Verified & Latest 1Z0-1089-21 Dump Q&As with Correct Answers [Q18-Q42]

4.5/5 - (2 votes)

Verified & Latest 1Z0-1089-21 Dump Q&As with Correct Answers

Latest 1Z0-1089-21 dumps – Instant Download PDF

NO.18 Which Message Passing Interface (MPI) distribution is NOT available today?

OCI MPI

Intel MPI

Platform MPI

Open MPI

NO.19 What are two different types of Storage offerings available to build a filesystem on Oracle Cloud infrastructure (OCI)?

Block Volume Service

Local NVMe SSDs on DenselO and HPC compute shapes

File Storage

Object Storage

NO.20 Acustomer is hesitating between running BM.Standard.E3.128 or HPC2.36 for his workload. Which statement would disqualify the AMD chip without any benchmark?

IO speed has a huge impact.

Total number of processes is between 48 and 128 cores and can useRDMA.

The workload requires about 4GB/core of memory.

The workload is embarrassingly parallel.

NO.21 Which two Oracle Cloud Infrastructure (OCI) shapes are a good choice for Spark based workloads?

HPC

VM Standard

BM Standard

GPU

NO.22 You are building a file system that needs to handle large files with a lot of nodes reading at the same time.
What should be your main goal?

Maximize Latency while keeping throughput above a certain threshold.

Maximize throughput.

Minimize latency.

Minimize throughput while keeping the latency low.

NO.23 You are architecting the infrastructure for a file system.
What are the different criteria you should use, and in what order to build a filesystem for optimal performance?

Network Bandwidth > Number of Compute Cores/RAM >Storage

Storage > Network Bandwidth > Number of Compute Cores/RAM

Number of Compute Cores/RAM > Storage > Network Bandwidth

Network Bandwidth > Storage > Number of Compute Cores/RAM

NO.24 What three data sources can Oracle Data Science draw from?

Google Cloud Object Storage

Autonomous Data Warehouse

On-premise Oracle database

Azure ObjectStorage Blob

AWS S3 buckets

NO.25 Whichis an accurate analogy for Amdahl’s law?

There Is no such thing as a free lunch.

An orchestra can only play well together If they have a great conductor.

When chopping vegetable to prepare a 4 person-meal, 2 persons are going twice as fast as one, but 8 people may not be 8 times faster.

In track and field, the 4xl00m relay is only 17% faster than the world record for individual 400m.

NO.26 You are building a filesystem of 10TB Storage Capacity with only one file server using “Balanced Elastic Performance tier” Oracle Cloud Infrastructure (OCI) Block Volume Storage. Balanced Elastic Performance tier throughtput scales at 480 KBPS/GB up to a maximum of 480 MBPS pervolume.
Which two options are correct about how many Block Volumes you can attach to the file server to get maximum IO throughput?

One Block Volume of 10TB size

10 Block Volumes of 1TB size

8 Block Volumes of 1.25TB size

20 Block Volumes of 500GB size

NO.27 Oracle Cloud infrastructure (OCI) Block Volume Storage supportssharing a Block Volume among multiple compute Instances in read/write or read only shareable mode.
Which two file systems should be used to allow multiple compute Instances to read/write data concurrently without any data loss?

Distributed File Systemslike Gluster, OCFS2, GFS2

Linux File Systems like XFS, EXT4, EXT3, etc.

Network File System (NFS)

Parallel File Systems like Lustre, IBM Spectrum Scale(GPFS), BeeGFS, etc.

NO.28 You are building a file system that needs to handle large files with a lot of nodes reading at the same time.
What minimum size of block volume maximizes throughput for large files?

800GB

1TB

32TB

500GB

NO.29 What HDFS replication factor should be used for locally attached storage In HDFS?

NO.30 Compute nodes are on subnet2(l72.16.1.0/24) and there is a BM.Dense.I02.8 on subnet1.
(1 /2.16.0.0/24) that will serve as a bastion, file server and license server. The bastion will be accessed through a public IP with SSH.
Which security lists should you choose for an HPC cluster? Ifmultiple options are possible, choose the one limiting access the most.

Subnet1: 0.0.0.0 for all TCP/UDP ports, subnet2: 172.16.0.0/16 for all TCP/UDP ports

Subnet1: 0-0.0.0 for TCP port 22, subnet 1 and 2: 172.16.0.0/23 for all TCP/UDP ports

Subnet1: 172.16-1.0/24 for port 22, subnet2: 172.16.1.0/24 for all TCP/UDP ports

Subnet1: 0-0.0.0 for TCP port 22, subnet 2: 172.16.0.0/16 for TCP port 22

NO.31 A customer has a very busy workload. The model is very large (1 PB range) and only some small files are updated for new jobs. Throughput needed during the run is roughly 25GB/s.
What Is a fast and cost-conscious way to handle the file system?

Put the data in object storage, and mount It using s3fs-fuse project.

Build a file-system using NVMe on Dense shapes. Then move the data to object storage when not needed.

Build a file system using Block volumes and Standard BMs, take advantage of the different block volume performances levels.

Use NVMe on HPC shapes to build a File Systemwith the RDMA connection.

NO.32 On Oracle Cloud Infrastructure (OCI), a customer wants to build a3TB filesystem for high IOPS workloads (like relational databases, small files, etc).
Which action provides the highest IOPS using OCI block volumes for storage?

Attach one Block volume of 3TB volume size and use filesystem Block size of 4K.

Attach one Block volume of 3TB volume size and use filesystem Block size of 2M or higher.

Attach three Block volumes of 1TB each and use filesystem Block size of 2M or higher.

Attach three Block volumes of 1TB each and use filesystem Block size of 4K.

NO.33 When an instance pool scales in, what is the correct order that the instances are terminated?

The number of instances is balanced across availability domains, the oldest instance within the fault domain is terminated, the number of instances Is balanced across fault domains.

The number of instances is balanced across availability domains, the number of instances is balanced across fault domains, the oldest instance within the fault domain is terminated.

The oldest instance within the fault domain is terminated, the number of instances is balanced across availability domains, the number of instances is balanced across fault domains.

The number of instances is balanced across fault domains, the number of instances is balanced acrossfault domains, the oldest instance within the fault domain is terminated.

NO.34 A two node clustered/parallel filesystem built using Baremetal Compute shape (BM.Standard2.52) and Block Volume Storage delivers maximum sustained aggregate IO throughput of 5GB/s for read and write, while storing a single copy of the data (replica=l) in filesystem.
What will be the maximum sustained aggregate IO throughput for read and write, when two copies of the data (replica=2) are stored in filesystem for High Availability?

5GB/s for both (read and write)

2.5GB/s for readand 5GB/s for write

2.5GB/S for write and 5GB/s for read

2.5GB/s for both (read and write)

NO.35 A customer has a Spark workload they areconsidering moving to Oracle Cloud Infrastructure (OCI). What option should they consider?

Oracle Data Flow

Self-managed Hadoop deployment

Self-managed Spark deployment

Big Data Service

NO.36 You are building a filesystem to maximize throughput of large files using high performance block volume and BM.Standard2.52 as file server.
Which block volume configuration should you choose to ensure aggregate Block volume throughput is higher than network bandwidth of a file server?

32 Block Volumes of 800GB

5Block Volume of 32 TB

6 Block Volumes of 32 TB

7 Block Volumes of 800GB

NO.37 When running a high memory workload, what should be your machine of choice?

BM.Standard.E2.64

BM.Standard.E3.128

BM.HPC2.36

BM.GPU3.8

NO.38 You are building a large filesystem to maximize throughput of large files using high performance block volume and BM.Standard2.52 as file servers.
Which block volumes configuration should you choose to ensure aggregate Block volume throughput is higher than network bandwidth of file servers?

7 Block Volumes of 800 GB

32 Block Volumes of 800 GB

5 Block Volumes of 32 TD

6 Block Volumes of 32 TB

NO.39 What does the Terasort phase of the Terasort benchmark do?

Itmaps and reduces source data in a parallel manner, leveraging data locality to minimize network transfer.

It randomly maps source data and reduces output to a smaller data set,

It maps and reduces one terabyte of data into a smaller data set.

It randomly maps source data and increases output to a bigger data set.