ua lly increases in HBase, it may be trimmed.

T rue

False

Page 2 of 12

11. The foll o wing figure shows the configuration of HDFS tiered storage. If t h e number of copies of the block is 4, w hich of the f o llowing statement is wrong?

All 4 blocks will be written to RAM_DISK

If i t i s the other copy fails to read except the first cop y , t r y to write to the s torage type specified in "Alternative Storage Policy for Copy"

The first copy of the b l ock will be w ritten into RAM_DISK, a n d the remain i ng copies will be written into DISK

If the fir s t copy of the bl o ck fails to write to RAM_DISK, t r y to write to the storage type specified by the "alternative storage strategy"

12. In a Hu a wei Fusioni ns ight HD cl u s te r , which of the followi n g servic e s c an the Spark s e rvice read data from? (Multiple choice)

Y ARN

HDFS

Hive

HBase

13. What p r ocesses are included in the HBase service of F u s ion i nsight HD? (M u ltiple choice)

HMaster

Sl a ve

HRegi o nServer

Data No d e

14. Big data requires a chan g e in tra d itional industry thinkin g . D ata collection and analysis should be reg a rded as an important part of business processes. The data end drives b usiness process optimization, realizes intelligence and a utomation, and realizes cross-bord e r expansion b y relying on data assets.

T rue

False

15. What do e s Fusioninsi g ht HD H B ase use by de f a ult as its underlying file s t orage sy s tem?

HDFS

Hadoop

Memory

MapRed u ce

16. Which of the following descriptions of Zookeep e r features is wro n g?

The number of Zookeeper n o des must be an odd numb e r .

Updates sent by the client w ill be applied in the order in which they were sent.

Message update can only succeed or fail, there is no intermediate state.

A messa g e needs to be recei v ed by more than half of the s e rvers. It c an be successf u lly written to di s k

17. The flume architecture shown in the figure be l ow is mainly used to import data from outside the cluster into the cluste r .

T rue

False

18. Spark a nd Ha d oop a re not suitable for iterative calculation computing.

T rue

False

19. Which i n stance must the Load e r i n stance in F u s ionlnsight HD be deplo y ed with?

DataNode

RegionS e rver

Resourc e Manag e r

NodeMa n ager

20. What items can Fusionlnsight Manag e r perfo r m health checks on? (Multiple choice)

Host

Serv i ce

Role

Examples

Page 3 of 12

21. In Y ARN, which of the following tasks is the "slave" node r es ponsible for? (Multiple choice)

Unified mana g ement a nd alloc a tion of all resources in the cl u s ter

Supervise the life cycle manageme n t of Container

Manage logs and ancillary services used by di f ferent applic a tions

Monitor the resource u sage (mem o r y , CPU, etc.) of each Container

22. The Container of the Spark task can run multiple tasks.

T rue

False

23. All Kafka messages will be persisted to the hard disk. A t t h e s ame time, Kafka en s ures data reliability by setting Replicetion on T opic Partitio.

T rue

False

24. The secondary index provides HBase with the ability to index according to the values of certain columns. T h e s econdary index fi r st c hecks the index table, and then locates the position in the data tabl e , without a full table scan.

T rue

False

25. What is the purpose of the HBase cluster to perform compac t ion regularl y ? (Multiple choice)

Reduce the number of files in t he s a me Region and Column F amily

Reduce the number of files in the s a me Region

Improve data reading performance

Improve data writing capabilities

26. Wh e n u sing the configuration pl a nning tool to plan the Fusioninsight HD V100R0 0 2 C 50 cluste r , you can use the V100R 2 00 2 C60 configu r ation p l anni n g tool to plan.

T rue

False

27. Fusioni ns ight is a unified pl a tform for Hu a wei's enterprise- lev el big data storage, qu e r y , and analysis. It can help companies q u ickly build a massive data information processing system, and discover new value points and busine s s oppor t unities through r eal-time and non - real-ti m e analysis a n d mining of massive information data.

T rue

False

28. In the task scheduling of Y ARN. Once the ApplicationMaster applies for the resource, it will communicate with the c o rrespo n ding ResourceM a nag e r and a s k it t o start the task.

T rue

False

29. Kafka is a high-throughput, d istributed, publish-and-subsc r ibe messaging system. Using kafka technolog y , a large-scale messag i ng system can be built on a c h eap PC Serve r .

T rue

False

30. What c onfiguration files can the Fusioninsight HD LLD c onf i guration pl a nning tool g e nerat e ? (Multiple choice)

Monitoring alarm threshold configu r ation file

Cluster i n s tallation t e mplate file

HDFS and Y ARN configuration fil e s

The configuration file CheckNode s .Config requ i red to execute Precheck

Page 4 of 12

31. Which of the following state m ents about Huawei's big data s olution is correct? (Multiple choice)

Farmer is a data service framework

GaussDB is an open-source data bas e product

Fusionins i ght Manager is a d i s tributed system management framew o rk, administra t ors can control distributed clusters through multiple access points

Fusioninsight HD is an enh a nced vers i on based on the ope n-s ource big data software Had o op

32. Which of the following is not a role or service i n v olved in the HBase read data process?

HDFS

Zookeeper

HMaster

HRegi o nServer

33. What is the resource manageme n t framework that Spark c o mes with?

Standalo n e

Mes o s

Y ARN

Docker

34. Which of the following OS ve r sions are recommended to build a Fusioninsight V1R2C60 cluster? (Multiple choice)

SU S E 1 1 SP1/SP2 / S P 3 for AMD64&lnter64

CentOS6.6

Redhat-6 . 4-x86_ 6 4

RedHat - 6.5-x86_ 6 4

RedHat-6 . 7-x86_ 6 4

Ubuntu6.3

35. What services can Huawei DWS provide to customers? (Multip l e choice)

Support GDS tools to speed u p data storage

E n s ure high reliability of data and system

Second-level response to trillions of data association anal y sis

Unified mana g ement console

36. In Fusionlnsight HD cluster planni n g, what kind of scenario is suitable for the i n tegrat e d deployme n t of managem e nt node & control node & d ata node?

Below 30 nodes

More than 100 nodes

100- 2 00 n odes

More than 200 nodes

37. Hive d o es not check whether the data conforms to the sch e ma when it is loaded. Hive follows the schema on read (r e ad m ode). Hive only checks a nd parses specific data fi e lds and schemas when it is read.

T rue

False

38. Which of the following scenarios is HDFS not suitable for? (Multiple choice)

Streaming data access

Large amount of small file storage

Large file storage and access

Random write

39. HiveSe rv er compiles and parses the HQL statements submitted by the user into correspond i ng Y arn tasks, Spark tasks or HDPS opera t io ns , thereby completing data extraction, conversion, and a n alysis.

T rue

False

40. Which of the following sub-p r oducts does the Fusi o ninsight family include? (Multiple choice)

Fusioninsight Miner

Fusioninsight Farmer

Fusioninsight HD

GaussDB 200

Page 5 of 12

41. In the Fusioninsight Hadoop cluste r , through df-hT query on a node, the partitions seen include the following:

/var/log

/srv / BigData

/srv/BigData/had o op/dat a 5

/srv / BigData/solr/solrser v er3

/srv/BigData/dbdat a om

What is the p lanned combination of the best Raid level for the disks corresponding to these partitions?

Raid0 Raid1 Raid0 No n-Rai d 1

Raid1 Raid1 No n -Raid Non- R aid R aid1

Raid0 Ra i d0 Raid0 R a i d0 Raid0

Non- R aid Non - Raid No n-Raid N on - Raid Rai d 1

42. U n der t h e HDFS fed e ration mech a nism, metadata between NameN o des is not share d .

T rue

False

43. When the number of n odes in the Zookeep e r cluster is 5 node s , how ma n y nodes are equivalent to the disaster tolerance of the cluster?

None of the above

44. In Hado o p, if yarn.sc h edul e r .capacit y .root.QueueA.minimum-u s er-limit-percent is set t o 50, which of the following is wrong?

A user s u bmits a task c an use 100% of QueueA's resources

E a c h user in QueueA can only get up to 50% of the resources

If there are already 2 user tasks running in QueueA, then t h e task submitted by the third user needs to wait for the release of resources

QueueA must ensure that each u s er gets at least 50% of the r esources

45. Which one of the following is not a role or service involved in t h e HBase re a d data proc e ss?

HDFS

HRegion S erver

HMaster

ZooKee p er

46. Spark o n Y ARN-client is sui t able for producti o n environm e n t s because you can see the output of the

APP faste r .

T rue

False

47. T o redu c e the amount of data w r itten to disk, mapr e duce supports MOF (mapoutfile) compression before writi n g.

T rue

False

48. In the MRS cluste r , which of the following com p onents does Spark mainly interact wit h ?

Zookeeper

Y arin

Hive

HDFS

49. In order to consider perfor m ance optimization, it is recom m end e d to de p loy LdapServer and

KrbServer o n the same node in all clusters.

T rue

False

50. Which of the following state m ent about ZKFC is incorrect?

ZKFC (ZKFailoverCon t roller) is used as a Zookeeper cluster client to monitor the stat u s information of the NameN o de

The ZKFC process needs to be deployed in the NameN o de node a nd the Zookeeper Lea d er no d e

The Standby NameN o de perceives the stat u s of the Active N ameNo d e th r ough Zook ee pe r . Once the Act i ve NameNode goes down, the Standby NameNode will per f orm the upg r ade o p erati o n

The ZKFC of HDFS NameNo d e connects to Z o okeeper a nd saves the host name and other information in Zookeep e r

Page 6 of 12

51. Spark d i vides the st a ges according to the dep e ndencies of t he RDD. The scheduler starts from the end of the DAG graph and travers e s the entire dependency ch a in in reverse. It will disconnect when it encounters a narrow dependenc y , and add it to the current stage when it encounters a wide de p en d e n c y .

T rue

False

52. When vi e wing the par t ition de t ails o f a topic in Kafka, which command is used as follows?

bin/kafka-topic.sh -create

bin/ka f ka- t opic.sh -li s t

bin/kafka - topic.sh -describe

bin/kafka - topic.sh -delete

53. In the Fusioninsight HD sy s tem, which component does the flume data stream not need to pass throu g h in the nod e ?

s ink

topic

Source

Chan n el

54. What methods or interfaces does Load e r provide to implement job man a gement? ( M ultiple choice)

WEB UI

Linux command line

REST int e rface

Java API

55. R e ga r ding the di f ference and connection bet w een Spark SQL & Hive, which of the f o llowing statements a re correct? (Multiple choice)

Spark SQL is compatible with most Hive syntax and functions

Spark SQL cannot use H ive's cust o m functions

The execution engi n e of Spark SQL is Spark c o re, and the d efault execution engi n e of Hive is MapRe d uce

Spark SQL relies on Hive metadata

56. Which o ption is wro n g in t he foll o wing d escription of Hive l o g collection on the Fusion i nsight Man a ger interface?

Spec i fic users can be designa t ed for log collect i on, for example, only logs gene r ated b y UserA can be down l oa d ed.

Y ou can specify a time period for l o g collection, for examp l e, only collect logs from 2016-1 - 1 to 201 6 -1- 1 0.

Y ou can specify the instance for log collection, for exampl e , specify the log collection o f metstore.

Y ou can specify the node IP for log collection, for example, only downlo a d the logs of a certain I

57. Wh e n Fusioninsight Manag e r is connected to an external management platform, which of the following int e rfaces does i t support? (Multiple choice)

SNMP

V PN

FTP

S y s log

58. The e nh anced featu r es of Zoo k eeper include a dding the a u dit log of the deleted e p h e meral node t o the audit log.

T rue

False

59. In the Fusioninsight HD system, using the St r eaming client Shell command to view the topology or submit the t o pology fails, which of the following po s itioning m ethods are cor r ect? (Multiple choice)

Check the client exception stack to determine w hether the r e is a problem with the client

Check the running log of the main Nimbus to determine wheth e r the N i mbus server is abnormal

Check the Supervisor running l o g to determine w hether the Supervisor is abn o rmal

Check W orker ru n ni n g log

60. The e m erg e nce of HFS s olv e s t h e need to s tore a lar g e n u mber of s mall files (below 10MB) in H D FS. At the same time, some mixed scenes of large fil e s (above 10MB) should be stored

T rue

False

Page 7 of 12

61. Which of the following are the main functions of Fusionlnsight Manager? ( Multiple choice)

Data integration

System Mana g ement

Security Managem e nt

Service g overnance

62. Which compon e nt c ontrols the primary and secondary arbi t ra t ion in H D FS?

Zooleeper Failover Controller

NodeMa n ager

Resourc e Manag e r

HDFS Client

63. Which of the following commands are of the set type? (Mult i ple choice)

scard

sunion

zcount

hexists

64. Re g ardi n g the Kafka message c o nsumer read i ng the message, which o f the following is wrong?

The blue is a topic of Kafk a , which can be un d er s tood as a queue, and ea c h grid re p resents a message.

The messages genera t ed by the producer a re placed at the end of the topic one by one.

The consumer re a ds the messages sequentially from right to left.

Consum e r uses o f fset t o record the read position

65. Re g ardi n g the Flink b arri e r , which of the following is inco r rect?

The barrier is periodically in s erted i n to the data stream a n d fl o ws with it as part of the data stream

Barrier is the core of Flink snapshot

A barrier separa t es the data of the snapshot of this cycle from the data of the next cycle of snapshots

When the barrier is inserted, the data flow will b e tempora r ily blocked

66. Which of the following functions c a n the Kafka Cluster Mirroring tool achieve?

Ka f ka cr o ss-cl u s ter data sy n chronization soluti o n

Kafka single cluster d ata backup

Kafka but data rec o ve r y in the cluster

None of the above is correct

67. Which of the following descriptions about Fusioninsight CTBase is wrong?

CTBase's read a nd wri t e data interf ac e, unified encapsulation of the line-defined interf ac e, automatic merging a n d parsing of cold fields, no need to me r ge and inte r p ret in the ap p lication

CTBase i s a clu s ter table development framework based on HBase

CTBase provides a s et of W ebUI for metadata d e finition, pro v ides a medical-only table design tool, and reduces the di f ficu l ty of t a ble design

CTBase's java API provides a s et o f interfaces for HBase co n nection pool manageme n t, and internal connection sharing r edu c es the di f f iculty of client application d ev elopment.

68. Which of the following state m ents i s correct about the int e raction between Flink and other components? (Multiple choice)

The impl e mentation of Flink's checkpoint relies on Zookeep e r

Fli n k can send the rec e ived components to Kafka

The oper a tion of Flink tasks r elies on Y ARN for resource sc h eduling ma n agement

Flink reads and writes da t a in the HDFS file system

69. Which of the following are the key features of the HDFS architecture? (Multiple choice)

Robust mechanism

Multi-mode access m e c hanism

HA high reliability

Metadata persistence mechanism

70. Y ARN mana g es cluster resources throu g h ResourceMan ag e r .

What are its main functions? (Multiple choice)

Cluster r e s ource sch e duling

Applicati o n managem e nt

Log man a gement

None of the abo v e statements

Page 8 of 12

71. What a r e the characteristics of Spark? (Multiple choice)

Fa s t

skil l ful

Spirit

Light

72. In creating a Loa d er j ob, wh i ch of the following steps can be used to set the filter typ e ?

Input settings

Conversion

Output

Basic inf o rmation

73. If you want to add 1 to the digital value stored in the Ke y , w hich comm a nd should be used?

incr

decr

incrby

decrby

74. Which of the following scenarios' data collection is F lume su i table for? (Multiple choice)

Collect data from data sourc e s su c h as Thrif t , A v ro, S y s log, K a f k a, e t c .

Local file data collecti o n

Log collection gene r at e d by the application syst e m

Real-time data collecti o n of large amounts of data

75. What is the purpose of the HBase cluster to perform compac t ion regularl y ? (Multiple choice)

Reduce the number of files in t he s a me Region and Column F amily

Improve data readi n g p erformance

Reduce the file data of the same ColumnFamily

Reduce the number of files in the s a me Region

76. If the phys i cal mem o ry util i z a tion of some Containers exc e e ds the configur e d me m ory threshold, but the total m e mory utiliz a tion of all Containers does not exceed the set Nod e Manag e r memory threshold, then the Containers with excess i ve m emory usage can still run continuousl y .

T rue

False

77. Which of the following commands delete files?

d f s -clear

d f s -del

dfs -rm

dfs - I s

78. The r e a r e 230 no d es i n the Fusionlnsight HD cluster of a universit y . When clustering is performed, which of the following pr oc ess e s should be deplo ye d on the data nodes? (Mu l tiple choice)

DataNode

NameNo d e

NodeMa n ager

RegionS e rver

DBServer

SolrServer

79. What stora g e formats does Hive support include? (Multiple c hoice)

T extFile

RCFile

Parquet

Sequenc e File

80. Which of the following descriptions of HBase Fi l eStream (H F S) are correct? (Multiple choice)

Applied in the upper ap plica t ion of Fusioninsight HD

HFS enc a psulates the i n terface between HBase and HDFS

Provide file storage, read, and dele t e functions for up p er- l e v el applicatio n s

HFS is an indepe n dent module of HBase

Page 9 of 12

81. Which of the following desc r iptions of the cloud search serv i ce CSS i s c orrect? (Multiple choice)

E xpansion of nodes and disk services will not b e interrupte d , and the thesaurus will be updated w i thout interruption

Compatible with Elasticsearch APIs

Does not support the fusion search of text such as sound

TB-level d ata retrie v al is returned in seconds

82. Which of the following state m ents about worke r , Executor and task are correct? (Multiple choice)

E a c h Ex e c utor can run a task

E a c h Ex e c utor can run tasks of di f ferent comp o nents (spout o r bolt)

Each worker can run multiple Ex e cutors

Each worker can only run Executor for one topo l ogy

83. Which of the following descriptions of Flink key features a re wro n g?

Compar e d with Flink, SparkStreaming has low e r latency

The Fli n k streaming e n gine can pr ov ide functions that suppo r t stream processing and batch processing applications at the same t i me

Compar e d with Streaming in Fusionlnght HD, Flink has hi g her thro u gh p ut

checkpoint realizes F li nk's fault tolerance

84. Which of the following are important roles of Spark? (Multiple choice)

DateNode

Nodema n ager

Driver

Resourc e Manag e r

85. Which of the following designs are mainly cons i dered in the planning process of the b ig data business co n s ulting servi c e plan? (Mul t iple choice)

Big data business mo d el design

Panoramic planning of business scenarios

T ypical scenario r equi r ements design

Scenario decomposition and busi n ess design

86. In order to improve the fault tolerance of ka f ka, kafka supports partition r eplication st r ateg y .

What is wro n g with the following description of Leader p a rtition and F o llower partition?

It is impossible for each node of a Kafka cluster to be the leader a n d flower of each other

If the leader fails, then ano t her follower will take over (becoming the new l ead e r)

Since the leader Server carr i es all the request pressure, from the overall consideration of the cluste r , Kafka will distribute the leader on each instance in a balanced manner to e nsure stable overall perform a nce

Kafka needs to sele c t a leader for partition repl ic ation, and the lead e r is respon s ible for the read a nd write operations of the partition, and other replica nodes are on l y responsible for da t a synchronizati o n

87. In the MapR e duce p r ocess, which of the following are shu f fle mechanisms? (Multiple choice)

partition

sort/merge

copy

combine

88. Which o perations in Hive can be merge d ? (Multiple choice)

UNION A LL

GROUP B Y

SELECT

JOIN

89. Using t h e Streaming client s hell command in FusionInsight HD system t o view the topology or submit the topology fails, which of the following positioning methods are correct? (Multiple choice)

Check the Supervisor ope r ation log to determine whether the Supervisor is abnormal

V iew worker runn i ng l o gs

Check the client exception stack to determine if there is a pr o blem with the client

Check the running log of the main nimbus to determine wheth e r the nim b us server is abn o rmal

90. Fusioni ns ight Spark SQL, like the community Spark JDBCServe r , only supports single tenant b i nding to a Y ARN resource queue multi-tenant, and does not support mu l ti-tenant p a rallel execution.

T rue

False

Page 10 of 12

91. The core c oncept of MapRe d uce is to decom p ose a large comp u ting task to each n o de of the cl u s te r , make full use of cluster resourc e s, and shorten the running tim e.

T rue

False

92. Assuming that the data volume i s about 200GB and the maximum fragment c apacity is limited to 3 0 GB, what is the a ppropriate design for the maximum number of fragments?

9 or 10

7 or 8

3 or 4

5 or 6

93. Which of the following s tatements about the re ad a n d write process of Zookeep e r's L ead e r n o de after

receiving the data change request is correct?

W rite to d i sk and memory at the s a me time

W rite to t h e disk first and then to the memory

W rite to m emory only

W rite to t h e memory fi r s t and then to the disk

94. H u awei has extensive experience in big data solutions in fi n ance, public securit y , railways, government a f fairs and other industri e s.

T rue

False

95. Fusioni ns ight Manger externally supp o rts REST interface, SNMP interface, and SYSLOG interface.

T rue

False

96. In the p r ocess of data transmission by Flume, after sink takes the data and writes i t to the destination, events will b e dele t ed from the channel.

T rue

False

97. The g ra phical health inspection tool Fusioni ns ight T ool is composed of FusionCare and SysChecke r .

T rue

False

98. Which of the following is not a required option when creating a Loa d er jo b ?

Name

Connection

T ype

Priority

99. After submitting the topology usi n g the Streaming client sh e ll command in the Fusioninsight HD system, usi n g the Strom UI to view i t, it i s found that the to p ology has not processed data for a long time.

What are the possible rea s ons? (Multiple choice)

Supervis o r is a component that receives data in the topology and then p e rforms processing

There is a logic error in the topology business, a nd it cannot run normally after submission

The topology is too c o mplica t ed or the concurr e ncy is too la r ge, which causes the worker startup t i me to be too long, exceeding the w a iting t i me of the S u pervisor

Supervisor's slot reso u rces a re exhausted, and no slots can be allocated to start the worker proce s s after the topology is submitted.

100. In the HDFS mechanism, the NameNode is responsible for man a gi n g metadata. Each read r eq uest on the client side needs to read met a data information from the metadata disk of the NameNo d e to obtain the location of the read file in the DataNode.

T rue

False

Page 11 of 12

101. Which of the following targe t s can Fusioninsight HD Loader export HDFS data to? (Multiple choice)

S F TP server

F T P s erver

Oracle database

DB2 database

102. Which two roles are r esponsible for HDFS's NameNo d e n o de ac t ive and standby st a te manag e ment and metada t a file me r ging respectively?

ZKFC and standby NameNode

Primary NameNo d e a n d standby NameNo d e

ZKFC and the main NameNode

Main NameNo d e and Jour n alNo d e

103. Wh e n performing H Base data reading service, which parts of data need to be read? (Multiple choice)

HFile

HLog

MemSto r e

HMaster

104. Which of the following nodes are included in the basic sys t e m architectu r e of HDFS? (Multiple choice)

ResourceManag e r

NameNo d e

NodeMa n ager

Data No d e

105. serve r .chan n els. c h1.transactioncapacity repr es ents the transaction s i ze, that is, the number of events that the current channel sup p o r ts for transaction proce s s i ng. It can be set to the s a me size a s t he batchsize of the source, but not less t h an the batchsize.

T rue

False

106. Which of the following options are included in the time operation types s u pport e d by Flink? (Multiple choice)

End t ime

Proce s s i n g time

Acquisiti o n time

Event time

107. Now ser v e r .chan n els=ch1 is avail a ble, set the Cha n nel type t o File Channel, which of the following configuratio n s are correct?

serve r .channels.ch1.type = file

serve r .channels.ch1.type = memo r y

serve r .channels.type = memory

serve r .channels.type = file

108. In the F usioninsight cluster networking de s i g n, the Layer 2 networking refers to the Layer 2 switching in the cluste r . The cl u s ter nodes are in a subnet, which is su i table for cluster scenarios with nodes less than 200.

T rue

False

109. Spark, like Hadoop, i s not suitable for iterative computing.

T rue

False

110. Which of the following sta t ements about Spar k s treaming and Streaming are w r ong?

SparkStreaming is a micro-ba t ch processing fr a mework. Ev e nts need to be processed only when t h ey are exhaust e d to a certain amount.

The execution logic of Strea m ing is to s tart im m ediatel y , a n d then recycle after running.

The throughput of SparkStreaming is about 2-5 times that of Streaming

SparkStreaming event processing latency is higher than Str e aming

Page 12 of 12

111. Spark Streaming calculation is based on DStr e am, which decomposes streaming c a lculations into a series of short batch processing jobs.

T rue

False

112. Which of the following can be viewed in the Loader history job record? ( M ultiple choice)

Job status

Job start/ r un time

Dirty data link

Number of error lines/f i les

113. Re g ardi n g the the configuration fu nc tion of Fusioninsight Manag e r to service, which of the following is incorrect?

Serv i ce-level configuration can take e f fe c t for all instances

I n stance-level configuration only takes e f fect for this instance

Instance- l evel configuration is also e f fect i v e for other in s tances

After the configurati o n is s a v ed, the service needs to be resta r ted to take e f fect

114. Which of the following is c o rrect about the LdapServer org an ization model? (Multiple choice)

No specific obje c ts can be defined u nder the o rg anization u nit.

E a c h node in the LDAPServer directory tree is called an entr y , and has its own uniq u e distinguished name DN (D is tinguished Name).

The root of the LDAPServer direct o ry tree gene r ally defines the domain n ame DC (D o main Compo n ent).

LDAPServer directory information is organized and stored ba s ed on a tr e e structure.

115. Which w a y is incorrect to load data to Hive table?

Load the files in the local pa t h directly into the Hive tab l e

Load the files on HDFS into the Hive table

Hive supports insert into a single record method, so you can insert a single rec o rd direct l y on the command line

Insert the result set of other tables into the Hive table

116. In Fusioninsight prod u cts, about K a fka T opic, w h ich of the following descr i ptions are incorrect?

E a c h topic can only b e d ivided into one partiti o n

The number of T opic p artitions can be configur e d at the time of creation

The storage level of each Par t ition corresponds to a log fi l e, which recor d s all information and data

E very message published to Kafka has a c ategor y , which is c a lled T opic, which c an also be understo o d as a queue for storing messages.

117. In the Fusioninsight HD syst e m, HBase supports dynamically expandi n g columns.

T rue

False

118. Wh e n importing da t a to the H i ve table, the validity of the d ata will not b e checked, o nly when the data is read.

T rue

False

119. The reliability of flum e 's da t a transmission me a ns that during the process of flume's data transmissio n , when the flume node of the next hop fails or the d ata is received ab n ormal l y , it can automatically switch to another way to continue tra n smission.

T rue

False

120. In the HDFS copy placement strateg y , what is the distance between di f ferent servers in the same rack?