Option 2 is the correct answer. To minimize the steps required to boot a well-configured VM, you need to boot multiple virtual machines with a machine image. First create a machine image with a non-standard library of images that is close to what you need. Then launch an instance with that machine image.

A machine image is a Compute Engine resource that stores all configurations, metadata, permissions, and data from multiple disks in a virtual machine (VM) instance. Machine images can be used in many system maintenance, backup and restore, and instance cloning scenarios.

Options 1 and 3 are incorrect. The VM generated from the optimal machine image provided by Google Cloud is not the optimal virtual machine configuration for the user. You need to install your own library to optimize the VM.

Options 1 and 4 are incorrect. A snapshot is a backup used when restoring an optical disc. Not available when replicating a virtual machine.

Option 4 is the correct answer. By default, Google Cloud Compute Engine instances are virtualized into their separate environments so they don't interfere with each other without you having to do anything.

Therefore, no special action is required.

【reference】

https://cloud.google.com/compute/docs/instances/

https://cloud.google.com/compute

Option 3 is the correct answer. You can add new members by specifying individual users (email addresses) and Google groups in the parameter "new members" on the IAM console screen.

Identity and Access Management (IAM) allows you to create and manage permissions for Google Cloud resources. IAM integrates access control for Google Cloud services into one system to provide consistent operations.

Options 1 and 2 are incorrect. You cannot enter a role in "New Member".

Option 4 is incorrect. The "New Member" parameter allows you to enter groups as well as individual users.

【reference】

https://cloud.google.com/iam/docs

Option 3 is the correct answer. Only basic and predefined roles can be searched for permission references to see if a role grants permissions. This process for checking the permissions of a custom role is incorrect. Use one of the following methods to determine if a role contains the appropriate permissions:

■ Run the gcloud iam roles describe command to list the permissions for the role.

■ Invoke the roles.get () REST API method to list the permissions of the roles.

■ Basic and predefined roles only: Search the permission reference to see if the role grants the permission.

■ Predefined roles only: See the description of the predefined roles to see the permissions contained in the role.

Also, select [Role] on the IAM and administrator screens. You can also select the check box next to a role to display the permissions for that role.

【reference】

https://cloud.google.com/iam/docs/understanding-roles?skip_cache=true

Option 5 is the wrong choice. The engine type is not used as a parameter when creating a cluster. A Kubernetes cluster is a collection of nodes that run containerized applications. You can containerize your application to package your application with application dependencies and multiple required services. Kubernetes Engine allows you to specify projects, regions, zones, machine types, number of nodes, etc. when creating a cluster.

It cannot be changed after the cluster is created. These choices affect cluster availability, version stability, and network.

Options 1, 2, 3 and 4 are items used when creating a cluster in Kubernetes Engine. Kubernetes Engine allows you to specify projects, regions, zones, machine types, number of nodes, etc. when creating a cluster.

【reference】

https://cloud.google.com/sdk/gcloud/reference/container/clusters/create?hl=j

Option 2 is the correct answer. The name of the pod spec you created in the previous step. The configuration information about how to communicate with the API is stored in the kubeconfig file. If you run multiple clusters within your Google Cloud project, you need to choose which cluster will communicate with kubectl. To do this, you can set the default cluster for kubectl by setting the current context in the Kubernetes kubeconfig file. Kubernetes uses a YAML file called kubeconfig, and stores the kubectl cluster credentials. kubeconfig contains a list of contexts referenced by kubectl when command is excecuted. By default, this file is saved in $ HOME / .kube / config

Option 1 is incorrect. config.yaml is a commonly used configuration file name, but not a Kubernetes configuration file.

Option 3 is incorrect. pod-file-name.yaml is a file that defines the name of the pod specification created in the previous step.

Option 4 is incorrect. The file kbconfig..yaml does not exist.

【reference】

https://cloud.google.com/kubernetes-engine/docs/how-to/cluster-access-for-kubectl

Option 3 is the correct answer. App Engine allows you to choose whether to use IP address splitting or HTTP cookie splitting when splitting traffic between two or more versions of your application. When you have specified two or more versions for splitting, you must choose whether to split traffic by using either an IP address or HTTP cookie. It's easier to set up an IP address split, but a cookie split is more precise.

■ IP address split For IP address splitting, when an application receives a request, it translates the IP address into a hash value in the range 0-999 and uses that number to route the request.

■ HTTP cookie split In the case of HTTP cookie splitting, the application routes the request based on the value stored in the cookie called GOOGAPPUID in the HTTP request header. If the cookie GOOGAPPUID does not exist, the request will be routed randomly.

The traffic split feature allows you to specify a traffic distribution ratio to distribute traffic to multiple versions of the same service.

【reference】

https://cloud.google.com/appengine/docs/standard/python3/splitting-traffic

Option 4 is the correct answer. Drives connected to virtual machines use block storage as disks. Virtual storage, which is a block storage resource, provides persistent disk space that does not lose its contents when a virtual machine instance is stopped / destroyed. Random access to files is possible by using block storage. Option 1 is incorrect. The connected drive is either an SSD or a hard drive. Therefore, it is not limited to SSD storage only.

Option 2 is incorrect. Object storage is a storage type used in Cloud Storage, etc. It cannot be configured as a drive directly attached to the virtual machine.

Option 3 is incorrect. File storage is a storage type used when making a file system type storage configuration. It cannot be configured as a drive directly attached to the virtual machine.

[Reference]

https://cloud.google.com/compute/docs/disks/performance

Option 3 is the correct answer. Containers provide maximum flexibility for efficient use of cluster resources. The orchestration platform reduces operational overhead. Therefore, efficient use of resources is possible by using containers in the cluster that is managed by the user.

Google Cloud uses Kubernetes to perform container orchestration. Kubernetes creates containers to increase the reliability of your compute configuration, reducing the time and resources you spend on DevOps, while also reducing the load associated with these tasks.

Option 1 is incorrect. The load balancer is used to distribute traffic between instances. Configuring a load balancer does not avoid the overhead of orchestration operations.

Option 2 is incorrect. It is not recommended to run in a single virtual machine as all services will go down if the server fails.

Option 4 is incorrect. As with option 2, it is not recommended to run in a single virtual machine as all services will go down if the server fails.

【reference】

https://cloud.google.com/containers/

Option 1 is the correct answer. BigQuery is a data warehousing and analytics database managed service that minimizes administrative overhead. You don't even have to configure storage. You can also perform expansion automatically when you expand your data warehousing infrastructure.

Option 2 is incorrect. Bigtable is a NoSQL database and cannot execute the large analytic queries possible in a data warehouse.

Option 3 is incorrect. Cloud Dataflow is a stream and batch processing service that cannot execute the large analytical queries possible in a data warehouse.

Option 4 is incorrect. Spanner is a relational database designed for scaling. All the features required of relational databases, such as schemas, SQL queries, and ACID transactions, have been proven and can be scaled globally. It can be used universally as an application database, but it is not a service for use as a data warehouse.

【reference】

https://cloud.google.com/bigquery