X-Past Exams Flashcards

1
Q

What happens to surplus cached data at Edge Locations?

A

If it hasn’t been used recently, it gets ejected

If a Regional Edge Cache is configured, then it is ejected into a local S3 bucket

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

How are High Availability and Fault Tolerance difference?

A

HA is resilient to component and AZ loss

Fault tolerant is HA but also no performance impact of losing an entire AZ

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How does root user and IAM user login to the console differ?

A

Root users use an email address, IAM users use a username (though it may be an email)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is notable about IPv6 addresses in VPCs?

A

They are all public

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

To what extent can NACLs be reused?

A

They can apply to multiple subnets, but only within one region

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Can an instance have more than one security group?

A

Yes - this allows more traffic

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Can a VPC have more than one IGW?

A

No

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Can a VPC have more than one NAT Gateway?

A

Yes - in fact, they should have one for each AZ

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

In what order at NACL rules evaluated?

A

Lowest to highest

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

To what extent can security groups be reused?

A

They can be shared across instances in multiple subnets if they are in the same VPC

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is Neptune?

A

A NoSQL graph database service

It uses SPARQL or Gremlin for queries

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is Amazon MQ

A

A managed service for Apache MQ - is supports push and pull based operation

It is single instance sensitive - a standby is available but it can’t scale

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is special about EMR?

A

It can use S3 for storage and allows for transient clusters

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Can can the performance impact of RDS back-ups be minimised?

A

With MultiAZ deployments, the backup can be taken from the secondary instance

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Does MultiAZ RDS improve performance?

A

Only in the sense that it minimises the impact of backups

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is the SLA for EC2?

A

99.99% availability per month for ECS and EBS

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What S3 data transfers are free?

A

All data in, transfers between buckets and other AWS services if in same region, transfers to CloudFront

(also the first 1 GB is free tier)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What is notable about the names for transfer acceleration buckets?

A

They can’t contain full-stops

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

Which AWS services give root access to the machine?

A

EC2 and EMR

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What types of virtualisation exist on EC2?

A

Hardware Virtual Machine (HVM) and Paravirtual Machine (PM)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

Is S3-Z-IA durable?

A

Yes, it maintains 11 9’s durability

22
Q

What are the steps to setting up EFS on an instance?

A

Open Port 2049, mount it, and set the Linux file permissions

23
Q

What happens when an EC2 instance is stopped and started again?

A

The underlying host changes and the instance store is lost

The public IP will change. The private IP won’t

24
Q

What happens when an EC2 instance is rebooted?

A

The host does NOT change and the instance store is not wiped

The public IP will not change

25
Q

What SNS alert is useful when debugging auto-scaling groups?

A

Look for EC2_INSTANCE_LAUNCH-ERROR

26
Q

What is different about MariaDB?

A

It can’t combine read replicas with MultiAZ

27
Q

Do changes to RDS require downtime?

A

Increasing storage does not but scaling up the instance does

28
Q

What are some special features of Aurora?

A

It has continuous backups. Read-replicas share the underlying storage so there isn’t replication lag

They support both MySQL and PostgreSQL

There is a serverless offering in which you don’t worry about instances

29
Q

What are ALB path patterns used for?

A

Routing different URL parts to different servers

30
Q

What could cause a Classic Load Balancer do not share load across its instances?

A

Cross-Zone Load Balancing has not been enabled

31
Q

Does RDS support custom plugins?

A

No

32
Q

What are the limits on SQS message size? How does this affect pricing?

A

They can be up to 256 KB but are charged in blocks of 64 KB

33
Q

How long are CloudWatch metrics of deleted instances retained for?

A

15 months

34
Q

What happens when SAML is used to log into the console?

A

The user identity is verified before generating a response

A sign-in URL to the console is set as a redirect

35
Q

What does a VPN into AWS require?

A

A Virtual Private Gateway, an on-premises Customer Gateway, and a VPC with Hardware VPN access

36
Q

What are the tenancy options for EC2?

A
  • Default (shared)
  • Dedicated (just you)
  • Host (you control config + same server each time)
37
Q

Can tenancy options on EC2 be changed?

A

You can only switch between dedicated and host, but this requires restarting the instance

38
Q

In practical terms, how resilient is S3?

A

All storage classes except S3-Z-IA are designed to sustain the complete loss of an AZ, or partial loss of two

39
Q

At what level do Route Tables apply?

A

They apply to subnets, not instances

40
Q

What are the steps to securely add an instance to the internet?

A
  • Create a route from the instance’s subnet to a NAT Gateway

- Create a route from the NAT Gateway to an IGW

41
Q

Are IGWs highly-available?

A

Yes, then span AZs

42
Q

What should be used to prevent against single-instance failure?

A

An autoscaling group

43
Q

How does EC2 auto-recover work?

A

It monitors CloudWatch alarms.

Note that it preserves the IP addresses of instances

44
Q

Should you put load balancers in a database subnet?

A

No - there’s no point

45
Q

How can traffic to host-names be blocked?

A

Use a proxy server - NACLs can’t block host-names

46
Q

How can network throughput between instances be maximised?

A

Use enhanced networking and a clustered placement group

47
Q

What performance does S3 provide?

A

5500 requests per second to read data; 3500 requests per second to write

48
Q

Why might newly created EBS volumes be slow?

A

If created from a snapshot, they are slow each time a block is read for the first time

49
Q

Can CloudFront cache dynamic content?

A

Yes if it is based on query strings

50
Q

Can load balancers work with more than one https hostname at once?

A

ALBs can; Classic Load Balancers can’t because they can only have one SSL certificate at a time

51
Q

Should on-prem be used with CloudFront?

A

Not really - it’s usually cheaper just to use S3

52
Q

Can the type of an EC2 instance be changed?

A

Yes - this requires a restart and that they have similar architectures etc.