Machine Learning in GCP Online Test

Sam Sankman works as Cloud Associate at FTXX Inc. FTXX’s core application is deployed in a Google Kubernetes Engine cluster. When a new version of the application is released, CI/CD tool updates the spec.template.spec.containers[0].image value to reference the Docker image of the new application version. When the Deployment object applies the change, Sam wants to deploy at least 1 replica of the new version and maintain the previous replicas until the new replica is healthy. Which change should Sam make to the Kubernetes Engine Deployment object shown below?

A: Set the Deployment Strategy to Rolling Update
B: Set the Deployment Strategy to Recreate

C: Set maxSurge to 0
D: Set maxSurge to 1

E: Set maxUnavailable to 0
F: Set maxUnavailable to 1

Depp works as Google Cloud Architect at Amber Inc. Deep created an instance with following firewall rules:

NAME: ssh-allow
NETWORK: testnet
DIRECTION: INGRESS
PRIORITY: 100
ALLOW: tcp:22

NAME: deny-everything
NETWORK: testnet
DIRECTION: INGRESS
PRIORITY: 400
DENY: tcp:0-65535,udp:0-6553

What would happen if Depp tries SSH to the instance now?

A: SSH would be denied as the deny rule overrides the allow
B: SSH would be allowed as the allow rule overrides the deny
C: SSH would be denied and would need instance reboot for the to allow rule to take effect
D: SSH would be denied and would need gcloud firewall refresh command for the allow rule to take effect

Kloo is a social media platform similar to Twitter in the EU market and uses Google Cloud for their infrastructure requirements. Kloo generates multiple files per day. Owing to data regulations, Kloo has to store these files in the EU regional bucket. Kloo’s data team needs only the files of last 14 days for their algorithms and APIs but need to store every single file for safe keeping and regulatory purposes even though the past data is rarely accessed. Given these constraints, how can Kloo’s data team lower the storage costs?

You work as a data manager for a large e-commerce company that heavily relies on a PostgreSQL database running on a GCP Compute Engine instance. The company has a well-established backup policy, which includes:

1. Full database backup every Monday at 1:00 AM.
2. Differential backup every day at 1:00 AM, except Monday.
3. Transaction log backup every 30 minutes.

These backups are stored in Google Cloud Storage. Today is Thursday, and a data corruption incident occurred at 4:15 PM. You have the following backup files available:

1. Full backup: Full_Backup_Mon.bak taken on Monday 1:00 AM.
2. Differential backups: Diff_Backup_Tue.bak, Diff_Backup_Wed.bak, Diff_Backup_Thu.bak taken at 1:00 AM on their respective days.
3. Transaction log backups: 30-minute interval backups from Monday 1:30 AM until Thursday 4:00 PM, like TLog_Backup_Thu_1530.bak, TLog_Backup_Thu_1600.bak.

Given the RPO (Recovery Point Objective) of 30 minutes, which of the following sequences of restore operations would ensure minimal data loss?

A: Full_Backup_Mon.bak, Diff_Backup_Thu.bak, then Transaction Log backups from Thursday 1:30 AM to 4:00 PM.
B: Full_Backup_Mon.bak, Diff_Backup_Wed.bak, then all Transaction Log backups from Wednesday and Thursday.
C: Full_Backup_Mon.bak, Diff_Backup_Thu.bak, then all Transaction Log backups from Thursday.
D: Full_Backup_Mon.bak, then all Transaction Log backups from Monday to Thursday 4:00 PM.
E: Full_Backup_Mon.bak, Diff_Backup_Tue.bak, Diff_Backup_Wed.bak, Diff_Backup_Thu.bak, then Transaction Log backups from Thursday 1:30 AM to 4:00 PM.